University of Ghana http://ugspace.ug.edu.gh ULTRASOUND AND PET-CT IMAGE FUSION FOR PROSTATE BRACHYTHERAPY IMAGE GUIDANCE BY FRANCIS HASFORD University of Ghana http://ugspace.ug.edu.gh ULTRASOUND AND PET-CT IMAGE FUSION FOR PROSTATE BRACHYTHERAPY IMAGE GUIDANCE This thesis is presented to the UNIVERSITY OF GHANA LEGON BY FRANCIS HASFORD [ ID: 10191693 ] In partial fulfillment of the requirement for the award of PHD MEDICAL PHYSICS DEGREE JUNE, 2015 i University of Ghana http://ugspace.ug.edu.gh DECLARATION Candidate’s Declaration This thesis is the result of research work undertaken by Francis HASFORD of the Department of Medical Physics, School of Nuclear and Allied Sciences, University of Ghana, under the supervision of Prof. John Humphrey AMUASI, Prof. Augustine Kwame KYERE and Prof. Mboyo Di Tamba VANGU. Signed:…………………………………. Date: ………………………… Francis HASFORD (PhD Candidate) Supervisor’s Declaration We hereby declare that the preparation and presentation of this thesis were supervised in accordance with guidelines on supervision of thesis laid down by the University of Ghana Signed:…………………………………. Date:………………………… Prof. John H. AMUASI (Principal Supervisor) Signed:…………………………………. Date:………………………… Prof. Augustine K. KYERE (Co-Supervisor) Signed:…………………………………. Date:………………………… Prof. Mboyo D.T. VANGU (Co-Supervisor) ii University of Ghana http://ugspace.ug.edu.gh ABSTRACT Fusion of medical images between different cross-sectional modalities is widely used, mostly where functional images are fused with anatomical data. Ultrasound has for some time now been the standard imaging technique used for treatment planning of prostate cancer cases. While this approach is laudable and has yielded some positive results, latest developments have been the integration of images from ultrasound and other modalities such as PET-CT to compliment missing properties of ultrasound images. This study has sought to enhance diagnosis and treatment of prostate cancers by developing MATLAB algorithms to fuse ultrasound and PET-CT images. The fused ultrasound-PET-CT image has shown to contain improved quality of information than the individual input images. The fused image has the property of reduced uncertainty, increased reliability, robust system performance, and compact representation of information. The objective of co-registering the ultrasound and PET-CT images was achieved by conducting performance evaluation of the ultrasound and PET-CT imaging systems, developing image contrast enhancement algorithm, developing MATLAB image fusion algorithm, and assessing accuracy of the fusion algorithm. Performance evaluation of the ultrasound brachytherapy system produced satisfactory results in accordance with set tolerances as recommended by AAPM TG 128. Using an ultrasound brachytherapy quality assurance phantom, average axial distance measurement of 10.11 ± 0.11 mm was estimated. Average lateral distance measurements of 10.08 ± 0.07 mm, 20.01 ± 0.06 mm, 29.89 ± 0.03 mm and 39.84 ± 0.37 mm were estimated for the inter-target distances corresponding to 10 mm, 20 mm, 30 mm and 40 mm respectively. Volume accuracy assessment produced measurements of 3.97 cm3, 8.86 cm3 and 20.11 cm3 for known standard volumes of 4 iii University of Ghana http://ugspace.ug.edu.gh cm3, 9 cm3 and 20 cm3 respectively. Depth of penetration assessment of the ultrasound system produced an estimate of 5.37 ± 0.02 cm, indicating the system’s ability to visualize low contrast objects 5.4 cm into a patient. PET-CT system’s performance evaluation also produced satisfactory results in accordance with set tolerances as recommended by IAEA Human Health Series 1. Computed tomography laser alignment test ensured that all CT gantry lasers were properly aligned with the patient bed. Image display width test ensured that volume of patient or organ being measured and displayed was equivalent to that selected on the CT scanner console, to a deviation of ± 1 mm. Results from CT image uniformity test showed that mean CT numbers in peripheral regions of interest deviated from the central mean to within recommended tolerance level of ± 5 HU, indicating a good level of uniformity. Computed tomographic dose indices for head and body phantoms were estimated as 44.30 mGy and 20.08 mGy, comparative to console displayed doses of 42.40 mGy and 19.49 mGy respectively. Registration accuracy for PET-CT images was to have displacements of less than 1 mm in x, y and z directions. Image quality of PET-CT images was performed to produce images simulating those obtained in a total body imaging study involving both hot and cold lesions. Percentage contrast estimates of 49.3% and 52.6% were obtained for hot spheres of diameters 1.3 cm and 2.2 cm respectively, while contrast estimates of 74.8% and 75.6% were obtained for cold spheres of diameters 2.8 cm and 3.7 cm respectively. The PET-CT system resolution was estimated as 0.5 ± 0.01 cm, indicating the system’s ability to image tumours of the size of about 5 mm. Satisfactory results from the performance evaluation of ultrasound and PET-CT systems, paved way for them to be used in acquiring prostatic images for the study. Developed MATLAB image enhancement algorithm enhanced the quality of prostatic images before fusion. The algorithm iv University of Ghana http://ugspace.ug.edu.gh was developed by mapping the intensity values in raw images to new values in a modified image using imadjust function. Contrast enhanced prostatic images of ultrasound and PET-CT were then co-registered with developed MATLAB fusion algorithm. The fusion algorithm was developed on the theory of mutual information and rigid body transformation. Fused image of ultrasound and PET-CT in this study has been assessed to have well defined and good visualized prostate capsule, urethra and implanted seeds, which would otherwise not be the case in either of the two images separately. The resultant image could therefore produce much more accurate results in treatment planning of prostate cancer cases. Assessment of image registration error for the ultrasound-PET-CT fused image produced a root mean square error estimate of 1.3 mm. v University of Ghana http://ugspace.ug.edu.gh DEDICATION This thesis is dedicated to the three special ladies in my life: my mum Margaret Eshun, my wife Stella A. Hasford and my daughter Katherine Yaa Eshun Hasford. Without their prayers, sacrifices and support I would not have achieved this dream. The study is also dedicated to the memory of the late Emmanuel Kwaku Nani, under whose encouragement and guidance this work began. It is unfortunate he is not alive to see the fruit of the seed he sowed in me. May his gentle soul rest in peace. vi University of Ghana http://ugspace.ug.edu.gh ACKNOWLEDGEMENT First of all I would like to thank the persons who made this work possible, my supervisors Prof. J.H. Amuasi, Prof. A.K. Kyere and Prof. M.D.T. Vangu. These distinguished persons have not only been supportive throughout my PhD studies, but have also been strong believers of my visions and ideas. Prof. Amuasi has been my advisor not only in the field of Medical Physics, but life in general, and has also been a spiritual father for me and my family. Prof. Kyere has been a father and great mentor in my life. His enthusiasm and sound advice has given me confidence in pursuing my ideas, and has given me excitement in my work at times when I encountered challenges. Prof. Vangu has been a wonderful counselor, ensuring the availability and accessibility of facilities for this study and also bringing his rich expertise as a nuclear medicine consultant on board. Words cannot express my gratitude to these great men making up a wonderful supervisory team. Mr. Bronwin Van Wyk deserves a special mention for his selfless and dedicated contribution to this study. I had the opportunity of firsthand experience on PET-CT systems under his guidance. Bronwin has been with me in all through the experimental work, sharing ideas on the direction of this study, and sometimes reviewing manuscripts produced out of our results. I am very grateful for having him as a colleague who has always been willing and ready to offer the necessary support in data collection. Going into the future, I look forward to working closely with him. I would like to thank management of the departments of Nuclear Medicine, Medical Physics, Radiation Oncology, and Radiology of the of CM Johannesburg Academic Hospital – South Africa, for permitting the use of their facilities for this study. Particular thanks to Mr. Thulani Mabhengu of the Medical Physics Department, technologists and registrars of the Nuclear vii University of Ghana http://ugspace.ug.edu.gh Medicine Department. Thanks also to the Radiotherapy Department of Korle-Bu Teaching Hospital for the use of their prostate brachytherapy unit for part of the studies. The effort of Mr. Ernest Eduful of Ghana Atomic Energy Commission is also acknowledged for his assistance in development of the MATLAB image enhancement algorithm. Next, I would like to thank my employer Ghana Atomic Energy Commission, my institute Radiological and Medical Sciences Research Institute, and my school Graduate School of Nuclear and Allied Sciences for the continued support in my pursuit of higher education. I acknowledge with thanks the immense contribution of the International Atomic Energy Agency (IAEA) for granting me fellowship to undertake part of this study at CM Johannesburg Academic Hospital / University of Witwatersrand in South Africa, under the supervision of Prof. M.D.T. Vangu. I also acknowledge the active support of the University of Ghana through the Office of Research Innovation and Development (ORID) in offering me the Faculty Development Grant Award for the study. Lastly, I would like to thank my family for their support and believe in me. Through thick and thin they have constantly stood by me. I will always be grateful to them especially my mother who has always supported me with prayers and encouragement. I am who I am because of her and her hard work. To my lovely wife who has been there for me the past three years, I say a very big thank you. No matter how difficult and frustrating times have been concerning my studies, she has always been there to calm my nerves and make me realize that sound academics thrive on love. viii University of Ghana http://ugspace.ug.edu.gh TABLE OF CONTENTS TITLE PAGE….. …… …… …… …… …… …… …… …… …… ……i DECLARATION …… …… …… …… …… …… …… …… …… …...ii ABSTRACT …… …… …… …… …… …… …… …… …… …… …..iii DEDICATION…. …… …… …… …… …… …… …… …… …… …..vi ACKNOWLEDGEMENT …… …… …… …… …… …… …… …… ….vii TABLE OF CONTENTS …… …… …… …… …… …… …… …… …..ix LIST OF FIGURES …… …… …… …… …… …… …… …… …… …xiv LIST OF TABLES …… …… …… …… …… …… …… …… …… …xvi LIST OF ABBREVIATIONS AND SYMBOLS …… …… …… …… …… ...xvii CHAPTER ONE: BACKGROUND 1.0 Introduction …… …… …… …… …… …… …… …… …… ……1 1.1 Thesis Motivation... …… …… …… …… …… …… …… …… ……3 1.2 Thesis Objectives... …… …… …… …… …… …… …… …… ……5 1.3 Benefits versus Risks …… …… …… …… …… …… …… …… ……6 1.4 Thesis Overview… …… …… …… …… …… …… …… …… ……7 CHAPTER TWO: LITERATURE REVIEW 2.0 Introduction …. …… …… …… …… …… …… …… …… ……8 2.1 The Prostate Gland …… …… …… …… …… …… …… …… ……8 2.2 Prostate Cancer… …… …… …… …… …… …… …… …… ……9 2.2.1 Prostate Cancer Staging…. …… …… …… …… …… …… …..14 2.3 Prostate Imaging Modalities …... …… …… …… …… …… …… …..16 2.3.1 Ultrasound …… …… …… …… …… …… …… …… …..16 2.3.2 Computed Tomography … …… …… …… …… …… …… …..18 2.3.3 Positron Emission Tomography..… …… …… …… …… …… …..21 2.3.4 Integrated PET-CT System …… …… …… …… …… …… …..23 2.3.5 Radionuclides in Prostate PET-CT Imaging …… …… …… …… …..24 2.4 Prostate Brachytherapy….. …… …… …… …… …… …… …… …..26 ix University of Ghana http://ugspace.ug.edu.gh 2.4.1 Low Dose Rate Brachytherapy….. …… …… …… …… …… …..28 2.4.1.1 Iodine-125 LDR Source…. …… …… …… …… …… …..31 2.4.1.2 Palladium-103 LDR Source …… …… …… …… …… …..32 2.4.2 High Dose Rate Brachytherapy..… …… …… …… …… …… …..34 2.4.2.1 Iridium-192 HDR Source... …… …… …… …… …… …..36 2.4.2.2 Cobalt-60 HDR Source..… …… …… …… …… …… …..37 2.4.3 Prostate Brachytherapy Seed Implant Techniques. …… …… …… …..38 2.4.4 Comparison of HDR and LDR Techniques …… …… …… …… …..40 CHAPTER THREE: THEORY OF IMAGE FUSION 3.0 Introduction …… …… …… …… …… …… …… …… …… …..41 3.1 Image Fusion Process …… …… …… …… …… …… …… …… …..41 3.1.1 Multiple Input Images Block …… …… …… …… …… …… …..42 3.1.2 Common Representational Format Block.. …… …… …… …… …..42 3.1.3 Fusion Block …… …… …… …… …… …… …… …… …..44 3.1.4 Display Block …… …… …… …… …… …… …… …… …..45 3.2 Principle of Mutual Information… …… …… …… …… …… …… …..45 3.2.1 Normalized Mutual Information… …… …… …… …… …… …..46 3.2.2 Calculation of Mutual Information …… …… …… …… …… …..47 3.2.3 Histogram… …… …… …… …… …… …… …… …… …..47 3.2.4 Parzen Windows…. …… …… …… …… …… …… …… …..48 3.2.5 Iso-intensity Lines.. …… …… …… …… …… …… …… …..50 3.2.6 Partial Volume Interpolation …… …… …… …… …… …… …..51 3.2.7 Artifacts…. …… …… …… …… …… …… …… …… …..52 3.3 Rigid Body Transformation …… …… …… …… …… …… …… …..53 3.3.1 Translation.. …… …… …… …… …… …… …… …… …..54 3.3.2 Rotation…. …… …… …… …… …… …… …… …… …..54 3.3.3 Zoom …… …… …… …… …… …… …… …… …… …..55 CHAPTER FOUR: METHODOLOGY 4.0 Introduction …… …… …… …… …… …… …… …… …… …..57 x University of Ghana http://ugspace.ug.edu.gh 4.1 Materials and Equipment… …… …… …… …… …… …… …… …..57 4.2 Method…... …… …… …… …… …… …… …… …… …… …..60 4.2.1 Performance evaluation of Imaging Systems …… …… …… …… …..61 4.2.1.1 Ultrasound Brachytherapy System Quality Control …… …… …..61 4.2.1.1.1 Axial and Lateral Distance Measurement Accuracy…... …..64 4.2.1.1.2 Volume Measurement Accuracy. …… …… …… …..65 4.2.1.1.3 Depth of Penetration….. …… …… …… …… …..66 4.2.1.1.4 Axial and Lateral Resolution….. …… …… …… …..67 4.2.1.2 PET-CT Quality Control… …… …… …… …… …… …..67 4.2.1.2.1 CT Laser Alignment…… …… …… …… …… …..68 4.2.1.2.2 Table (Bed) Position Accuracy… …… …… …… …..69 4.2.1.2.3 CT Display Width …… …… …… …… …… …..70 4.2.1.2.4 CT Image Uniformity…. …… …… …… …… …..72 4.2.1.2.5 Computed Tomographic Dose Index (CTDI)…. …… …..73 4.2.1.2.6 PET-CT Image Uniformity …… …… …… …… …..76 4.2.1.2.7 PET-CT Image Registration Accuracy… …… …… …..78 4.2.1.1.8 PET-CT Image Quality... …… …… …… …… …..80 4.2.1.2.9 PET-CT System Resolution …… …… …… …… …..82 4.2.2 Acquisition of Prostatic Images… …… …… …… …… …… …..84 4.2.3 Development of MATLAB Algorithms…. …… …… …… …… …..87 4.2.3.1 Contrast Enhancement Algorithm.. …… …… …… …… …..87 4.2.3.2 Image Fusion Algorithm… …… …… …… …… …… …..88 4.2.4 Estimation of Image Registration Error…. …… …… …… …… …..88 CHAPTER FIVE: RESULTS AND DISCUSSION 5.0 Introduction …… …… …… …… …… …… …… …… …… …..89 5.1 Performance Evaluation Results…. …… …… …… …… …… …… …..89 5.1.1 Ultrasound Brachytherapy System …… …… …… …… …… …..89 5.1.2 PET-CT System….. …… …… …… …… …… …… …… …..95 5.2 Acquired Prostatic Images. …… …… …… …… …… …… …… …105 xi University of Ghana http://ugspace.ug.edu.gh 5.3 MATLAB Algorithms…… …… …… …… …… …… …… …… …107 5.3.1 Contrast Enhancement Algorithm.. …… …… …… …… …… …107 5.3.2 Image Fusion Algorithm… …… …… …… …… …… …… …108 5.4 Error Assessment… …… …… …… …… …… …… …… …… …111 CHAPTER SIX: CONCLUSION AND RECOMMENDATIONS 6.0 Introduction …… …… …… …… …… …… …… …… …… …114 6.1 Conclusion.. …… …… …… …… …… …… …… …… …… …114 6.2 Recommendations.. …… …… …… …… …… …… …… …… …116 REFERENCES… …… …… …… …… …… …… …… …… …… …118 APPENDIX A: MATLAB ALGORITHMS…. …… …… …… …… …… …130 Appendix A1: MATLAB Contrast Enhancement Algorithm …… …… …… …… …130 Appendix A2: MATLAB Image Fusion Algorithm …… …… …… …… …… …135 APPENDIX B: PUBLISHED AND ACCEPTED ARTICLES …… …… …… …141 Journal Articles 1. F. Hasford, B. Van Wyk, T. Mabhengu, M.D.T. Vangu, A.K. Kyere, J.H. Amuasi. Determination of dose delivery accuracy in CT examinations. Journal of Radiation Research an Applied Sciences (2015), http://dx.doi.org/10.1016/j.jrras.2015.05.006. 2. F. Hasford, B. Van Wyk, T. Mabhengu, M.D.T. Vangu, A.K. Kyere, J.H. Amuasi. Effect of Radionuclide Activity Concentration on PET-CT Image Uniformity. World Journal of Nuclear Medicine. WJNM_44_15. Accepted for publication. 3. F. Hasford, J.H. Amuasi, A.K. Kyere, M.D.T. Vangu, Quantitative Assessment of Radionuclide Uptake and Positron Emission Tomography-Computed Tomography Image Contrast. World Journal of Nuclear Medicine. WJNM_78_15. Accepted for publication Conference Presentations 1. F. Hasford, J.H. Amuasi, A.K. Kyere, M.D.T. Vangu. Ultrasound and PET-CT Image Fusion for Prostate Brachytherapy Image Guidance. International Conference on Clinical PET-CT and Molecular Imaging (IPET2015), 5 – 9 October 2015, Vienna – Austria. Poster presentation xii University of Ghana http://ugspace.ug.edu.gh 2. F. Hasford, B. Van Wyk, T. Mabhengu, M.D.T. Vangu, A.K. Kyere, J.H. Amuasi. Quantitative Assessment of PET-CT Image Uniformity. International Conference on Clinical PET-CT and Molecular Imaging (IPET2015), 5 – 9 October 2015, Vienna – Austria. Poster presentation 3. F. Hasford, J.H. Amuasi, A.K. Kyere, M.D.T. Vangu. Ultrasound and PET-CT Image Fusion for Prostate Brachytherapy Image Guidance. Maiden University of Ghana Doctoral Research Conference, 5 – 6 November 2015, Accra – Ghana. (Winner: Best Poster presentation) xiii University of Ghana http://ugspace.ug.edu.gh LIST OF FIGURES Figure 2.1: The prostate (circled) and other surrounding glands …… …… …… ..….9 Figure 2.2: (a) Normal prostate and (b) cancerous prostate…. …… …… …… ….10 Figure 2.3: (a) Ultrasound generated showing regions of compression and rarefraction; (b) Ultrasound transducer assembly …… …… …… …… …… .…17 Figure 2.4: Acquisition of x-ray CT…… …… …… …… …… …… …… ….19 Figure 2.5: Emission and detection of photons from annihilation point …… …… ….22 Figure 2.6: Schematic diagram of a PET-CT scanner design by Siemens Medical Solutions… …… …… …… …… …… …… …… …… ….23 Figure 2.7: Prostate brachytherapy procedure.. …… …… …… …… …… .…27 Figure 2.8: Decay scheme of I-125…... …… …… …… …… …… …… .…30 Figure 2.9: Decay scheme of Pd-103… …… …… …… …… …… …… ….32 Figure 2.10: Decay scheme of Ir-192…… …… …… …… …… …… …… ….35 Figure 3.1(a): Generic image fusion processing chain…. …… …… …… …… .…41 Figure 3.1(b): Shows the pixel m,n with the gray-levels Gk , k 1,2,3,4….. …… .…49 Figure 3.2: Showing four pixel locations Qk  uk ,vk  , k 0,1,2,3. …… …… ….51 Figure 4.1: Ultrasound system with transrectal probe. …… …… …… …… .…57 Figure 4.2: PET-CT system….. …… …… …… …… …… …… …… .…58 Figure 4.3: Multi-modality prostate phantom... …… …… …… …… …… ….59 Figure 4.4(a): Prostate brachytherapy QA phantom …… …… …… …… …… .…61 Figure 4.4(b): Schematic diagram of phantom…. …… …… …… …… …… .…62 Figure 4.5: Setup for ultrasound system QC tests …… …… …… …… …… .…63 Figure 4.6: Ultrasound image showing N-shaped targets for axial and lateral distance measurements…… …… …… …… …… …… …… .…64 Figure 4.7: Cross-sectional image of QA phantom’s volume insert for volume estimation..65 Figure 4.8: Set-Up for CT Laser Alignment Test …… …… …… …… …… .…67 Figure 4.9(a): Set-up of quality assurance phantom for display width test …… …… ….70 Figure 4.9(b): Acquired image of orthogonal section for display width test …… …… ….70 Figure 4.10: Cross-sectional slice with regions of interest…… …… …… …… .…71 Figure 4.11: Setup for CTDI determination…… …… …… …… …… …… .…73 xiv University of Ghana http://ugspace.ug.edu.gh Figure 4.12(a): Cylindrical phantom under scanning for PET-CT uniformity test …… .…76 Figure 4.12(b): Reconstructed slice of cylindrical phantom with 1010 mm square ROIs ….76 Figure 4.13(a): Image quality phantom under scanning… …… …… …… …… .…78 Figure 4.13(b): Cross-sectional reconstructed PET-CT image of image quality phantom ….79 Figure 4.14(a): Jaszczak phantom under scanning.. …… …… …… …… …… .…81 Figure 4.14(b): Cross-sectional reconstructed PET-CT image of Jaszczak phantom…… ….82 Figure 4.14(c): Profile plot of counts vs. distance for estimation of PETC-CT system Resolution… …… …… …… …… …… …… …… …… .…83 Figure 4.15(a): Multi-modality prostate phantom under scanning on ultrasound brachytherapy system…… …… …… …… …… …… …… .…84 Figure 4.15(b): Multi-modality prostate phantom with needle inserts under scanning on ultrasound system…… …… …… …… …… …… …… .…84 Figure 4.16: Multi-modality prostate phantom under scanning on PET-CT system…. ….85 Figure 5.1: Ultrasound image of prostate phantom… …… …… …… …… ...105 Figure 5.2: PET-CT image of prostate phantom …… …… …… …… …… ...106 Figure 5.3: Graphic user interface for contrast enhancement algorithm …… …… ...107 Figure 5.4: Graphic user interface for image fusion algorithm …… …… …… ...108 Figure 5.5(a): Fused US-PET-CT image in grayscale display….. …… …… …… ...109 Figure 5.5(b): Fused US-PET-CT image in RGB display.. …… …… …… …… ...110 Figure 5.6: Coordinates for image registration error assessment …… …… …… ...111 xv University of Ghana http://ugspace.ug.edu.gh LIST OF TABLES Table 2.1: AJCC’s TNM staging of cancer…… …… …… …… …… …… .…15 Table 2.2: Classification of brachytherapy treatments based on dose rate…. …… .…27 Table 2.3: American Brachytherapy Society’s (ABS’) recommendations for LDR brachytherapy. …… …… …… …… …… …… …… .…28 Table 2.4: Indications and contraindications for LDR brachytherapy monotherapy according to ABS and GEC-ESTRO recommendations… …… …… ….29 Table 2.5: Properties of iodine (I-125) LDR brachytherapy seed…… …… …… ….31 Table 2.6: Properties of palladium (Pd-103) LDR brachytherapy seed …… …… .…33 Table 2.7: Properties of iridium (Ir-192) HDR brachytherapy seed.... …… …… .…36 Table 2.8: Properties of cobalt (Co-60) HDR brachytherapy seed…. …… …… .…37 Table 2.9: Comparison of LDR and HDR implants… …… …… …… …… .…39 Table 3.1: Method for Calculating Optimum One-Dimensional Bandwidth  …… .…48 Table 5.1: Axial distance measurement accuracy results…… …… …… …… ….89 Table 5.2: Lateral distance measurement accuracy results…. …… …… …… .…91 Table 5.3: Volume measurement accuracy results…. …… …… …… …… ….92 Table 5.4: Depth of penetration results …… …… …… …… …… …… .…93 Table 5.5: Axial and lateral resolution test results….. …… …… …… …… ….93 Table 5.6: CT laser alignment test results…… …… …… …… …… …… .…95 Table 5.7: Table (bed) position accuracy test results.. …… …… …… …… .…95 Table 5.8: CT Display width test results …… …… …… …… …… …… .…96 Table 5.9: CT uniformity and image noise test results …… …… …… …… ….97 Table 5.10(a): CTDI test results for head phantom at 120 kVp and 150 mAs…. …… .…97 Table 5.10(b): CTDI test results for body phantom at 120 kVp and 100 mAs…. …… .…98 Table 5.11(a): PET-CT image uniformity test results…… …… …… …… …… .…99 Table 5.11(b): Coefficient of variation for non-uniformity …… …… …… …… ...100 Table 5.12: PET-CT image registration accuracy …… …… …… …… …… ...101 Table 5.13(a): Percentage contrast test results…… …… …… …… …… …… ...102 Table 5.13(b): Background variability test results.. …… …… …… …… …… ...103 Table 5.14: PET-CT system resolution test results…… …… …… …… …… ...104 Table 5.14: Image registration error assessment results …… …… …… …… ...112 xvi University of Ghana http://ugspace.ug.edu.gh LIST OF ABBREVIATIONS AND SYMBOLS 2D Two-dimensional 3D Three-dimensional AAPM American Association of Physicists in Medicine ABS American Brachytherapy Society AJCC American Joint Committee on Cancer ASTRO American Society of Therapeutic Radiology and Oncology C-11 Carbon-11 Co-60 Cobalt-60 CRT Conformal radiation therapy CT Computed tomography CTDI Computed tomography dose index CV Coefficient of variation DHT Dihydrotestosterone DNA Deoxyribonucleic acid DRE Digital rectal examination EBRT External beam radiation therapy F-18 FDG Fluorine-18 fluoro-deoxy-glucose FCH Fluorocholine FOV Field-of-view FWHM Full width at half maximum GEC-ESTRO Groupe Europeen de Curietherapie – European Society for Therapeutic Radiology and Oncology H Entropy HDR High dose rate HHS Human Health Series HU Hounsfield unit HVL Half value layer I-125 Iodine-125 IAEA International Atomic Energy Agency xvii University of Ghana http://ugspace.ug.edu.gh IARC International Agency for Research on Cancers IEC International Electrotechnical Commission Ir-192 Iridium-192 keV Kilo electron-volt kVp Kilo volts peak LDR Low dose rate LOR Line of response LSO Lutetium oxyorthosilicate MBq Mega Becquerel MDR Medium dose rate MeV Mega electron-volt mGy milli Gray MI Mutual information MRI Magnetic resonance imaging NEMA National Electrical Manufacturers Association NU Non-uniformity OaR Organ at risk Pd-103 Palladium-103 PET Positron emission tomography PET-CT Positron emission tomography – computed tomography PPV Positive predictive value PSA Prostate-specific antigen Pt-192 Platinum-192 PVI Partial volume interpolation QA Quality assurance QC Quality control Rh-103 Rhodium-103 RMSE Root mean square error ROI Region of interest SNR Signal to noise ratio SPECT Single photon emission computed tomography xviii University of Ghana http://ugspace.ug.edu.gh Te-125 Tellurium-125 TG Task Group TNM Tumor, nodes and metastases TRUS Transrectal ultrasound US Ultrasound USPSTF United States Preventive Services Task Force xix University of Ghana http://ugspace.ug.edu.gh CHAPTER ONE BACKGROUND 1.0 INTRODUCTION Prostate cancer is the abnormal growth and variable cellular differentiation of prostate tissues (ACS, 2012). One treatment option is to expose the cancerous cells to ionizing radiation to cause a rapid breakdown in the cell’s deoxyribonucleic acid (DNA) structures resulting in cell death. This treatment option is broken down into external beam radiation therapy and brachytherapy (Demanes D.J. et al, 2005). Brachytherapy is the clinical use of small encapsulated radioactive sources at a short distance from the target volume for irradiation of malignant tumors (Nath R. et al, 1997). Interstitial prostate brachytherapy is the permanent implantation of radioactive seeds directly into the prostate. Brachytherapy delivers dose locally to the prostate, but dose gradients are much higher than that for external beam treatment (Nath R. et al, 1997). Over the last decade, technical innovations, three-dimensional (3D) image-based planning, template guidance, computerized dosimetry analysis, and improved quality assurance (QA) practice have converged in synergy in modern prostate brachytherapy which promise to lead to increased tumor control and decreased toxicity (Yu Y. et al, 1999). Prostatic imaging modalities such as ultrasound (US), x-ray, computed tomography (CT) and radionuclide imaging have been employed over the years in the quest to improve cancer patient healthcare (Halpern E.J., 2006; Terris M.K. et al, 1992; Langen K.M. et al, 2003; Henderson E. et al, 2003; Korporaal J.G. et al, 2010). Advancement in diagnostic and therapeutic procedures 1 University of Ghana http://ugspace.ug.edu.gh have led to the evolvement of powerful and sophisticated imaging modalities such as magnetic resonance imaging (MRI), single photon emission computed tomography (SPECT), positron emission tomography (PET), and more recently PET-CT (Seo Y. et al, 2006; Ivan J. et al, 2010; Kaplan I. et al, 2002; Kurhanewicz J. et al, 2002; Beyersdorff D. et al, 2005; Toshihiko H. et al, 1998; Oyama N. et al, 2002; Schmid D.T. et al, 2005). These imaging modalities have contributed to 3D image-based planning leading to a much more accurate prostate dosimetry in pre-plan, real time and post-operative protocols. Fusion of medical images between different cross-sectional modalities is widely used, mostly where functional images such as SPECT or PET are fused with anatomical data from MRI, CT or US. In the area of off-line image fusion, a number of integrations have been performed between several imaging modalities that have in one way or the other improved medical diagnosis and therapy. Transrectal ultrasound (TRUS) and CT or MRI have been used for planning of brachytherapy and also for biopsy-guidance, as focal lesions and brachytherapy seeds are better visualized on CT or MRI images (Lonescu G. et al, 1999; Rifkin M.D. et al, 1990; Reynier C. et al, 2004). Holupka et al (1996) described a real-time US imaging and targeting system for the treatment of prostate cancer. In the study, patients were scanned on both CT and US system with TRUS probe in place (in situ) in order to assess the 3D shape of the prostate. The TRUS and CT images were fused by using a two-point matching technique. The accuracy of the method was found to be 2 mm. In a recent study by Steggerda et al (2005), prostate seed implants were evaluated by co- registering CT and TRUS images with probe in situ. The registration error was estimated to be less than 1 mm and the co-registered images offered both optimal prostate and seed visibility. Another research group, Fuller et al (2005), tested a system for dynamic dosimetry feedback 2 University of Ghana http://ugspace.ug.edu.gh based on CT-US fusion. The group reported a median discrepancy between the images of 0 – 4 mm and the method improved the position and dose loading of brachytherapy needles. A TRUS- MRI image fusion has been done by Reynier et al (2004), who made the co-registration based on a stepper that gave a constant inter-slice distance of the TRUS images. The volume estimates of the prostate by TRUS compared with MRI showed that TRUS underestimated prostate volume leading to an overestimated brachytherapy dose. PET-CT imaging modality is now widely used because the integration of PET and CT scanners to provide co-registered images combine molecular quantifiable images obtained by PET with high spatial resolution and anatomical detail of CT (Ewertsen C. et al, 2010; Cimitan M. et al, 2006; Gaa J. et al, 2004; Wachter S. et al, 2006). PET-CT has been proven today to be one of the most sensitive and specific examination for tumour staging, therapy planning and prognosis through the complementary nature of the two systems in many tumours (von Schulthess G.K., 2004). Attenuation correction of PET images using the CT data enables PET-CT to process images faster than PET alone, thereby improving imaging efficiency and patient throughput (Landis K.G., 2005). The value of this technique has been proven in tumour imaging and new applications are emerging rapidly. PET-CT is today the fastest-growing imaging modality worldwide. 1.1 Thesis Motivation Globocan 2002 data by the International Agency for Research on Cancers (IARC) estimated that out of 165,000 prostate cancer incidences annually, approximately 91,000 were likely to result in mortality (Parkin D.M. et al, 2005). The data showed prostate cancer as the leading cause of 3 University of Ghana http://ugspace.ug.edu.gh cancer mortality in Ghanaian males followed by liver cancer. A 10-year study (1991 – 2000) performed at the Korle-Bu Teaching Hospital in Ghana, and published in 2006 by Wiredu et al (2006) found prostate cancer (with 17.35% mortality rate) to be the second most deadly cancer in men only after liver cancer. These two reports concur that prostate cancer is one of the topmost causes of cancer deaths in Ghanaian males. In 2007, another survey at the Korle-Bu Teaching Hospital revealed that Ghana had exceeded global prostate limits. By the survey, out of every 100,000 men, the country recorded 200 prostate cancer incidences as against 170 incidences worldwide (Daily Graphic, 2007). A more recent publication by IARC as captured in Globocan 2008 report shows that out of 1,200 prostate cancer incidences annually, 962 (80%) are likely to result in mortality in Ghana (IARC, 2008). These revelations present a worrying feature and hence the need to advance in prostate diagnosis and therapy to meet the challenge. Treatment of the prostate disease or cancer like many other organs in the human body requires highly skilled personnel with carefully optimized therapeutic protocols. Treatment planning for a particular therapeutic protocol that is adopted for the treatment of prostate cancer depends on the imaging modality employed. Hence in order to achieve optimal results in prostate cancer therapy, particular attention has to be paid to the imaging modality used and its accuracy. For instance, because the prostate gland is situated close to the rectum and the bladder, which are all critical organs, any minute overdose of radiation is enough to offset clinical complication. In the quest to optimize dose delivery to prostate, a number of imaging modalities have been employed over time to obtain images used for treatment planning (Langen K.M. et al, 2003; Henderson E. et al, 2003; Seo Y. et al, 2006; Beyersdorff D. et al, 2005; Toshihiko H. et al, 1998; Oyama N. et al, 2002). Fusion of images of different modalities have contributed to improved treatment outcomes in prostate brachytherapy, however, the accuracy of the image co-registration has been 4 University of Ghana http://ugspace.ug.edu.gh variable over time, depending on the protocol used by individual researchers. The ability to improve the accuracy of a US-PET-CT image co-registration to levels around 1 mm would likely be of great value for real time and post-implant dose distribution evaluations. Presently in Ghana, prostate brachytherapy is performed using ultrasound alone as the means of image guidance during seed implantation. This approach is laudable and easily applicable. However, for pre-implant assessment of prostate volume and post-implant evaluation of dose distribution, improved means of image fusion is necessary and required for accurate treatment outcomes. 1.2 Thesis Objectives Multiple integrated imaging modalities have shown significant improvement over single imaging modalities in patient studies, diagnosis and therapy. The primary objective of this study is to improve on diagnosis and therapy of prostate cancers by fusing ultrasound and PET-CT images through the use of MATLAB computer software. To effectively achieve this goal, a number of tasks would be addressed in this study, which include:  Performance evaluation of ultrasound and PET-CT systems.  Enhancement of the quality (contrast) of acquired prostatic images by writing suitable MATLAB algorithms.  Fusion of enhanced ultrasound and PET-CT images with MATLAB algorithms.  Assessment of image fusion accuracy. 5 University of Ghana http://ugspace.ug.edu.gh 1.3 Benefits versus Risks Since its introduction in the mid 1980s, prostate brachytherapy has become a well-established treatment option for patients with early localized disease. In the United States alone, over 50,000 eligible prostate cancer patients a year are treated using this method (Prostate Brachytherapy Advisory Group, 2013). In 2012, Ghana recorded 62 prostate brachytherapy cases, which were undertaken at the Korle-Bu Teaching Hospital (IAEA RAF 6044/6045 Ghana Report, 2012). Prostate brachytherapy is a very effective treatment modality for early, localized prostate cancer, with patients rapidly returning to normal activities (Langley S.E. et al, 2002). Although patients sometimes experience problems like urinary incontinence or difficulty with urination in the first 6 months after their implant, these usually settle down and lasting problems are rare, only occurring in about 1–2% of patients (Crook J. et al, 2008). Erectile dysfunction or impotence is another side-effect associated with all treatments for prostate cancer. Other treatments for prostate cancer cause problems with erectile dysfunction in 30–60% of men, but these problems are much less common after LDR brachytherapy, and only occur in about 10–30% of men under the age of 60 years, who were potent before treatment (Prostate Brachytherapy Advisory Group, 2013). In a recent study looking at patients’ quality of life, LDR brachytherapy compared favourably with other treatment options (Buron C. et al, 2007). Although these side effects are not desirable, patients give it less consideration when comparing it to the importance of living a long and cancer-free life. Also, most prostate cancer patients are generally older men past the age of active procreation. 6 University of Ghana http://ugspace.ug.edu.gh 1.4 Thesis Overview This study seeks to improve diagnosis and therapy of prostate cancers through the use of MATLAB image enhancement toolbox. Enhancement and fusion of ultrasound, PET and CT images has been performed by writing rigorous algorithms in MATLAB. Accuracy of the co- registered images has been determined to test the credibility of the MATLAB algorithm used for this study. Chapter one of this thesis gives a general overview of the research topic. It highlights on the motivation and objectives of the study, and on the benefits and risks of prostate brachytherapy. Chapter two reviews literature pertinent to the study. These include a review of literature on prostate cancer, prostate imaging modalities such as ultrasound and PET and CT, and prostate brachytherapy techniques. The techniques reviewed are low dose rate (LDR) and high dose rate (HDR) brachytherapy. Chapter three of the thesis highlights the theory of MATLAB image fusion, which is based on the ‘principle of mutual information’ and ‘rigid body transformation’. Chapter four addresses materials and methods employed in the study. The chapter addresses QC tests performed on the imaging equipment, the experimental set-up and the processes of image acquisition. Two MATLAB algorithms are written in this chapter to (a) enhance the contrast of the acquired images and (b) fuse the required images. Error analysis has been performed on the fused images to assess the level of registration accuracy and integrity of fusion algorithm. Results and discussion are presented in chapter five and the study is concluded in chapter six, with relevant recommendations. 7 University of Ghana http://ugspace.ug.edu.gh CHAPTER TWO LITERATURE REVIEW 2.0 INTRODUCTION Multi-modality imaging plays an increasingly important role in the diagnosis and treatment of a large number of diseases such as prostate cancer, particularly if both functional and anatomical information are acquired and accurately co-registered. Imaging with multi-modality systems could easily help locate tumour within a particular region of the prostate, guide treatment procedures, and also help detect local recurrence earlier than is currently possible. This chapter reviews current developments in the field of prostate imaging and brachytherapy. 2.1 THE PROSTATE GLAND The prostate is a gland found only in males, located in front of the rectum and below the urinary bladder. Figure 2.1 shows the prostate gland and its surrounding critical organs. The size of a prostate varies with age. In younger men, it is about the size of a walnut, but it is much larger in older men. A typical prostate in an adult male measures about 4 cm  2 cm  3 cm and weighs approximately 20 grams (Masters Men’s Clinic, 2013; Behre H.M. et al, 1994). The prostate's function is to produce some of the fluid that protects and nourishes sperm cells in semen, making the semen more liquid. Just behind the prostate are glands called seminal vesicles that produce most of the fluid for semen. The urethra, the tube that carries urine and semen out of the body through the penis, goes through the center of the prostate. 8 University of Ghana http://ugspace.ug.edu.gh Fig. 2.1: The prostate and other surrounding glands (Masters Men’s Clinic, 2013) The prostate, which starts its development before birth, grows rapidly during puberty. The growth is fueled by androgens, which are male hormones in the body. The main androgen called testosterone is produced in the testicles. Testosterone is converted into dihydrotestosterone (DHT) by the enzyme 5-alpha-reductase. Dihydrotestosterone is the main hormone that signals the prostate to grow. The prostate usually stays at about the same size or grows slowly in adults, as long as male hormones are present (ACS, 2012). 2.2 PROSTATE CANCER The human body is made up of trillions of living cells. Normal body cells grow, divide, and die in an orderly fashion. During the early years of a person's life, normal cells divide faster to allow the individual to grow. However, in the adult age, most cells divide only to replace worn-out or dying cells or to repair injuries. Cancer begins when cells in a part of the body start to grow out 9 University of Ghana http://ugspace.ug.edu.gh of control. Cancer cell growth is different from normal cell growth. Instead of dying, cancer cells continue to grow and form new abnormal cells. Cancer cells most often invade other neighbouring tissues, a character that normal cells do not have. Prostate cancer is a form of cancer that develops in the prostate as shown in Figure 2.2(b). Most prostate cancers are slow growing (Lister S., 2009), however, there have sometimes been cases of aggressive prostate cancers (ACS, 2012). The cancer cells may metastasize from the prostate to other parts of the body, particularly the bones and lymph nodes (ACS, 2012; Lister S., 2009). Prostate cancer sometimes causes pain, difficulty in urinating, problems during sexual intercourse, and/or erectile dysfunction. Other symptoms potentially develop during later stages of the disease (Lister S., 2009). Fig. 2.2: (a) Normal prostate and (b) cancerous prostate (Masters Men’s Clinic, 2013) Rates of detection of prostate cancers vary widely across the world, with South and East Asia detecting less frequently than in Europe, and especially the United States (IARC, 2001). Prostate 10 University of Ghana http://ugspace.ug.edu.gh cancer tends to develop in men over the age of 50 years (Siegel R., 2011). Globally, it is the sixth leading cause of cancer-related death in men (Jemal A. et al, 2011). Prostate cancer is most common in the developed world, with increasing rates in the developing world (Baade P.D. et al, 2009). However, many men with prostate cancer never have symptoms, undergo no therapy, and eventually die of other unrelated causes (Siegel R., 2011). Many factors, including genetics and diet, have been implicated in the development of prostate cancer (Jemal A. et al, 2011). The presence of prostate cancer may be indicated by symptoms, physical examination, prostate- specific antigen (PSA) or biopsy. Prostate specific antigen testing increases cancer detection but does not decrease mortality (Djulbegovic M. et al, 2010). The United States Preventive Services Task Force (USPSTF) in 2012 recommended against screening for prostate cancer using the PSA testing, due to the risk of over-diagnosis and over-treatment with most prostate cancer remaining asymptomatic. The USPSTF concludes that the potential benefit of testing for prostate cancer does not outweigh the expected harms (Prostate Cancer Screening, 2013). Several types of cells are found in the prostate, but almost all prostate cancers develop from the gland cells. Gland cells produce the prostate fluid that is added to the semen. The medical term for a cancer that starts in gland cells is adenocarcinoma (College of American Pathologists, 2013). Other types of cancer can also start in the prostate gland, including sarcomas, small cell carcinomas, and transitional cell carcinomas. But these types of prostate cancers are so rare that if a person is diagnosed of prostate cancer it is almost certain to be an adenocarcinoma. Some prostate cancers grow and spread rapidly, but most of them grow slowly (Virtual Medical Centre, 2013). Autopsy studies by some researchers have shown that many older and some younger men who died of other diseases also had prostate cancer that never affected them during their lives (ACS, 2012). In many cases neither they nor their doctors knew they had prostate cancer. 11 University of Ghana http://ugspace.ug.edu.gh Management strategies for prostate cancer are guided by the severity of the disease. Many low- risk tumors can be safely followed with active surveillance. Curative treatment generally involves surgery, various forms of radiation therapy, or less commonly, cryosurgery. Hormonal therapy and chemotherapy are generally reserved for cases of advanced disease, although hormonal therapy may be given with radiation in some cases (Virtual Medical Centre, 2013). Several studies suggest that frequent ejaculation reduces the risk of prostate cancer (Leitzmann M.F. et al, 2004; Giles G.G. et al, 2003; Dimitropoulou P. et al, 2009). The age and underlying health of the man, the extent of metastasis, appearance under the microscope and response of the cancer to initial treatment are important in determining the outcome of the disease. The decision whether or not to treat localized prostate cancer with curative intent is a patient trade-off between the expected beneficial and harmful effects in terms of patient survival and quality of life (Virtual Medical Centre, 2013). As yet, the main cause of prostate cancer, like many other cancers, is not completely understood. However, researchers have identified several risk factors as associated with the development of the disease (ACS, 2012; Patel A.R. et al, 2009). Some risk factors such as smoking, diet and obesity can be changed, while others such as age, race, ethnicity and family history (gene) cannot be changed. It is not conclusive to predict whether an individual may develop prostate cancer in future, using risk factor alone.  Age: The chance of developing prostate cancer rapidly increases after age 50 years. Almost two-thirds of reported prostate cancer cases are found in men above 65 years (ACS, 2012; Patel A.R. et al, 2009). 12 University of Ghana http://ugspace.ug.edu.gh  Race / Ethnicity: Prostate cancer is more often detected in men of African-American descent than in men of other races. Also, the disease occurs less often in men of Asian- American and Hispanic descent compared with non-Hispanic whites (ACS, 2012; Patel A.R. et al, 2009).  Nationality: Prostate cancer is most common among nationals of North America, North- Western Europe, Australia, and Caribbean islands. It is relatively less common in Asia, Africa, Central America, and South America. Some researchers attribute this observation to reasons such as lifestyle differences and the availability of intensive screening exercises in some of the developed countries. Studies have revealed that men of Asian descent living in the United States have a lower risk of prostate cancer than white Americans, but their risk is higher than that of men of similar backgrounds living in Asia (ACS, 2012; Patel A.R. et al, 2009).  Family History: Prostate cancer has been observed to run in certain families, suggesting an inherited or genetic factor in those cases. The risk is higher for a man with an affected brother rather than an affected father. The risk is much higher for a man with several affected relatives, particularly if the relatives were affected at young ages (ACS, 2012; Patel A.R. et al, 2009).  Diet: Certain studies have suggested that men who eat lots of red meat, high-fat dairy products, and fewer fruits and vegetables, appear to have slightly higher chance of getting prostate cancer. Although calcium is known to provide important health benefits, some researchers suggest that men who consume lots of calcium through food supplements 13 University of Ghana http://ugspace.ug.edu.gh may also have a high risk of developing prostate cancer (ACS, 2012; Patel A.R. et al, 2009).  Obesity: A few studies have been able to suggest that obese men have a higher risk of getting an aggressive prostate cancer, however, more studies are needed in this area to confirm the observation (ACS, 2012; Patel A.R. et al, 2009). 2.2.1 Prostate Cancer Staging An important part of evaluating prostate cancer is determining the stage, or how far the cancer has spread. Knowledge of the stage helps to define prognosis and is useful when selecting therapies. The most common staging system is the TNM System, where T, N and M represent tumor, nodes and metastases respectively. The components of this staging system include the size of the tumor, the number of involved lymph nodes, and the presence of any other metastases (BMJ Group, 2009). The most important distinction made by any staging system is whether or not the cancer is confined to the prostate. In the TNM System, clinical T1 and T2 cancers are found only in the prostate, while T3 and T4 cancers have spread elsewhere (Sobin L.H. et al, 2009). Several tests are used to look for evidence of spread. Table 2.1 shows the American Joint Committee on Cancer (AJCC) TNM staging of cancers (National Cancer Institute, 2013). Medical specialty professional organizations recommend against the use of PET, CT, or bone scans when a physician stages early prostate cancer with low risk for metastasis (American Society of Clinical Oncology, 2012; Makarov D.V. et al, 2012). Those three imaging modalities 14 University of Ghana http://ugspace.ug.edu.gh are appropriate in such cases when a CT scan evaluates spread within the pelvis. Bone scan reveals osteoblastic appearance due to increased bone density in the areas of bone metastasis, opposite to what is found in many other cancers that metastasize. Endorectal coil magnetic resonance imaging is sometimes used to closely evaluate the prostatic capsule and the seminal vesicles for staging. Table 2.1: AJCC’s TNM staging of cancer (National Cancer Institute, 2013) T: Size or direct extent of the primary tumor Tx Tumor cannot be evaluated Tis Carcinoma in situ T0 No signs of tumor T1 Tumor ≤ 2 cm in greatest dimension T2 Tumor ≥ 2 cm but ≤ 5 cm in greatest dimension T3 Tumor > 5 cm in greatest dimension T4 Tumor of any size with direct extension to the chest wall and/or to the skin N: Degree of spread to regional lymph nodes Nx Lymph nodes cannot be evaluated N0 No regional lymph node metastases N1 Regional lymph node metastasis present (tumor spread to closest or small number of regional lymph nodes) N2 Tumor spread to an extent between N1 and N3 (N2 is not used at all sites) N3 Tumor spread to more distant or numerous regional lymph nodes (N3 is not used at all sites) M: Presence of distant metastasis M0 No distant metastasis M1 Metastasis to distant organs (beyond regional lymph nodes) Prostate specific antigen (PSA) is currently the main serological marker routinely used in the diagnosis, staging and monitoring of treatment response or failure in prostate cancer (Polascik T.J. et al, 1999). The PSA test remains a better predictor of prostate cancer than digital rectal examination (DRE) or transrectal ultrasound (Christensson A. et al, 1990). It is a prostatic 15 University of Ghana http://ugspace.ug.edu.gh secretory protein whose physiological function is to dissolve the coagulated semen. Its half-life in serum is about 3 days (Christensson A. et al, 1990; Zhang W.M. et al, 1995). In prostate biopsy, a pathologist or urologist observes samples of the prostate tissue under microscope, and reports on the presence or absence of tumour. The grade of tumour tells how much its tissue differs from normal prostate tissue and suggests how fast the tumor is likely to grow. Gleason system (Gleason D.F., 1992; Billis A. et al, 2008) has been used over the years to grade prostate tumours from 2 to 10, where a Gleason score of 10 indicates the most abnormalities. A pathologist assigns a number from 1 to 5 for the most common pattern observed under microscope, then does the same for the second-most-common pattern. The sum of the two numbers is the Gleason Score. 2.3 PROSTATE IMAGING MODALITIES The evolution of medical imaging modalities allows for better detection and staging of prostate cancer, thus directing appropriate treatment and follow-ups. However, the appropriate use of imaging is difficult to define, as many controversial studies regarding each of the modalities and their utilities can be found in literature. There exists a variety of imaging modalities employed in diagnosing prostate tumours and aiding in prostate brachytherapy, but of importance to this study are ultrasound, CT and PET. 2.3.1 Ultrasound Ultrasound, otherwise called ultrasonography, is a technique in which high-frequency sound 16 University of Ghana http://ugspace.ug.edu.gh waves exceeding the range of human hearing are bounced off internal organs, with the echo pattern converted into a two-dimensional image of the structures beneath a transducer. Ultrasound energy is generated by mechanical displacement in compressible medium, which is modeled as an elastic spring (Figure 2.3(a)). Energy propagation is shown as a function of time, resulting in areas of compression and rarefaction with corresponding variations in positive and negative pressure amplitude. The transducer or the ultrasound probe made up of a piezoelectric material, as shown in Figure 2.3(b) converts electrical energy into mechanical energy to produce ultrasound and mechanical energy back into electrical energy for ultrasound detection. Medical ultrasound usually uses frequencies in the range of 2 to 10 MHz, with specialized ultrasound applications up to 50 MHz (Bushberg J.T. et al, 2011). Fig. 2.3: (a) Ultrasound generated showing regions of compression and rarefraction – Left; (b) Ultrasound transducer assembly – right (Bushberg J.T. et al, 2011) Transrectal ultrasound uses a probe that is inserted into the rectum to create an image of the prostate. It is the most commonly used modality for imaging the prostate gland. TRUS enables 17 University of Ghana http://ugspace.ug.edu.gh determination of prostate size and demonstration of the zonal anatomy, and prostate cancer lesions usually appear hypoechoic relative to normal tissue. TRUS has many advantages, including its portability, ease of use, lack of ionizing radiation, low cost, and its capability to perform real-time imaging. Ultimately, however, TRUS is not a high-resolution imaging modality. The inherent speckle pattern of ultrasound limits the detection and definition of the margins of cancers. The capability of TRUS to delineate cancer foci is limited, and its sensitivity and specificity are low: small cancer foci are often not visible at all, and the majority of hypoechoic foci detected by TRUS are not malignant (Hricak H. et al, 2007). The addition of colour Doppler and/or power Doppler has the potential of increasing the rate of tumour visualization by detecting regions of hypervascularity; however, the sensitivity of TRUS is not increased with Doppler because most small tumours are not angiogenic (Cornud F. et al, 2000). Contrast-enhanced TRUS with microbubbles has been found to provide higher sensitivity for the detection of cancer foci than standard TRUS, and has shown to increase the detection rate of clinically significant prostate cancer in several studies (Yang J.C. et al, 2008; Tang J. et al, 2007; Halpern E.J. et al, 2005; Taymoorian K. et al, 2007). 2.3.2 Computed Tomography Computed tomography is a technique that uses multiple two-dimensional (2D) x-ray images, taken 3600 around a single axis of rotation, to generate a three-dimensional (3D) image of the inside of an object, mostly human beings as shown in Figure 2.4. Physically, x-ray CT is the measurement of the object’s x-ray attenuation along straight lines. For incident photons ( I 0 ) and an object layer of thickness ( t ) and attenuation coefficient (  ), the number of photons reaching 18 University of Ghana http://ugspace.ug.edu.gh the detector ( I ) is given by the exponential attenuation law as I  I t0e (2.1) For non-homogeneous objects the attenuation coefficient is a function of x , y , and z . The projection value P corresponds to the line integral along line L of the object’s linear attenuation coefficient distribution x, y, z . Hence I L PL  In  dLx, y, z (2.2) I0 L The process of computing the CT image f x, y, z , which is an accurate approximation to the attenuation coefficient distribution x, y, z , from the set of measured projection values PL is called CT image reconstruction. Fig. 2.4: Acquisition of x-ray CT 19 University of Ghana http://ugspace.ug.edu.gh The image values f x, y, z are converted into CT values prior to storage by applying the linear function (2.3). f   CT  water 1000 HU (2.3) water CT modality is widely used in both patient diagnosis and follow-up of nearly all malignancies, but it has a slight limited role in the imaging of prostate cancer owing to its poor soft-tissue contrast resolution, which does not allow precise distinction of the internal or external anatomy of the prostate (Hricak H. et al, 2007). The major role of CT in patients with prostate cancer is for the detection of bony involvement and in nodal staging; however, CT only detects the enlargement of involved nodes, which is a late finding in patients with prostate cancer (Narayana V. et al, 1997). In prostate brachytherapy, CT has proven to provide good visualization of implanted seeds comparative to other imaging modalities (Narayana V. et al, 1997). At present, CT has very little role in the diagnosis of prostate cancer or in staging of known cancers in patients with a low clinical suspicion of metastatic disease (Hricak H. et al, 2007). CT detects nodal metastases based on size, and in general CT has a very low diagnostic yield in low- risk patients due to a low incidence of large nodal metastases in these patients. The primary role of CT in prostate cancer is in staging for patients with suspected metastatic disease, for which it has variable sensitivity and specificity (Oyen R.H. et al, 1994; Walsh J.W. et al, 1980). O’Dowd et al (1980) recommend that CT be used for high-risk patients with PSA > 20 ng/mL, Gleason score > 7, or at least clinical stage T3 disease. For the purpose of detecting pelvic lymph node metastases, the sensitivity has been reported from 25% to 85% (Rørvik J. et al, 1998), but is generally approximately 36% (Wolf J.S. et al, 20 University of Ghana http://ugspace.ug.edu.gh 1995), which is not sufficiently accurate to justify widespread use except for selected high-risk patients. Similarly, for diagnosing bone metastases, CT is inferior to other modalities such as PET or SPECT (Hricak H. et al, 2007). 2.3.3 Positron Emission Tomography Positron emission tomography is a form of imaging modality that uses positron emitting radionuclides; the positrons annihilate electrons, which results in the production of two simultaneous and oppositely directed gamma rays. A positron emitting radioactive tracer is introduced into the body using a biologically active molecule and tomographic reconstruction is applied to yield a 3D dataset. While many radionuclides decay via positron emission, only a few have been used much for PET imaging. The most commonly used are: Nuclide Half-life Carbon (C-11) 20.3 min Nitrogen (N-13) 10 min Oxygen (O-15) 124 s Fluorine (F-18) 110 min Rubidium (Rb-82) 75 s After a positron is emitted, it travels a short distance (~1 mm) in tissue, losing energy by exciting and ionizing nearby atoms. Once it has lost almost all its kinetic energy, it annihilates with a nearby electron. The product of this annihilation is a pair of photons. Conservation of energy and momentum dictates that the two photons depart in opposite directions, each with energy 511 keV. The emission of two photons from positron decay is recorded with detectors in opposite directions in a process called coincidence detection, as shown in Figure 2.5. The detectors 21 University of Ghana http://ugspace.ug.edu.gh consist of a scintillator, which converts energy from the 511 keV photons into light photons, and a photomultiplier converts the light into an electronic pulse. The creation of two electronic pulses at the same time (“in coincidence”) is a signal that there has been an annihilation somewhere in the column or line of response (LOR) connecting the associated detectors. The number of coincidence counts obtained on a particular LOR indicates the amount of radioactivity present along that line during the scan. Positron emission tomography has emerged as a very important imaging tool in assessing mostly cancer diseases. Positron emission tomography images are obtained in 3D, and the intensity of the signal is proportional to the amount of tracer; therefore, the technique is potentially quantitative. While routine PET has limited spatial resolution (approximately 4–6 mm), its resolution is better than that of single photon emission computed tomography (SPECT) (Suri J.S. et al, 2005). Positron emission tomography has additional characteristics of imaging specific physiological processes, such as the rate of glucose or fatty acid metabolism, which cannot be achieved with other imaging techniques. Fig. 2.5: Emission and detection of photons from annihilation point (Suri J.S. et al, 2005) 22 University of Ghana http://ugspace.ug.edu.gh 2.3.4 Integrated PET-CT System Integrated PET-CT scanner (as schematically shown in Figure 2.6) is developed with the objective of integrating CT and PET within the same device, to use the CT images for the attenuation and scatter correction of the PET emission data, and to explore the use of anatomic images to define tissue boundaries for PET reconstruction. Thus, the goal was to construct a device with both clinical CT and clinical PET capability so that a full anatomic and functional scan is acquired in a single session, obviating the need for patients to undergo an additional clinical CT scan after PET imaging. Fig. 2.6: Schematic diagram of a PET-CT scanner design by Siemens Medical Solutions (Suri J.S. et al, 2005) 23 University of Ghana http://ugspace.ug.edu.gh 2.3.5 Radionuclides in Prostate PET-CT Imaging Several radiotracers, mainly carbon-11 and fluorine-18 based agents, have been used to study the prostate and the results documented in literature (Liu I.J. et al, 2001; Hofer C. et al, 1999; Shreve P.D. et al, 1996; Yoshimoto M. et al, 2001; Kotzerke J. et al, 2002; Fricke E. et al, 2003). These include fluorine-18 fluoro-deoxy-glucose (F-18 FDG), carbon-11 methionine, F-18 fluorocholine, carbon-11 choline, and carbon-11 acetate, as tracers for detecting prostate cancer. The most common radiotracer for PET imaging, F-18 FDG, is not effective in the diagnosis of localized prostate cancer. The prostate tumours are not glucose dependent, and the tracer does not accumulate in the prostate (Liu I.J. et al, 2001). F-18 FDG does, however, become more useful in patients with androgen-independent tumours such as breast and lungs (Liu I.J. et al, 2001). Some localized prostate cancers are inexplicably highly glucose dependent and will be positive on F-18 FDG PET scans obtained for reasons other than investigation for prostate cancer, although the finding is the exception rather than the norm (Shreve P.D. et al, 1996). In several studies, remarkable similarities in the F-18 FDG uptake in prostate tumours and benign prostate hyperplasia (BPH) have been demonstrated. Additionally, intense bladder activity is usually present, which obscures the prostate and interferes with the identification of pelvic lymph node metastases (Liu I.J. et al, 2001; Hofer C. et al, 1999; Shreve P.D. et al, 1996). For such reasons, F-18 FDG is not the ideal PET tracer for the initial staging and follow-up of prostate cancer. Acetate is a tracer that is continuously being investigated. In four clinical trials, C-11 acetate PET was shown to be useful in the diagnosis, staging and restaging of prostate cancer (Oyama N. et al, 2002; Yoshimoto M. et al, 2001; Kotzerke J. et al, 2002; Fricke E. et al, 2003). Although these few preliminary studies reveal promising results for C-11 acetate PET (Scheidler J. et al, 24 University of Ghana http://ugspace.ug.edu.gh 1999), prospective clinical trials are needed to properly evaluate its role in localized disease. Elevated levels of choline have been detected in prostate cancer cells, and hence several choline analogs have been labeled with C-11 or F-18 for PET imaging. C-11 choline has limited urinary excretion, and is thus well suited for prostate imaging. C-11 choline PET has been shown to be useful in the diagnosis and staging of prostate cancer (Farsad M. et al, 2005; Schiavina R. et al, 2008; Scher B. et al, 2007). Schiavina et al. (2008) found that C-11 choline PET performed better than clinical nomograms for predicting nodal metastases, with a sensitivity and specificity of 60% and 98% respectively. The most promising results with PET-CT are in the field of detecting recurrent disease after primary therapy for prostate cancer (Bouchelouche K. et al, 2009). Rinnab et al (2008) initially found that C-11 choline was useful for targeted salvage lymph node dissection after treatment for prostate cancer (Rinnab L. et al, 2008), and then more recently that C-11 choline has a high sensitivity of 93% and positive predictive value of (PPV) of 80% for detecting local recurrence or distant metastases (Rinnab L. et al, 2009). This more recent study confirmed the results of several previous studies (Krause B.J. et al, 2008; Reske S.N. et al, 2008) that PET-CT using C-11 choline may be a useful approach for detecting and localizing recurrent disease, and also showed that integrated PET-CT systems may be particularly helpful at low PSA levels. Li et al (2008) in their study have proposed that using a ratio of uptake values between prostate tissue and muscle as a primary measurement, C-11 choline PET with sensitivity and specificity values of 90% and 86% respectively, may be feasible to differentiate benign from malignant prostate lesions. With a relatively short half-life of 20 minutes, all C-11 labeled compounds are limited in their use. Longer-lived radioisotopes are now being used in the development of novel tracers, such as F-18 fluorocholine (F-18 FCH), which has a half-life of 110 min and rapidly clears from the 25 University of Ghana http://ugspace.ug.edu.gh blood stream into the urine. Several researchers (Kwee S.A. et al, 2008; Cimitan M. et al, 2006; Schuster D.M. et al, 2007; Larson S.M. et al, 2004; Dehdashti F. et al, 2005; Nunez R. et al, 2002; Richter J.A. et al, 2010) have in recent times used other forms of C-11 and F-18 analog pharmaceuticals to assess diagnostic abilities of the prostate and have obtained promising results, giving additional hope for the future in terms of prostate PET-CT imaging. 2.4 PROSTATE BRACHYTHERAPY Brachytherapy is a term derived from the Greek word brachys, which means brief or short. It refers to cancer treatment with ionizing radiation delivered by radioactive material placed a short distance from, or within, the tumor. In prostate cancer, brachytherapy involves the use of imaging modalities such as ultrasound, CT and MRI, and a template-guided insertion of radioactive seeds into the gland. Brachytherapy as a treatment modality for prostate cancer has been used with varying degrees of enthusiasm for many years. Early techniques in the 1970’s used direct open implantation with live source permanent implants, typically gold grains or iodine seeds (Khan K. et al, 1983). Unfortunately, long-term follow-up revealed less than satisfactory results in terms of cancer control. The poor results were attributed to major 2 problems. Firstly, the technical inability to accurately implant the sources, and secondly, the relative paucity of objective dosimetric criteria by which to analyze the radiation dose in that era (Cheuck L, 2015). Interest in brachytherapy waned at that time and the attention shifted to external beam radiation therapy (EBRT) and nerve-sparing radical prostatectomy. 26 University of Ghana http://ugspace.ug.edu.gh In about two decades now developments in both imaging techniques and brachytherapy equipment have led to prostate brachytherapy treatment technique becoming an important integral part of the management of localized prostate cancer. The technique is supported by improved dosimetry and offers the potential advantage of delivering a higher radiation dose to the prostate than would be possible with EBRT (Cheuck L, 2015). The latter consideration is particularly important in view of the high rate of positive prostate biopsy findings following conventional EBRT. Along with treatment modalities such as radical prostatectomy, cryotherapy, and EBRT, interstitial brachytherapy is a potentially curative treatment for localized prostate cancer. For appropriately selected patients, brachytherapy appears to offer cancer control comparable to that achieved with the other techniques (Cheuck L, 2015). Prostate brachytherapy, as shown in Figure 2.7, comes in two forms; low-dose rate (LDR), in which seed sources are implanted permanently in the prostate, and high-dose-rate (HDR), in which seed sources are implanted temporarily in the prostate capsule. Classification of brachytherapy treatments with respect to dose rate is shown in Table 2.2. Permanent LDR and temporary HDR brachytherapy are competitive techniques for clinically localized prostate radiotherapy. Comparative analysis of the two treatment options is useful in understanding some of their intrinsic differences, several of which could be exploited to improve outcomes of treatment. 27 University of Ghana http://ugspace.ug.edu.gh Fig. 2.7: Prostate brachytherapy procedure (Masters Men’s Clinic, 2013) Table 2.2: Classification of brachytherapy treatments based on dose rate Technique Dose rate at the dose specification point(s) Low dose rate (LDR) Between 0.4 and 2 Gy/h Medium dose rate (MDR) Between 2 and 12 Gy/h High dose rate (HDR) Greater than 12 Gy/h Medium dose rate is not in common use. In those few cases when it has been used, the treatment results have been rather poor compared to LDR or HDR treatments (Cheuck L, 2015). 2.4.1 Low-Dose-Rate Brachytherapy Low dose rate brachytherapy is one of the radiation methods that has been known for almost 30 years in treatment of localized prostate cancer. The main idea of this method is to implant small 28 University of Ghana http://ugspace.ug.edu.gh radioactive seeds as a source of radiation, directly into the prostate gland. LDR brachytherapy is applied as a monotherapy, and also used along with EBRT as a boost. It is used as a sole radical treatment modality, however, not as a palliative treatment. The application of permanent seeds implants is a curative treatment alternative in patients with organ-confined cancer, without extracapsular extension of the tumour (Nag S. et al, 1999; Porter A.T. et al, 1995; Garbaulet A. et al, 2002; Thompson I. et al, 2007). The American Brachytherapy Society’s (ABS) recommendations for LDR brachytherapy is given in Table 2.3. Table 2.3: ABS’ recommendations for LDR brachytherapy (Nag S. et al, 1999) Selection Criteria Brachytherapy Brachytherapy Brachytherapy recommended optional investigational (Good) (Fair) (Poor) PSA (ng/ml) < 10 10-20 > 20 Gleason score 5-6 7 8-10 Stage T1-T2 T2 T3 3 Prostate volume (cm ) < 40 40-60 > 60 Maximum urinary flow rate (ml/s) > 15 15-10 < 10 A number of societies such as ABS and the Groupe Europeen de Curietherapie – European Society for Therapeutic Radiology and Oncology (GEC-ESTRO) have made recommendations on indications and contraindications in the choice of LDR brachytherapy, which have been confirmed by several societies (Nag S. et al, 1999; Ash D. et al, 2000; Mohler J. et al, 2010). Table 2.4 demonstrates some of the recommendations. LDR brachytherapy technique is particularly a favorite in advanced countries such as USA, Japan, Netherlands and Spain. In Eastern European countries, however, HDR brachytherapy method is more popular in early staged prostate cancer treatment. As a monotherapy, LDR brachytherapy seems to be a reliable choice for early stage prostate cancer, according to low morbidity rate, good results and short 29 University of Ghana http://ugspace.ug.edu.gh hospitalization. It is a curative alternative of radical prostatectomy or EBRT (3-D conformal radiotherapy or intensity modulated radiotherapy) with comparable long-term survival and biochemical control, and most favorable toxicity (Grimm P. et al, 2012; Crook J.M. et al, 2011; Ferrer M. et al, 2008; Sylvester J.E. et al, 2011; Polascik T.J. et al, 1998). Low dose rate brachytherapy represents the most conformal radiation therapy (CRT), and the number of patients referred for this type of radical treatment has grown rapidly over the last 2 decades. Table 2.4: Indications and contraindications for LDR brachytherapy monotherapy according to ABS and GEC-ESTRO recommendations (Nag S. et al, 1999; Ash D. et al; 2000) Selection criteria ABS (low risk group) GEC-ESTRO PSA (ng/ml) < 10 < 10 Gleason score 2 – 6 5 – 6 Stage T1 – T2a T1c – T2a 3 Prostate volume (cm ) < 60 < 50 Life expectancy < 5 years < 5 years Distant metastasis + + 3 Gland size (cm ) > 60 > 50 Pubic arch interference + + Bleeding disorder – + There are several reasons why LDR brachytherapy achieved such popularity. In comparison with EBRT, LDR brachytherapy has better toxicity profile with higher dose to prostate gland. Comparing with radical prostatectomy, permanent seed implantation is a short, one day therapy with lower complication rate (bleeding, urinary incontinence, impotence) during and after the procedure. Specific selection of radioactive isotopes such as iodine-125 (Table 2.5) (Williamson J.F., 2002) and palladium-103 (Table 2.6) (Williamson J.F., 2000), and their correct localization, allows the deposition of high dose into the prostate tumor with rapid dose fall-off outside the 30 Contraindications Indications University of Ghana http://ugspace.ug.edu.gh area of treatment, which results in the preservation of organs at risk (OaRs). LDR brachytherapy has been a gold standard for prostate brachytherapy in low risk patients for many years. 2.4.1.1 Iodine-125 LDR Source Iodine-125 is a commonly used source for permanent implanted interstitial brachytherapy. Iodine-125 is manufactured into resin spheres which are encapsulated within a thin titanium shell. It decays through electron capture to Te-125, a stable isotope with the release of gamma photons (Figure 2.8). As iodine is for permanent implant, they do not require specific disposal, although if the patient dies within a year of insertion cremation is not recommended unless the implants are removed from the body first. The characteristics, advantages and disadvantages of iodine-125 brachytherapy seed are presented in Table 2.7. Fig 2.8: Decay scheme of I-125 31 University of Ghana http://ugspace.ug.edu.gh Table 2.5: Properties of iodine (I-125) LDR brachytherapy seed Characteristic Value 1. Energy of emitted photons - keV (number of photons per decay) 27.4 keV (1.15) 31.4 keV (0.25) 35.5 keV (0.067) Mean = 28 keV 2. Outer dimensions of source 4.5 mm x 0.8 mm 3. Half Life 59.4 days 4. Typical Prescription 150 – 160 Gy 5. Initial dose rate 7.0 cGy/hr 6. Exposure rate at 1 m < 0.3 mR/hr 7. HVL in tissue 1.8 cm 2 8. Specific Activity 6.4 x 10 TeV/g 9. 90% of total dose delivered in 197 days Advantages 1. Well Characterized dosimetry 2. Long term use (Permanent implant) 3. Pure gamma emitter Disadvantages 1. Very anisotropic dose distribution. 2. Relative long half-life (compared with Pd-103), which may complicate dose calculations over time. 3. Rounded seed ends make them mobile. 4. Fragile capsule. 5. Single use only. 2.4.1.2 Palladium-103 LDR Source Palladium-103 is a relatively new sealed source that is used in place of I-125, particularly for permanent interstitial prostate implants. It is coated onto graphite pellets and encapsulated within a titanium shell as seeds. Pd-103 is a pure gamma emitter, decaying through electron capture to Rh-103, as shown in Figure 2.9. 32 University of Ghana http://ugspace.ug.edu.gh Like I-125, Pd-103 has a relatively fragile capsule which may be damaged. It is safely stored within a lead safe at least 3 mm thick. Table 2.8 gives the characteristics, advantages and disadvantages of palladium-103 brachytherapy source. Fig. 2.9: Decay scheme of Pd-103 33 University of Ghana http://ugspace.ug.edu.gh Table 2.6: Properties of palladium (Pd-103) LDR brachytherapy seed Characteristic Value 1. Energy of emitted photons - keV (number of photons per decay): 20.1 keV (0.656), 23.0 keV (0.125) Mean = 21 keV 2. Outer dimensions of source 4.5 mm x 0.81 mm 3. Half life 16.97 days 4. Typical prescription 115 – 120 Gy 5. Initial dose rate 19.6 cGy/hr 6. Exposure rate at 1m < 0.15 mR/hr 7. HVL in tissue 1.1 cm 3 8. Specific activity 2.8 x 10 TeV/g 9. 90% of total dose delivered in 56 days Advantages 1. Short Half Life 2. Cupped seed ends tend to anchor seeds 3. Relatively safe for staff to handle during insertion due to low dose rate 4. Patients discharged from hospital immediately after procedure Disadvantages 1. Single use only 2. Fragile source capsule 3. Activity decays 4% a day 4. Dosimetry based on only two studies 5. Edema not resolved for 30 days may cause issues with implants 2.4.2 High-Dose-Rate Brachytherapy HDR brachytherapy is a temporary type of brachytherapy where a high-dose rate radioactive source, usually iridium-192 or cobalt-60 is placed in the prostate gland through an applicator implantation procedure. In a little over 3 decades now, HDR brachytherapy has been developed parallel to the LDR technique (Kovacs G. et al, 2005; Chiche A. et al, 2009; Martin T. et al, 2004; Galalae R.M. et al, 2002), and in the last few years with growing interest in the USA. 34 University of Ghana http://ugspace.ug.edu.gh HDR equipment is commonly available and the radioactive source used for treatment is the same as in the case of other neoplasms. The dwell-time position of the source in the applicators may be freely programmed during the procedure. The dwell time may be adapted to the requirements of treatment. In the course of treatment and real-time planning, the possibility of imprecise indication of the applicators position in relation to the treated gland is minimal, which ensures high precision of the treatment. HDR brachytherapy was initially introduced as a high-dose-rate supplement for EBRT, and proved to be an effective and safe method of treatment (Kovacs G. et al, 2005; Martinez A.A. et al, 2002; Demanes D.J. et al, 2000; Demanes D.J. et al, 2005). The technique enables small diameter catheters of around 2 mm external diameter to be used for interstitial implantation, with the great flexibility of dose delivery inherent in the stepping source system. Catheter placement exploiting the advantages of transrectal ultrasound and computed tomography (CT) imaging to provide accurate positioning and reconstruction of the implant can ensure high-quality implants and precise dosimetry related to anatomical structures. Delivery of radiation at high dose rate, however, carries biological implications which demand careful attention to dose fractionation and dose distribution. The challenge with HDR brachytherapy in prostate cancer is to deliver a safe, effective dose using a technique which will enable several fractions of treatment to be given over several days whilst retaining the high quality of the implant throughout. The presence of both LDR and HDR brachytherapy techniques in some countries is interesting to compare. 35 University of Ghana http://ugspace.ug.edu.gh 2.4.2.1 Iridium-192 HDR Source Iridium-192 is the most commonly used isotope for HDR brachytherapy applications using remote afterloading techniques. It is produced by neutron bombardment of stable Ir-191, and manufactured as a wire. The wire is coated in platinum to filtrate the electrons produced by the iridium decay process. Iridium-192 has an interesting decay pattern. As displayed in Figure 2.10, 95% of the time, Ir-192 decays through negative beta emission to Pt-192, and 5% of the time Ir- 192 decays through electron capture to Os-192. Fig. 2.10: Decay scheme of Ir-192 36 University of Ghana http://ugspace.ug.edu.gh Table 2.7 gives the characteristics, advantages and disadvantages of iridium-192 brachytherapy source. Table 2.7: Properties of iridium (Ir-192) HDR brachytherapy seed Characteristic Value 1. Energy of emitted photons - MeV Max = 0.612 MeV Min = 0.206 MeV Mean = 0.353 MeV 2. Half life 73.83 days 3. Typical prescription 36 – 40 Gy 4. Initial dose rate 4.28 mGy/hr 5. Exposure rate at 1m 4.81 mSv/hr 6. HVL in Lead 5.5 mm 2 7. Air kerma rate constant 108 μGym /GBq/hr Advantages 1. Small source size 2. Stable daughter product 3. Re-usable 4. Relative ease of manufacture Disadvantages 1. Replacement every 3 - 4 months 2. Frequent recalibrations due to radioactive decay 2.4.2.2 Cobalt-60 HDR Source Cobalt-60 is a synthetic radioactive isotope of cobalt which is produced artificially by neutron activation of the isotope Co-59. Co-60 decays by beta decay to the stable isotope nickel-60. The activated nickel nucleus emits two gamma rays with energies of 1.17 and 1.33 MeV, hence the overall nuclear equation of the reaction is: 59 27Co  n 60 60 27Co28Ni  e   e  gamma rays 37 University of Ghana http://ugspace.ug.edu.gh Properties of Co-60 are presented in Table 2.8. Table 2.8: Properties of cobalt (Co-60) HDR brachytherapy seed Characteristic Value 1. Energy of emitted photons - MeV Max = 1.33 MeV Min = 1.17 MeV Ave = 1.25 MeV 2. Half life 5.27 Years 3. Typical activity 18.5 GBq (0.5 Ci) 2 4. Air kerma rate constant 309 μGym /GBq/hr 5. Exposure rate at 1m 13 mSv/hr 6. HVL in Lead 11 mm Advantages 1. Long half life, hence source can be reused for a long time. Disadvantages 1. Higher emitting photon energy of Co requires more shielding. 2.4.3 Prostate Brachytherapy Seed Implant Techniques Most of both brachytherapy technical steps are similar. Many centers have improved and introduced their own techniques. In the operating room, a patient undergoing prostate brachytherapy is placed under general or spinal anesthesia in dorsal lithotomic position. After catheterization, contrast or air filled gel is usually used to visualize the urethra, and to differentiate the bladder from the prostate. First step of the procedure is the necessity to determine the shape and size of the gland by initial TRUS examination before needle insertion. This can be done a few days before insertion of needles or seeds (pre-implant treatment planning) or can be performed on the day of the procedure (intraoperative treatment planning). A bi-planar probe at 5, 6, or 7.5 MHz frequency, gather ultrasound visualization of prostate localization at 0.5 cm intervals, compared with the one after needle insertion. Treatment plan 38 University of Ghana http://ugspace.ug.edu.gh should contain several information such as needle location, number and strength of seeds (or number and position of HDR needles), and shape and volume of the target. To achieve the exact dose inside the prostate it is essential to use nomograms combined with real-time TRUS and treatment planning system (Pujades M.C. et al, 2011). Transrectal ultrasound equipment is combined with special template, and by guiding creates stepping unit. Before proper procedure it is important to measure the distance from bladder base to the template. Two needles are then inserted through the template just posterior to the urethra on either side of the midline to stabilize the prostate’s movement. Because of the movement during the procedure, a pre-plan can be created in order to minimize the risk of positioning errors. The loading pattern indicates coordinates in the computer planning system in connection with the templates stepping unit. That gives the physicians exact points to insert each needle. In this step, brachytherapy techniques differ one from the other. When the pre-plan is done, 20 cm long needles are inserted, and after consulting two plans (before and after insertion), radioactive seeds are placed into the prostate gland. In the case of LDR, where the sources are permanently left in the prostate, withdrawing each needle is done very carefully to avoid source migration inside the gland. Once the procedure has been completed, the position of seeds must be observed under fluoroscopy and ultrasound, in LDR technique. Usually, there is no possibility of removing the seeds after insertion and if “cold spots” are observed, a few extra seeds are added to cover them. Performing a final CT scan of the prostate and post-implant dosimetry ends up the whole procedure of LDR seeds implantation in prostate cancer treatment. The patient leaves the theatre catheterized, and after removing it, can be discharged home the next day. 39 University of Ghana http://ugspace.ug.edu.gh 2.4.4 Comparison of LDR and HDR Techniques Table 2.9 is a summarized comparison of LDR permanent implants and HDR temporal implants as compiled by the HDR Prostate Working Group (Skowronek J., 2013) and presented to radiation oncologists at the American Society of Therapeutic Radiology and Oncology (ASTRO) meeting in Phoenix – USA in October 1998. Table 2.9: Comparison of LDR and HDR implants (Skowronek J., 2013) Low Dose Rate High Dose Rate Conformal treatment + + + + + + + + Target accuracy + + + + + + + + Ability to treat extracapsular extension + + + + + Ability to treat seminal vesicles + + + + + + Ease of control of radiation + + + + + + Lack of cold/hot spots + + + + + + Control of critical organ dose + + + + + + Modify dose distribution + + + + + Need for external beam No / Sometimes Yes / Sometimes Monotherapy + + + + Experience of physician Crucial Crucial Pre-planning dosimetry Extensive (TRUS) Not needed Post implant dosimetry Extensive (CT) Not needed Stages treated T1 – T2 All (T1 – T3) 3 Prostate volume > 60 cm at time of implant More difficult Less difficult Pubic arch interference at time of implant Cannot be done Less of a problem Final dose verification Post-treatment Pre-treatment Symptom duration Months Weeks Implant cost Lower Higher + + + + (very high) + (low) 40 University of Ghana http://ugspace.ug.edu.gh CHAPTER THREE THEORY OF IMAGE FUSION 3.0 INTRODUCTION Image fusion is defined as the process of combining multiple input images into a single composite image. The aim of this is to create from the collection of input images a single output image which contains a better description of the scene than the one provided by any of the individual input images. The output image is therefore of improved quality information and more useful for human visual perception or for machine perception (Mitchell H.B., 2010). The basic problem of image fusion is one of determining the best procedure for combining the multiple input images. The view adopted in this study is that combining multiple images with a priori information is best handled within a statistical framework. In particular this study is restricted to classical and robust statistical approaches, Bayesian methods, sub-space and wavelet techniques, through the principles of mutual information and rigid body transformation. MATLAB image processing toolbox provides capabilities for image enhancement and image fusion. The benefits of image fusion include (i) reduced uncertainty, (ii) increased reliability, (iii) robust system performance, (iv) compact representation of information, and (v) extended range of operation. 3.1 IMAGE FUSION PROCESS The principal process in a generic image fusion processing chain is shown in Figure 3.1, for the case when the output is a single fused image ~I . 41 University of Ghana http://ugspace.ug.edu.gh Fig. 3.1: Generic image fusion processing chain (Mitchell H.B., 2010) The fusion processing chain consists of four principal blocks namely (i) multiple input images block, (ii) common representational format block, (iii) fusion block, and (iv) display block. 3.1.1 Multiple Input Images Block This block in the fusion processing chain is where multiple images of an external scene are captured by multiple cameras or scanners. The external scene could be an organ or a body part and scanner could be any of the imaging modalities such as ultrasound, PET, CT, or MRI. 3.1.2 Common Representational Format Block The input images are transformed into a common representational format in this block. In other words, the images are said to be compatible or “speak the same language” before fusion could be 42 University of Ghana http://ugspace.ug.edu.gh possible. This involves several processes including: spatial and temporal alignment, semantic equivalence, radiometric calibration, feature extraction and decision labeling. The principal functions in the common representational format block are:  Spatial Alignment: The input images are spatially aligned into the same geometric base. Without a common geometric base any information derived from a given input image cannot be associated with other spatial information. The accurate spatial alignment of the input images is therefore a necessary condition for image fusion. After spatial alignment the input images are re-sampled and if necessary the gray-levels of the input images are interpolated.  Temporal Alignment: The spatially aligned input images are temporally aligned to a common time. This step is only required if the input images are changing or evolving in time. In this case the accurate temporal alignment of the input images is a necessary condition for image fusion.  Feature Extraction: A feature is any distinguishing property or attribute of an image. Characteristic features are extracted from the spatially and temporally aligned input images. The output is one or more feature maps for each input image. Examples of features used in image fusion are edges, lines, patterns and colour.  Decision Labeling: Pixels in each spatially and temporally aligned input image or feature map are labeled according to a given criteria. The output is a set of decision maps.  Semantic Equivalence: In order for the input images, feature maps or decision maps to be fused together they must refer to the same object or phenomena. The process of causally 43 University of Ghana http://ugspace.ug.edu.gh linking the different inputs to a common object or phenomena is known as semantic equivalence.  Radiometric Calibration: The spatially, temporally and semantically aligned input images and feature maps are converted to a common measurement scale. This process is known as radiometric calibration. 3.1.3 Fusion Block After conversion of the images into a common representational format, the spatially, temporally, semantically and radiometrically aligned images are fused together in the fusion block. The fusion process may be classified into three classes: pixel fusion, feature fusion and decision fusion. The output is a fused image ~I , feature map ~F or decision map ~D (Gang H., 2007). For environments which are essentially static and in which the output is a single image ~I , three requirements are often imposed on the image fusion algorithms (Rockinger O., 1998): (i) Pattern conservation - The fusion process should preserve all relevant information on the input imagery in the composite image. (ii) Artifact free - The fusion scheme should not introduce any artifacts or inconsistencies which would distract the human observer or subsequent image processing stages. (iii) Invariance - The fusion scheme should be shift and rotational invariant, i.e. the fusion result should not depend on the location or orientation of an object in the input imagery. 44 University of Ghana http://ugspace.ug.edu.gh For environments which are evolving in time the input is a set of input sequences I k t , t T1,T2 , k 1,2,..., K and the output is a fused image sequence ~ I t . In this case, the following additional requirements are imposed on the image fusion algorithms: (iv) Temporal stability – The fused output should be temporally stable, that is, gray level changes in ~I t , should be present in at least one of the input sequences I k t . (v) Temporal consistency – Gray level changes which occur in the input sequences I k t must be present in the fused sequence ~I t . 3.1.4 Display Block The display block is where the fused output is processed for viewing. 3.2 PRINCIPLE OF MUTUAL INFORMATION Given a reference image A and a spatially aligned and re-sampled floating image B, the mutual information of A and B is defined as   MI A, B   pAB a,blog pAB a,b 2 dxdy (3.1) pA apB b Where; pA a is the probability a pixel x, y in A has a gray-level a, pB b is the probability a pixel x, y in B has a gray-level b and pAB a,b is the joint probability a pixel x, y in A has a gray-level a and the same pixel in B has a gray-level b. 45 University of Ghana http://ugspace.ug.edu.gh 3.2.1 Normalized Mutual Information The integral in Equation 3.1 is taken over the pixels which are common to both A and B. As a result, MIA, B may vary if the number of pixels which are common to A and B changes. In general the variations in MIA, B are small but they may lead to inaccuracies in a spatial alignment algorithm. To avoid these inaccuracies, we often use a normalized mutual information similarity measure in place of MIA, B . Four commonly used normalized MI measures are (Hossny M. et al, 2008):  MI A, B  ; H A H B    MI A, B ; minH A, H B NMI A, B  (3.2) MI A, B  ;  H A, B    MI A, B   H AH B Where H A and H B denote the entropies of image A and B respectively, and denoted by equations: H A  pA alog 2 pA adxdy (3.3) HB  pB blog2 pB bdxdy (3.4) HA, B  pAB a,blog2 pAB a,bdxdy (3.5) 46 University of Ghana http://ugspace.ug.edu.gh The integration in the equations is performed over the overlap region of the images A and B. 3.2.2 Calculation of Mutual Information The most common approach to calculate the mutual information MIA, B and the normalized mutual information NMIA, B is to calculate the entropies H A , H B and H A, B using the marginal probabilities pA a and pB b and the joint probability pAB a,b. The joint probability pAB a,b is the only factor considered for calculation, because the marginal probabilities pA a and pB b may be derived from the joint probability pAB a,b: i.e. pA a  pAB a,bdb , pB b  pAB a,bda . The most straight forward way to calculate the joint probability distribution pAB a,b is to use a discrete histogram H AB . 3.2.3 Histogram In calculating pAB a,b, the gray-levels in A and B are quantized into P and Q bins respectively. pAB a,b is then approximated using the 2D histogram H AB  hAB 1,1,hAB 1,2,....,hAB P,QT , where hAB p,q is the number of pixels whose gray-levels in A fall in the th p bin and whose gray-levels in B fall in the thq bin. In this case, the formula for the mutual information is 47 University of Ghana http://ugspace.ug.edu.gh  , h p,q  MI A B  h  AB AB p,qlog 2   hAB p,q, (3.6)  p,q   hA phB q p,q  where h p h p,q and A AB hB qhAB p,q . q p Although widely used, the histogram method suffers from several drawbacks: It yields a discontinuous density estimate and there is no principled method for choosing the size and placement of the bins. For example, if the bin width is too small, the density estimate is noisy while if the bin width is too large the density estimate is too smooth. Legg et al. (2007) recommends using Sturges’ rule for the optimal bin width: r w  (3.7) 1 log 2 K  where r is the range of gray-level values, K is the number of elements in the input image. In this case the optimal number of bins is r w . A partial solution to these problems is to calculate pAB a,b using the method of Parzen windows. 3.2.4 Parzen Windows Instead of using discrete histogram bins to calculate the joint probability distribution pAB a,b, we use continuous bins. This is known as kernel, or Parzen-window density estimation (Scott 48 University of Ghana http://ugspace.ug.edu.gh D.W., 1992; Silverman B., 1986) and is a generalization of histogram binning. If A and B each contain K pixels with gray-levels ak ,bk ,k 1,2,..., K, then the estimated joint density pAB a,b is given by: 1 K K  a  a   b  b  p  k   l AB a,b  2 H H  (3.8) K   k1 l1   A   A B B  where H x denotes a kernel function which satisfies H xdx 1. In general a density x estimate px is more sensitive to the choice of the bandwidth  and less sensitive to the choice of the kernel H x . For this reason we often use a zero-mean Gaussian function with standard deviation  for the kernel H x . Table 3.1 demonstrates a scheme which is commonly used to calculate the optimal bandwidth  . Table 3.1: Method for Calculating Optimum One-Dimensional Bandwidth  Name Description Rule-of-Thumb Suppose the input data (consisting of N measurements ai ,i1,2,..., N), is generated by a given parametric density function, e. g. a Gaussian function. In this case  1.06ˆN 1 5 , where ̂ is the sample standard deviation. Robust versions of this bandwidth are available:  1.06minˆ,Qˆ 1.34N 1 5  1.06sˆN 1 5 and , where Q̂ is the sample interquartile distance and sˆ  med j a j med i ai . 49 University of Ghana http://ugspace.ug.edu.gh 3.2.5 Iso-intensity Lines Iso-intensity lines (Rajwade A. et al, 2009) is a new scheme developed specifically for calculating the joint probability density pAB a,b. Suppose the gray-levels in A and B are quantized, respectively, into P and Q bins. For each pixel location m,n , the gray-level values G1, G2, G3, G4 of its four neighbors which lie at a horizontal or vertical distance of half a pixel from m,n is estimated. The square defined by these neighbors is divided into a pair of triangles as shown in Figure 3.2. Within the triangle, it is supposed the gray-level values vary linearly as follows: Amx,n y aAx bA y  cA (3.9) Bmx,n y aB x bB y  cB (3.10) Fig. 3.2: Shows the pixel m,n with the gray-levels Gk , k 1,2,3,4 50 University of Ghana http://ugspace.ug.edu.gh where Amx,n y and Bmx,n y denote, respectively, the gray-level of a point mx,n y in the triangle and 0.5  x,y  0.5 . To calculate the joint distribution of the two images A and B we sequentially consider the PQ different gray-level pairs denoted as  ,  . For each pixel m,n we see whether the pair of corresponding triangles contains a point mx,n y which has a gray-level value  in A and  in B. Such a point mx,n y contributes a vote to the entry  ,  in pAB a,b. 3.2.6 Partial Volume Interpolation The histogram, Parzen and iso-intensity line algorithms all assume the images A and B are spatially aligned and if necessary image interpolation has been performed. The partial volume interpolation (PVI) is an alternative technique which does not assume spatial alignment or image interpolation (Tsao J., 2003). It works as follows: Suppose T represents a mapping of the pixel x, y in B into the corresponding location u,v in A. In general u,v will not correspond to a pixel location in A. Suppose Qk  uk ,vk  , k 0,1,2,3, are the four pixel locations in A which surround u,v as shown in Figure 3.3. Then if Auk ,vk  has a quantized gray-level  k and Bx, y has a quantized gray-level  , then H AB K ,  receives a fractional vote equal to 3 r 1k  r 1 h , h0 51 University of Ghana http://ugspace.ug.edu.gh where rk    2 2 u u  v  vk k  Fig. 3.3: Showing four pixel locations Qk  uk ,vk  , k 0,1,2,3, in A which surround the transformed point u,v . Also shown is the Euclidean distance 2r3  u3 u v3  v from Q3 to u,v . 3.2.7 Artifacts The success of the mutual information algorithms image registration lies in their inherent simplicity. It makes few assumptions regarding the relationship that exists between different images. It only assumes a statistical dependence. The idea is that although different cameras may produce very different images, since they are imaging the same underlying scene, there will exist some inherent mutual information between the images. When the images (or image patches) are spatially aligned, then the mutual information is maximal. To be an effective similarity measure, however, it is required that the mutual information falls monotonically to zero as we move away 52 University of Ghana http://ugspace.ug.edu.gh from perfect alignment. In practice, the mutual information does not fall monotonically to zero. These artifacts are due to inaccuracies in estimating the marginal densities pA a and pB b and the joint density pAB a,b. The two types of artifacts which appear in mutual information image registration are interpolation effects (Tsao J., 2003) and small size effects (Andronache A. et al, 2006). 3.3 RIGID BODY TRANSFORMATION Rigid-body transformations consist of only rotations and translations, and leave given arrangements unchanged. They are a subset of the more general affine transformations. Affine transformation means that parallel lines remain parallel after the transformation. For each point x1, x2 , x3  in an image, an affine mapping can be defined into the co-ordinates of another space y1, y2 , y3  . This is expressed as: y1  m11x1  m12x2  m13x3  m14 y2  m21x1  m22x2  m23x3  m24 (3.11) y3  m31x1  m32x2  m33x3  m34 which is often represented by a simple matrix multiplication y  Mx:  y1  m11 m12 m13 m14  x1        y  2 m m m m x    21 22 23 24  2  (3.12) y3  m31 m32 m33 m34 x3      1   0 0 0 1     1   53 University of Ghana http://ugspace.ug.edu.gh The elegance of formulating these transformations in terms of matrices is that several of them can be combined, simply by multiplying the matrices together to form a single matrix. This means that repeated re-sampling of data can be avoided when re-orienting an image. Inverse affine transformations are obtained by inverting the transformation matrix. 3.3.1 Translation If a point x is to be translated by q units, then the transformation is simply: y  x  q (3.13a) In matrix terms, this can be considered as:  y1  1 0 0 q1  x1        y  2 0 1 0 q x   2   2  (3.13b) y3  0 0 1 q3  x3    1    0 0 0 1     x4  3.3.2 Rotation In two dimensions, a rotation is described by a single angle. For a point at co-ordinate x1, x2  on a two-dimensional plane, a rotation of this point to new co-ordinates y1, y2 , by  radians around the origin, can be generated by the transformation: y1  cos x1  sin x2 (3.14) y2   sin x1  cos x2 54 University of Ghana http://ugspace.ug.edu.gh For a three-dimensional case, there are three orthogonal planes that an object can be rotated in. These planes of rotation are normally expressed as being around the axes. A rotation of q1 radians about the first x axis is normally called pitch, and is performed by:  y1  1 0 0 0 x1     y2 0 cosq  sin  0    q x     1 1   2  (3.15a) y3  0  sinq1  cosq1  0 x3    1    0 0 0 1      1  Similarly, rotations about the second y and third z axes (called roll and yaw respectively) are carried out by the following matrices:  cosq2  0 sinq2  0  cosq3  sinq3  0 0  0 1 0 0  sinq3  cosq3  0 0    and   (3.15b)  sinq2  0 cosq2  0  0 0 1 0   0 0 0 1     0 0 0 1   Rotations are combined by multiplying these matrices together in the appropriate order. The order of the operations is important. For example, a rotation about the first axis of  2 radians followed by an equivalent rotation about the second would produce a very different result to that obtained if the order of the operations was reversed. 3.3.3 Zoom The affine transformations described so far will generate purely rigid-body mappings. Zooms are needed to change the size of an image, or to work with images whose voxel sizes are not 55 University of Ghana http://ugspace.ug.edu.gh isotropic, or differ between images. These represent scalings along the orthogonal axes, and can be represented via:  y1  q1 0 0 0 x1        y  2 0 q 0 0 x    2   2  (3.16) y3   0 0 q3 0 x3   1     0 0 0 1      1  A single zoom by a factor of -1 will flip an image. Two flips in different directions will merely rotate it by  radians (a rigid-body transformation). In fact, any affine transformation with a negative determinant will render the image flipped. 56 University of Ghana http://ugspace.ug.edu.gh CHAPTER FOUR METHODOLOGY 4.0 INTRODUCTION Methodology provides the systematic theoretical analysis of the methods applied to the field of study. This chapter presents materials and equipment used for the research study and the methods employed in achieving the results. 4.1 MATERIALS AND EQUIPMENT Materials and equipment involved in this study among others include transrectal ultrasound (TRUS) system, calibrated stepper with template grid, prostate brachytherapy quality assurance (QA) phantom, PET-CT system, PET-CT quality control (QC) kits, prostate phantom, and MATLAB software.  Transrectal Ultrasound System: Ultrasound system (BK Medical Falcon 2101 EXL, USA) used for acquiring ultrasound images in the study is shown in Figure 4.1. The system is connected to a transrectal probe, which is held in position and its movement controlled by a calibrated stepper. The stepper enables accurate and reproducible positioning of the probe within the rectal canal. The system has multitude of uses with an imaging frequency range of 1 to 16 MHz. 57 University of Ghana http://ugspace.ug.edu.gh Fig. 4.1: Ultrasound system with transrectal probe (at NCRNM, Korle-Bu Hospital)  PET-CT System: PET-CT system (Biograph 40, Siemens, USA) was used in acquiring the PET and CT images for the study. The system, as shown in Figure 4.2, has a 40-slice CT component, which provides a 360o full scan of an average patient in an approximate time of 42 seconds. The PET component uses Lutetium oxyorthosilicate (LSO) detector which produces faster scintillation, higher light yield, and faster light decay, compared with other detectors. Dimensions of the LSO detectors are 4420 mm3, giving it a high resolution capability. The system has a 21.8 cm axial field-of-view (FOV) and 70 cm spacious gantry opening for easy and flexible patient positioning, enhanced patient comfort, and convenience. 58 University of Ghana http://ugspace.ug.edu.gh y x z Fig. 4.2: PET-CT system (at Nuclear Medicine Dept., CM Johannesburg Academic Hospital)  Prostate Phantom: The multi-modality prostate phantom (CIRS Model 053-MM, USA), as shown in Figure 4.3, was imaged on the ultrasound and PET-CT systems. The phantom is disposable and developed for practicing procedures which involve scanning under ultrasound, CT and MRI systems. The prostate along with structures simulating the rectal wall, seminal vesicles and urethra is contained within a 11.579.5 cm3 clear acrylic container. A 3 mm simulated perineal membrane enables various probes and surgical tools to be inserted into the prostate capsule, which contains simulated lesions. 59 University of Ghana http://ugspace.ug.edu.gh Fig. 4.3: Multi-modality prostate phantom (at Medical Physics Dept., SNAS, Univ. of Ghana)  MATLAB Software: MATLAB (Version 7.8, MathWorks Inc., USA) image processing toolbox was employed in developing contrast enhancement and fusion algorithms for the study. The software is a high-performance technical computing language which is employed for tasks such as algorithm development, modeling, simulation, and data analysis. Extensive set of functions are available in MATLAB for processing multi-dimensional arrays, of which matrices (two-dimensional numerical arrays) have a one-to-one correspondence with digital images. 4.2 METHOD A number of stages were employed in achieving the main objective of fusing images from ultrasound and PET-CT modalities. Extensive quality control tests were initially performed on 60 University of Ghana http://ugspace.ug.edu.gh the ultrasound and PET-CT systems after which they were used in acquiring images of the prostate phantom. Algorithms were then written in MATLAB to enhance the contrast of acquired images and subsequently fuse the ultrasound and PET-CT images together. Error analysis was performed on the fused images to assess the integrity of the fusion algorithm. 4.2.1 Performance Evaluation of Imaging Systems State-of-the art imaging systems require periodic calibrations. The primary purpose of performing quality control is to ensure optimal operation of the imaging equipment in terms of image quality and accuracy, as well as safety of patients and operators. Quality control tests were performed on the ultrasound and PET-CT systems to assess their performances. Satisfactory results from the QC tests gave credence for the use of the imaging systems in this study. 4.2.1.1 Ultrasound Brachytherapy System Quality Control Accurate image guidance and dosimetry calculation in prostate brachytherapy depend upon the quality and accuracy of ultrasound images used. Hence, robust quality assurance program recommended by Task Group (TG) 128 of the American Association of Physicists in Medicine for ultrasound brachytherapy system is essential. Major QC tests performed to assess the functionality of the ultrasound brachytherapy system in this study were axial and lateral distance measurement accuracies, volume measurement accuracy, depth of penetration, and axial and lateral resolution tests. Prostate brachytherapy QA phantom (CIRS Model 045, USA), shown in Figure 4.4(a), was used in performing the tests. Schematic diagram of the phantom showing the 61 University of Ghana http://ugspace.ug.edu.gh top, front, left and right views is presented in Figure 4.4(b). The phantom is manufactured of zerdine®, a tissue mimicking material with sound speed of 1540 m/s, and contains targets to assess volume measurements, depth of penetration, volume measurements, among others. Fig. 4.4(a): Prostate brachytherapy QA phantom (at Medical Physics Dept., SNAS, Univ. of Ghana) The phantom consists of 13 parallel target wires of 0.1 mm diameter, making an N-shape and positioned at 10 mm intervals in the lateral and axial directions. Position of a target wire is defined by a combination of rows (1 – 5) and columns (B – F). Three volumes made of zerdine® with sizes 4 cm3, 9 cm3 and 20 cm3 located in the phantom are used for volume measurement accuracy. 62 University of Ghana http://ugspace.ug.edu.gh 20 cc 20 cc Fig. 4.4(b): Schematic diagram of phantom For all tests conducted, condom sleeve was put around the probe to reproduce clinical settings and coupling gel smeared on scanning surface to allow for smooth transmission of ultrasound waves, removal of air spaces and reduction of artifacts. All tests were performed in accordance with the American Association of Physicists in Medicine (AAPM) Task Group (TG) 128 report on quality assurance tests for prostate brachytherapy ultrasound systems (Pfeiffer D. et al, 2008). Images were recorded to document the measurements made to aid reproducibility and also serve as points of reference. Setup for the measurements is shown in Figure 4.5. 63 University of Ghana http://ugspace.ug.edu.gh Fig. 4.5: Setup for ultrasound system QC tests 4.2.1.1.1 Axial and Lateral Distance Measurement Accuracy Distance measurements in the axial and lateral directions are of importance to make accurate measurements during a procedure, such as to determine the distance of a needle from another needle or from the rectal wall (Pfeiffer D. et al, 2008). In taking measurements, scan parameters were set consistent with clinically used parameters. Ultrasound probe was held in position on the QA phantom such that axis of the ultrasound beam was perpendicular to the wires. Once the N- shaped targets were observed in the ultrasound image as shown in Figure 4.6, the image was 64 University of Ghana http://ugspace.ug.edu.gh frozen, and distances between the targets in the axial and lateral directions measured with electronic caliper. Tolerance levels of ± 2 mm and ± 3 mm were respectively set for the axial and lateral distance tests as recommended in the AAPM TG 128 (Pfeiffer D. et al, 2008). Fig. 4.6: Ultrasound image showing N-shaped targets for axial and lateral distance measurements Measurements were taken at 0.5 cm stepwise movements of the ultrasound probe, from an initial position of 0.5 cm to 10.5 cm into the phantom. Results of axial and lateral measurements are presented in Table 5.1 and Table 5.2 respectively. Each measurement was repeated three times and their average estimates presented in the tables. 4.2.1.1.2 Volume Measurement Accuracy Volume measurement was performed to ascertain the trueness of organ volumes as they appear in the ultrasound images (Pfeiffer D. et al, 2008). After setting up the QA phantom with the 65 University of Ghana http://ugspace.ug.edu.gh probe in position, the probe was turned 60o clockwise and anticlockwise to visualize the three volume inserts on the ultrasound monitor. The standard sizes of the three volumes are 4 cm3, 9 cm3 and 20 cm3. Axial images of the volumes were contoured from the base to the apex as shown in Figure 4.7, and their respective volume estimates determined and presented in Table 5.3. For each volume insert, the test was repeated five times and the average volumes estimated. Tolerance level of ± 5% deviation between estimated volume and nominal target volume, as suggested by AAPM TG 128 (Pfeiffer D. et al, 2008), was set for this test. Fig. 4.7: Cross-sectional image of QA phantom’s volume insert for volume estimation 4.2.1.1.3 Depth of Penetration This test measures the sensitivity of the system, which determines how deep into the patient a low contrast object can be reliably visualized (Pfeiffer D. et al, 2008). Ultrasound probe was 66 University of Ghana http://ugspace.ug.edu.gh placed on the scanning surface of the QA phantom during set up and coupling gel applied to remove air bubbles. A location was identified in the image relatively clear of highly reflective targets. Using frequency of 6.5 MHz, which is most commonly used for implants on the system, the maximum depth at which the static ultrasound speckle pattern of the phantom was clearly distinguished from the dynamic electronic noise, was identified. The depth of penetration was measured with electronic caliper. The test was repeated three times and in each case measurement was taken at five different measuring points along the boundary of the speckled pattern. Average depth of penetration was estimated as presented in Table 5.4. Tolerance level set for the depth of penetration test is a change of less than 1 cm between the maximum depth of penetration and its corresponding baseline value (Pfeiffer D. et al, 2008). 4.2.1.1.4 Axial and Lateral Resolution After set up, the QA phantom was scanned in the axial array direction to display the N-targets. The image was frozen and dimensions of the targets in the image were measured axially and laterally. The probe was switched into orthogonal (longitudinal array) direction and steps repeated. Results for the test are presented in Table 5.5. A tolerance level of 1 mm deviation between measured axial and lateral resolution from baseline was set for this test in accordance with recommendation of AAPM TG 128 (Pfeiffer D. et al, 2008). 4.2.1.2 PET-CT Quality Control Quality control tests performed on the PET-CT were CT laser alignment, table (bed) position accuracy, CT display width, CT image uniformity, CT dose index (CTDI), PET-CT image 67 University of Ghana http://ugspace.ug.edu.gh uniformity, PET-CT image registration accuracy, PET-CT image contrast, and PET-CT image resolution. Setups and procedures for performing the tests are outlined under this section and results presented in Chapter Five. 4.2.1.2.1 CT Laser Alignment The laser alignment test was performed to ensure that the gantry lasers were properly aligned with CT gantry and patient table. Tolerance level for this study was set to ±1 mm as recommended in IAEA Human Health Series No. 1 (IAEA HHS 1, 2009). An in-house developed laser alignment phantom and a metre rule were the materials used for the test. y z x Fig. 4.8: Set-Up for CT Laser Alignment Test 68 University of Ghana http://ugspace.ug.edu.gh The alignment test phantom, which has reference scribe marks to check laser alignment, was placed on top of the table (i.e. PET-CT bed) and aligned with the sides of the table. It was centered and aligned orthogonally with the long axis of the table, using the gantry lasers. The set- up was then checked to see if all gantry lasers (in the x, y and z planes) correctly aligned with the reference scribe marks, as shown in Figure 4.8. The metre rule was used to check for any shift in laser position from the phantom scribe marks and results recorded. Results for this test are presented in Table 5.6. 4.2.1.2.2 Table (Bed) Position Accuracy The aim of the table (patient bed) position accuracy test is to ensure that the table’s horizontal and vertical motions according to digital indicators are accurate and reproducible; table indexing and positioning under scanner control is accurate. This test is especially important if the images are being used for radiation therapy treatment planning. Equipment used for this test were laser alignment phantom, graph paper, measuring tape and 70 kg weight. Tolerance level of ±1 mm was set for this test as recommended in the International Atomic Energy Agency – Human Health Series (IAEA HHS) No. 1 (IAEA HHS 1, 2009). The table was positioned centrally, otherwise called zero position, within the CT gantry and a 70 kg weight placed at the end of the patient bed. To determine the table movement and position accuracy under manual and computer control, the graph paper and measuring tape were positioned midline at the head end of the table, and accurately aligned with the long axis of the table using CT lasers. With the aid of lasers, the central portion of the graph paper was marked as a reference point. The table was repeatedly moved in the horizontal ( z -plane) and vertical ( y - 69 University of Ghana http://ugspace.ug.edu.gh plane) directions to defined distances from the central reference point using the system’s control console, and the table displacement was measured with the measuring tape. Measurements were taken for  z ,  z ,  y and  y table movements. Results for the test are presented in Table 5.7. 4.2.1.2.3 CT Display Width Display width test is performed to ensure that the volume of the patient being measured and displayed is similar to that selected on the CT scanner console. Materials used for this test were CT quality assurance phantom and metre rule. Tolerance level for this test was that the measured width of the displayed volume should be within ±1 mm of the width selected on the CT operator’s console, as recommended in IAEA HHS No. 1 (IAEA HHS 1, 2009). The quality assurance phantom was mounted to the CT table and aligned such that the volume of interest was parallel to the acquisition plane. A series of CT images were acquired in spiral mode using the nominal slice thicknesses used for clinical brain technique of 1 mm. The widths of the octagonal plane and the smallest circular section of the phantom were physically measured using the metre rule, and compared with their corresponding display widths in the acquired images. The set-up and acquired image are shown in Figure 4.9(a) and 4.9(b) respectively. Results for this test are presented in Table 5.8. 70 University of Ghana http://ugspace.ug.edu.gh 20.5 cm (a) Fig. 4.9(a): Set-up of quality assurance phantom for display width test 20.4 cm (b) Fig. 4.9(b): Acquired image of orthogonal section for display width test 71 University of Ghana http://ugspace.ug.edu.gh 4.2.1.2.4 CT Image Uniformity The aim of this test is to ensure that CT numbers (pixel values) are uniform over the image. Tolerance level of ± 5 HU between periphery CT numbers and the central CT number was set for this study (IAEA HHS 1, 2009). Material used for the test is the CT quality assurance phantom. Fig. 4.10: Cross-sectional slice with regions of interest The quality assurance phantom was mounted to the patient bed and positioned using lasers. The phantom was scanned with the CT part of the system to obtain tomographic images across it. A slice which is mid-way along the uniform part of the phantom across the scan length was 72 University of Ghana http://ugspace.ug.edu.gh selected, and five regions of interest (ROIs) drawn over it. ROI-3 was drawn at the centre of the slice, and ROI-4, ROI-5, ROI-6 and ROI-7 drawn at the top, right, bottom and left peripheries respectively, as shown in Figure 4.10. The PET-CT system software displayed the CT numbers in the respective regions. Uniformity of the slice was estimated as the difference between mean CT numbers of the peripheral ROIs and the mean CT number of the central ROI. Noise in the image was estimated as the standard deviation of the CT numbers divided by the average of the CT numbers in each of the ROIs. Results for the uniformity and noise tests are presented in Table 5.9. 4.2.1.2.5 Computed Tomographic Dose Index (CTDI) Computed tomography dose index determination was performed to ensure that appropriate radiation doses are being used for patient CT scans. Head (16-cm diameter) and body (32-cm diameter) acrylic dosimetry phantoms, pencil ion chamber, electrometer, thermometer and barometer were the instruments used for this test. Tolerance level of ± 20% between estimated dose and CT console displayed dose was set for this study as recommended in IAEA Human Health Series No. 1 (IAEA HHS 1, 2009). Measurements were taken by setting up the head and body phantoms in succession. The head phantom was first setup on the PET-CT couch and centred at the isocentre of the scanner with the long axis of the phantom aligned with the z-axis of the scanner. Scoutview and single 1 mm slice image of the phantom was acquired for alignment purposes. The ion chamber was placed in the center of the phantom and a scoutview image used to select the volume or slice to be imaged. Dosimeter readout was set to zero, and exposure in axial mode at brain CT scan technique (120 73 University of Ghana http://ugspace.ug.edu.gh kV, 150 mAs) was used for the head phantom study. Charges (in nanocoulombs) were measured in the central and peripheral holes by changing the ion chamber position from one hole to the other as shown in Figure 4.11. After taking head phantom measurements, the procedure was repeated on the body phantom. Pelvic CT scan techniques (120 kV, 100 mAs) were set for the body phantom measurements. All the measurements were taken at temperature of 21 oC and pressure of 832 hPa. Results from the CTDI test are presented in Table 5.10(a) and Table 5.10(b). Fig. 4.11: Setup for CTDI determination 74 University of Ghana http://ugspace.ug.edu.gh Electrometer readings were taken in charge mode, corrected for temperature and pressure, and converted into exposure (rad) using equation 4.1a. CTDI100, CTDIw and CTDIvol were estimated using equations 4.1b, 4.1c and 4.1d respectively. The average absorbed radiation dose over the x, y, and z directions in the body is given by CTDIvol (AAPM TG 23, 2008). The estimated values were compared with the computer displayed CTDIvol for the two examinations. Q Q 1 X (rad)  (C/kg)   4 (R). f (rad/R) (4.1a) 2.58 10 medmair mair  X (rad) C f .L(mm) CTDI100  (4.1b) N T (mm) 1 2 CTDI  CTDI centre  CTDI peripheryw 100 100 (4.1c) 3 3 CTDI CTDIvol  w (4.1d) p f where fmed - [Exposure to dose conversion factor] = 0.78 rad/R, C f - [Electrometer / ion chamber calibration factor] = 1, L - [Ion chamber length] = 100 mm, T - [Width of one slice or tomographic selection] = 4 mm, N - [Number of slices or tomographic sections imaged in a single axial scan] = 40, X - [Estimated exposure] = Q m p f - [Pitch factor] = 0.938. mair - [Mass of air irradiated] = air vair air - [Density of air at Standard Temperature and Pressure] = 1.293 kg/m 3 vair - [Vol. of irradiated air for single slice] = 4 mm  v = 0.04  v 100 mm c c vc - [Vol. of ion chamber] = 3.14 cm 3 = 3.14  10-6 m3 Hence, mair = 1.2930.043.1410 6 = 1.624107 kg (at STP) 75 University of Ghana http://ugspace.ug.edu.gh 4.2.1.2.6 PET-CT Image Uniformity Uniformity of a reconstructed PET-CT image was measured to assess the system’s response to a homogeneous activity distribution in the axial field of view. Cylindrical uniform phantom with internal diameter of 190 mm and a volume of 6.58106 mm3 was used in performing the uniformity test. As recommended by HHS 1, the set tolerance for this study was %NUmeasured < 1.05%NUreference (IAEA HHS 1, 2009). The cylindrical phantom was filled with water and 55.5 MBq F-18 FDG activity injected into it. The cylinder was covered with the lid and the phantom was repeatedly inverted to thoroughly mix the activity with the water. Activity concentration of 8.43 kBq/mL was attained by adding water to completely fill the phantom. The phantom was mounted to the patient bed (as shown in Figure 4.12(a)) and positioned using lasers. The phantom was scanned both with CT and PET to obtain tomographic images acquired across it. Slices corresponding to the central long active part of the phantom was reconstructed with all the corrections applied using the system software. Calculated attenuation correction was applied using the corresponding CT images to avoid noise propagation. Reconstructed transaxial and sagittal slices of the images were displayed and carefully inspected visually for artifacts. Estimation of quantitative index of non-uniformity was performed by drawing a circular area of 175 mm diameter centered inside the mid-transaxial slice of the phantom. Orthogonal grid of square regions of interest, approximately 1010 mm2, were drawn on the slice inside the circular area as shown in Figure 4.12(b). From Figure 4.12(b), for each square ROI k in slice i , the average, maximum and minimum counts inside the thk ROI were obtained. Non-uniformity and the coefficient of variation inside 76 University of Ghana http://ugspace.ug.edu.gh the slice were evaluated using equation 4.2 and 4.3, and their results respectively presented in Table 5.11(a) and Table 5.11(b). (a) Fig. 4.12(a): Cylindrical phantom under scanning for PET-CT uniformity test 10x10 mm ROIs 175 mm diameter Maximum NU ROI (b) Fig. 4.12(b): Reconstructed slice of cylindrical phantom with 1010 mm square ROIs 77 University of Ghana http://ugspace.ug.edu.gh 100 MAX Ck AVE Ck  AVE Ck   NU slicei  MAX   AVE C MIN C  (4.2) 100 k k AVE Ck   SDslicei CVslicei 100  (4.3) AVECk  where 1 k 2SDslicei  Ck  AVECk  (4.4) NROIs 1 k1 NROIs denote the number of the square ROIs inside the 175 mm diameter circle. 4.2.1.2.7 PET-CT Image Registration Accuracy This test was performed to assess quantitatively the accuracy of the registration of the images obtained with the PET and CT scanners. Since the fusion of PET and CT images assumes correct registration of the two modalities, it is crucial to ensure that the two studies are perfectly registered for a given patient weight. Tolerance level for the registration accuracy was set to ± 1 mm or ± 1 pixel as recommended in IAEA Human Health Series No. 1 (IAEA HHS 1, 2009). The image quality phantom (NEMA 2001 body phantom, USA) (IEC, 2008) was used for the image registration accuracy test. The phantom was also used for the image contrast assessment in this study. The phantom was filled with F-18 solution of 5.3 kBq/mL activity concentration. Two spheres in the phantom with diameters 3.7 cm and 2.8 cm were filled with cold water to mimic cold lesion 78 University of Ghana http://ugspace.ug.edu.gh imaging. Two other spheres with diameters 2.2 cm and 1.3 cm were filled with F-18 solution of concentration 42.4 kBq/mL, such that ratio of concentration between the hot spheres and the background was approximately 8:1. One other sphere of diameter 1.7 cm was left empty. The phantom was set up in a supine position on the patient bed such that the centres of all the spheres were in the same transverse slice as shown in Figure 4.13(a). Lead bricks were uniformly distributed over 1.5 m length of the bed adjacent to the image quality phantom. The phantom was scanned on the PET and CT scanners and the respective images acquired in 512512 matrix. The reconstructed PET and CT volumes were displayed simultaneously using the display-fusion software of the PET-CT system as shown in Figure 4.13(b). Displacements between PET and CT images at the edges of the hot spheres and the image quality phantom were measured with electronic ruler in the fusion software. Results of the deviations are presented in Table 5.12. (a) Fig. 4.13(a): Image quality phantom under scanning 79 University of Ghana http://ugspace.ug.edu.gh (b) Fig. 4.13(b): Cross-sectional reconstructed PET-CT image of image quality phantom 4.2.1.1.8 PET-CT Image Quality Images obtained from the set-up of the PET-CT image registration accuracy test (Section 4.2.1.2.7) were used for the image quality assessment, which involves contrast and background variability. One transverse slice centered on the cold and hot spheres as shown in Figure 4.13(b) was selected for the analysis. Comparative to the other slices, the selected slice visually had the highest contrast between the hot and cold spheres. Circular regions of interest (ROIs) were drawn on the hot and cold spheres, and also on the background in the image slice. A total of twelve ROIs were drawn throughout the background, and counts of activity determined for each ROI. 80 University of Ghana http://ugspace.ug.edu.gh ROIs of the same sizes as the smaller spheres (1.3, 1.7, 2.2, and 2.8 cm) were drawn concentric to each of the 3.7 cm ROIs on the background. The average counts in each background ROI was recorded, and percentage contrast QH , j for each hot sphere n was calculated by Equation 4.5. C 100 H ,n CB,n  CB,n QH ,n   (4.5) aH  aB  aB where CH ,n denotes the average counts in the ROI for sphere n , CB,n is the average of the background ROI counts for sphere n , aH is the activity concentration in the hot spheres, and aB is the activity concentration in the background. The percent contrast QC ,n for each cold sphere n is computed as CB,n C C ,nQC ,n 100  (4.6) CB,n where CC ,n is the average counts in the ROI for sphere n , and CB,n is the average of the background ROI counts for sphere n . The percentage background variability N n for sphere n is calculated as 1 K   2 C 1 B,n,k CB,n  K  100 k1N n   (4.6) CB,n where K is equal to the number of background ROIs. The percentage contrast and the percentage background variability results are presented in Table 5.13(a) and Table 5.13(b). 81 University of Ghana http://ugspace.ug.edu.gh 4.2.1.2.9 PET-CT System Resolution Resolution test was performed to assess the smallest tumour size the PET-CT system would be able to pick up during imaging. Jaszczak phantom was used in performing this test. The phantom, which contains parallel rods to simulate different sizes of tumours, was filled with water to half its volume and 74 MBq F-18 solution added. The phantom was shaken to achieve uniform mixture and fully filled with water to the brim. The phantom was then scanned on the PET-CT system as shown in Figure 4.14(a) and the resultant image displayed in Figure 4.14(b). (a) Fig. 4.14: (a) Jaszczak phantom under scanning 82 University of Ghana http://ugspace.ug.edu.gh Sector with least resolved rods (b) Fig. 4.14(b): Cross-sectional reconstructed PET-CT image of Jaszczak phantom Resolution of the system was assessed as the full width half maximum (FWHM) estimate for the least resolved rod in the image. The rods are grouped into six sectors raging from smallest to biggest. The acquired image was imported into Image J software and lines of interest drawn over eight randomly selected rods in each of the sectors to obtain profile plots as shown in Figure 4.14(c). Full width half maximum estimates for the selected spheres were recorded and their averages calculated and presented in Table 5.14. The average FWHM estimated for the least resolved rods in the image signifies the resolution of the PET-CT system. The recommended tolerance for this test according to HHS 1 is Rmeasured < 1.05Rexpected (IAEA HHS 1, 2009). 83 University of Ghana http://ugspace.ug.edu.gh Fig. 4.14(c): Profile plot of counts vs. distance for estimation of PETC-CT system resolution 4.2.2 Acquisition of Prostatic Images Acquisition of images of the prostate phantom was done in succession on the ultrasound and PET-CT systems in two setup conditions (with and without needle insertions). Ultrasound image of the phantom was first acquired without needle insertion. The phantom was positioned on table and transrectal probe positioned at the prostate region as shown in Figure 4.15(a). Axial images were acquired in equal and stepwise increments of 5 mm from base of the prostate capsule to the apex, acquiring twelve equal depth slices. From the perineal membrane portion of the phantom, four prostate brachytherapy needles were then pushed through the prostate capsule and procedure repeated as shown in Figure 4.15(b). Acquired prostatic images, as presented in chapter five (section 5.2) were transferred unto a computer system for contrast enhancement and fusion procedures. 84 University of Ghana http://ugspace.ug.edu.gh Fig. 4.15(a): Multi-modality prostate phantom under scanning on ultrasound brachytherapy system Fig. 4.15(b): Multi-modality prostate phantom with needle inserts under scanning on ultrasound system 85 University of Ghana http://ugspace.ug.edu.gh Similar to the ultrasound system, images were acquired on the PET-CT system first without needles and subsequently with needles as shown in Figure 4.16. Activity solution of 2 mCi F-18 was introduced into the needles after pushing them through the prostate capsule. The needles were placed at the exact spots in the phantom as in the case of the ultrasound imaging procedure. The phantom was set up on the patient table and axial PET-CT images acquired from the base to apex at slice thickness of 5 mm, producing twelve image slices. Acquired images, as shown in chapter five (section 5.2) were also copied unto the computer system for contrast enhancement and fusion processes. Fig. 4.16: Multi-modality prostate phantom under scanning on PET-CT system 86 University of Ghana http://ugspace.ug.edu.gh 4.2.3 Development of MATLAB Algorithms Two MATLAB algorithms were developed in this study using MATLAB’s image processing toolbox. The first algorithm enhances contrast of images acquired from the prostate phantom and the second algorithm co-registers the ultrasound and PET-CT images together. 4.2.3.1 Contrast Enhancement Algorithm The MATLAB image contrast enhancement algorithm was developed by mapping the intensity values in raw images to new values in a modified image. The function imadjust was written as image processing tool for intensity transformations of grayscale images of MATLAB. The function was used to improve the contrast image and has the syntax below: Modified image = imadjust (Raw image, [low_in high_in], [low_out high_out], gamma). The full algorithm is presented in Appendix A1, and by running this algorithm in MATLAB, the resultant graphic user interface in Figure 5.3 is produced. The function imadjust maps the intensity values in raw image to new values in the modified image, such that the values between low_in and high_in map to values between low_out and high_out. Values below low_in and above high_in are clipped, hence values below low_in map to low_out and those above high_in map to high_out. The output intensity of images is reversed if the high-out is less than low-out. The parameter gamma specifies the shape of the curve that maps the intensity values in the raw image to create the modified image. If gamma is less than 1, the mapping weighs toward higher output values, making the image brighter. If gamma is greater than 1, the mapping weighs toward lower output values, making image darker. 87 University of Ghana http://ugspace.ug.edu.gh 4.2.3.2 Image Fusion Algorithm The MATLAB image fusion algorithm is presented in Appendix A2. Fusion of the images is based on the theory of mutual information and rigid body transformation, as expressed in Chapter 3. The fusion algorithm consists of four principal blocks. The first block is where multiple images are input into the algorithm. Second block is where the input images are converted into a common representational format such that they are compatible and able to communicate. Block three is where transformation matrices are established and images are fused together. A 22 transformation matrix was established between the ultrasound and PET-CT images. By setting the ultrasound image as reference and PET-CT image as floating, the PET-CT image was transposed unto the ultrasound image in the algorithm. Block four is where the fused image is displayed for viewing on screen. Running the algorithm in MATLAB produces the graphic user interface shown in Figure 5.4. 4.2.4 Estimation of Image Registration Error Error analysis was performed to estimate the level of accuracy of the fusion algorithm. This was done by establishing coordinates of the prostate capsule in the ultrasound and PET-CT images as well as the coordinates of the phantom’s housing in the two images. Deviations in the respective coordinates were estimated and root mean square error (RMSE) formalism used to calculate the entire image registration accuracy. The RMSE formalism is expressed in equation 4.7 and results for the error analysis are presented in Table 5.14. 1 n 2 RMSD  cUS  cPET CTi i  4.7 n i1 88 University of Ghana http://ugspace.ug.edu.gh CHAPTER FIVE RESULTS AND DISCUSSION 5.0 INTRODUCTION Results from the study and their analyses are presented in this chapter. The results are grouped into quality control tests of imaging devices, acquisition of prostatic images, development of MATLAB algorithms, and image registration error assessment. 5.1 PERFORMANCE EVALUATION RESULTS Results obtained from the QC tests on the ultrasound and PET-CT systems are presented under this section. Ultrasound brachytherapy system’s test results are presented in Section 5.1.1 and PET-CT system’s test results are presented in Section 5.1.2. 5.1.1 Ultrasound Brachytherapy System Quality control test results on the ultrasound brachytherapy system are presented in Table 5.1 – Table 5.5. For the purpose of this study, five tests were performed on the ultrasound brachytherapy system to assess its functionality and suitability to be used in this study. Results for axial distance measurement accuracy, lateral distance measurement accuracy, volume measurement accuracy, depth of penetration, and axial and lateral resolution tests are presented in Table 5.1, Table 5.2, Table 5.3, Table 5.4 and Table 5.5 respectively. Axial Distance Measurement Accuracy: Axial distances between targets along the B and F columns were measured from the ultrasound image using electronic caliper and compared with 89 University of Ghana http://ugspace.ug.edu.gh standard inter-target distances of 10 mm. Along column B, average inter-target distances between B1–B2, B2–B3, B3–B4 and B4–B5 were 10.12 ± 0.08 mm, 10.12 ± 0.13 mm, 10.11 ± 0.11 mm and 10.12 ± 0.10 mm respectively. The overall average inter-target distance for column B was estimated as 10.11 ± 0.10 mm, amounting to approximately 1.1% increase in the standard inter-target distance of 10 mm. Table 5.1: Axial distance measurement accuracy results Probe Measured Distance (mm) Position Column B Column F (cm) B1 - B2 B2 - B3 B3 - B4 B4 - B5 F1 - F2 F2 - F3 F3 - F4 F4 - F5 0.5 10.15 9.94 9.99 10.02 9.95 10.09 10.17 10.08 1.0 10.17 10.09 10.07 10.17 10.09 10.09 10.15 10.57 1.5 10.09 10.17 9.95 10.17 10.10 10.03 10.02 10.17 2.0 10.17 9.95 10.16 10.23 10.17 10.23 10.10 10.23 2.5 10.09 10.09 10.09 10.23 9.94 10.10 10.23 10.30 3.0 9.94 9.94 10.10 10.02 10.02 10.30 10.15 9.95 3.5 10.02 10.30 10.10 10.30 10.33 10.12 10.09 10.23 4.0 10.23 10.23 10.30 10.09 10.10 10.01 10.23 9.95 4.5 10.10 10.15 10.17 10.10 10.17 10.23 10.23 10.30 5.0 10.17 10.09 10.23 9.95 9.95 10.10 10.03 9.95 5.5 10.23 10.43 10.35 10.30 10.17 10.17 10.08 9.64 6.0 10.17 10.30 10.09 10.30 10.29 10.09 10.02 10.29 6.5 10.02 10.09 10.10 10.03 10.10 10.02 10.09 9.95 7.0 10.09 10.09 10.09 10.03 10.03 10.02 10.02 10.03 7.5 10.03 10.03 9.95 10.10 10.02 10.03 10.10 9.95 8.0 10.10 10.23 10.03 10.09 10.09 10.09 10.09 10.09 8.5 10.15 10.09 10.03 10.03 10.09 10.22 10.09 10.09 9.0 10.09 10.13 10.09 10.09 9.95 9.95 10.09 10.03 9.5 10.29 10.09 10.03 10.09 10.16 10.08 10.02 10.10 10.0 10.15 10.03 10.22 10.09 10.15 10.03 10.16 10.09 10.5 10.01 9.95 10.10 10.03 10.09 10.09 10.09 10.09 Average 10.12 10.12 10.11 10.12 10.09 10.10 10.11 10.10 Std Dev 0.08 0.13 0.11 0.10 0.10 0.09 0.07 0.19 Along F column, average inter-target distances between F1–F2, F2–F3, F3–F4 and F4–F5 were 10.09 ± 0.10 mm, 10.10 ± 0.09 mm, 10.11 ± 0.07 mm and 10.10 ± 0.19 mm respectively. The 90 University of Ghana http://ugspace.ug.edu.gh overall average inter-target distance for column F was estimated as 10.10 ± 0.12 mm, deviating by approximately 1.0% from the 10 mm standard inter-target distance. With tolerance level of ± 2 mm error margin recommended by AAPM TG 128 (Pfeiffer D. et al, 2008), it was deduced from the results that the ultrasound system produces a good level of accuracy in axial measurements. This result gives indication that for patient studies, the system can be relied upon to measure accurate axial distances between different brachytherapy seeds or between seeds and the rectal wall. Lateral Distance Measurement Accuracy: Lateral distances between targets along the rows 1, 2, 3, 4 and 5 were measured with electronic caliper from the ultrasound image and presented in Table 5.2. Measured distances were compared with known inter-target distances of 10 mm (for B2–C2 and E4–F4), 20 mm (for B3–D3 and D3–F3), 30 mm (for C2–F2 and B4–F4), and 40 mm (for B1–F1 and B5–F5). Estimated average distances for B2–C2 and E4–F4 were 10.07 ± 0.07 mm and 10.08 ± 0.07 mm, deviating from the standard inter-target distance of 10 mm by approximately 0.07 mm and 0.08 mm respectively. Average distances for B3–D3 and D3–F3 each deviated from the 20 mm standard inter-target distance by 0.01 mm. For targets C2–F2 and B4–F4, there were deviations of – 0.37 mm and – 0.55 mm respectively, between the measured and 30 mm standard inter- target distance. Each of the targets B1–F1 and B5–F5 also recorded deviations of – 0.46 mm from the standard inter-target distance of 40 mm. 91 University of Ghana http://ugspace.ug.edu.gh Table 5.2: Lateral distance measurement accuracy results Measured Distance (mm) Probe Position Row 1 Row 2 Row 3 Row 4 Row 5 (cm) B1 - F1 B2 - C2 C2 - F2 B3 - D3 D3 - F3 B4 - E4 E4 - F4 B5 - F5 0.5 39.73 10.10 29.33 20.03 19.95 29.60 10.03 39.73 1.0 39.93 10.02 30.00 20.03 20.01 29.80 10.02 39.93 1.5 39.80 10.08 29.80 20.10 19.96 29.33 10.01 39.80 2.0 39.00 10.10 29.67 19.90 20.05 29.00 10.10 39.00 2.5 39.33 10.02 29.80 19.90 20.03 30.00 10.08 39.33 3.0 39.33 10.10 29.33 20.03 20.00 29.33 10.10 39.33 3.5 40.00 10.00 29.00 20.10 19.98 29.33 10.23 40.00 4.0 40.00 10.08 29.40 19.90 19.90 30.00 10.08 40.00 4.5 39.80 10.10 30.00 20.03 19.97 29.33 10.02 39.80 5.0 39.27 10.30 29.32 20.03 19.97 29.67 10.09 39.27 5.5 39.07 10.08 29.63 20.10 20.10 29.40 10.08 39.07 6.0 39.33 10.10 30.03 20.03 19.99 30.00 10.02 39.33 6.5 40.00 10.02 29.70 20.10 20.01 29.67 10.08 40.00 7.0 39.00 10.02 29.50 19.90 20.12 29.33 10.11 39.00 7.5 39.00 10.02 29.33 19.97 20.02 30.00 10.30 39.00 8.0 39.67 10.08 29.54 19.97 20.01 29.33 10.02 39.67 8.5 40.00 10.08 29.53 20.03 20.00 29.33 10.10 40.00 9.0 39.33 10.10 29.60 20.10 19.99 30.00 10.05 39.33 9.5 40.00 10.02 30.02 20.10 19.98 29.80 10.08 40.00 10.0 39.33 10.02 29.97 19.97 20.03 29.00 10.10 39.33 10.5 39.33 10.02 29.70 19.97 20.04 29.53 10.04 39.33 Average 39.54 10.07 29.63 20.01 20.01 29.56 10.08 39.54 Std Dev 0.38 0.07 0.28 0.07 0.05 0.33 0.07 0.38 Deviations in all eight sets of lateral distance measurements fell within the tolerance level of ± 3 mm recommended by AAPM TG 128 (Pfeiffer D. et al, 2008). This indicates a good level of accuracy in the measurement of lateral distances within the prostate organ during brachytherapy procedures. Tolerance level for lateral distance accuracy is greater than axial accuracy due to reduced spatial resolution in the lateral direction, leading to increased uncertainty in the position of the target. 92 University of Ghana http://ugspace.ug.edu.gh Volume Measurement Accuracy: Important for real time dosimetry is the ability of an ultrasound brachytherapy system to accurately determine the volume of a target. Results for the three dimensional target volumes measured in the QA phantom are presented in Table 5.3. Table 5.3: Volume measurement accuracy results 3 Standard Measured Volume (cm ) Target Volume % 3 V (cm ) V1 V2 V3 V4 V5 Vave Std Dev Deviation 4.0 3.76 4.08 3.98 3.81 4.20 3.97 0.16 - 0.76 9.0 9.00 8.35 9.23 8.80 8.90 8.86 0.29 - 1.58 20.0 20.40 19.00 19.40 20.00 22.00 20.11 1.04 0.55 The volumes 4 cm3, 9 cm3 and 20 cm3 produced average measurements of 3.97 ± 0.16 cm3, 8.86 ± 0.29 cm3 and 20.11 ± 1.04 cm3, resulting in approximate deviations of – 0.76%, – 1.58% and 0.55% respectively. Percentage deviations of all three target volumes were within the tolerance of ± 5% specified by AAPM TG 128 (Pfeiffer D. et al, 2008). Depth of Penetration: Results for the depth of penetration is presented in Table 5.4. The test determines how deep into a patient a low contrast object can be visualized. Average depth of penetration estimated for the ultrasound system was 5.37 ± 0.04 cm. This indicates that a low contrast object or organ can be visualized approximately 5.4 cm into a patient on the ultrasound system. A change in signal to noise ratio (SNR) may have impact on the measurement. Clinically, a decrease in SNR causes a decrease in the maximum depth of penetration, which makes it difficult to visualize the anterior boundary of an imaged prostate gland. Comparison of results in this study with baseline figures was not possible due to unavailability of the baseline data during the time of performing this test. 93 University of Ghana http://ugspace.ug.edu.gh Table 5.4: Depth of penetration results Depth of penetration (cm) D1 D2 D3 D4 D5 Dave St. Dev Test 1 5.48 5.33 5.39 5.37 5.40 5.39 0.06 Test 2 5.35 5.29 5.40 5.35 5.38 5.35 0.04 Test 3 5.34 5.39 5.33 5.36 5.40 5.36 0.03 Overall average 5.37 ± 0.04 cm Axial and Lateral Resolution: Resolution of the ultrasound system determines its ability to distinguish between two closely spaced objects. The axial and lateral resolution tests were performed with the probe in the axial and longitudinal directions. Results for the tests are presented in Table 5.5. By scanning with the probe in axial direction, average axial and lateral resolution of 0.60 ± 0.07 mm and 1.89 ± 1.20 mm respectively were estimated. In the same probe direction, lateral resolution results in column B and F showed increase in the size of targets in the image as one moves further away from the probe (from row 1 to row 5). By turning the probe into the longitudinal array, axial resolution of 0.33 ± 0.13 mm was estimated while resolution of 1.76 ± 0.26 mm was estimated in the lateral direction. Tolerance level for this test could not be assessed due to the unavailability of baseline data to compare results with. Table 5.5: Axial and lateral resolution test results Axial Array Longitudinal Array Axial Lateral Axial Resolution (mm) Lateral Resolution (mm) Resolution Resolution Column B Column F Column B Column F (mm) (mm) Row 1 0.61 0.62 0.74 0.92 0.15 1.55 Row 2 0.64 0.41 0.74 1.16 0.30 1.70 Row 3 0.61 0.62 0.74 1.60 0.44 1.85 Row 4 0.62 0.62 3.05 3.19 0.46 2.16 Row 5 0.62 0.65 3.24 3.52 0.30 1.54 Ave 0.60 1.89 0.33 1.76 Std Dev 0.07 1.20 0.13 0.26 94 University of Ghana http://ugspace.ug.edu.gh While ultrasound guidance of prostate implants does not depend as critically on spatial resolution as do some diagnostic studies, certain aspects of the procedure can place a demand on the resolution of the system. Poor resolution makes it difficult to properly identify and locate implanted seeds, or the image of a needle may be too spread out to accurately register its location. Spatial resolution can be negatively impacted by poor probe condition and faulty pulse/receiver electronics. Results from the tests show a loss of lateral resolution with depth in the axial array direction due to the targets being out of the focal zone of the transducer. Axial resolution is seen to be constant with depth in axial array direction because axial resolution is primarily dependent upon the frequency of the transducer. 5.1.2 PET-CT System Quality control test results of the PET-CT system are presented in tables 5.6 – 5.14. Satisfactory results from the tests ensured suitability of the system to be used in the research study. Results for the tests include CT laser alignment, bed position accuracy, CT display width, CT image uniformity, CT dose index, PET-CT image uniformity, PET-CT image registration accuracy, PET-CT image contrast, and PET-CT image resolution. CT Laser Alignment: Laser alignment test was performed to ensure that the gantry lasers were properly aligned with CT gantry and bed. Results for the test are presented in Table 5.6. All the CT lasers (in the x, y and z planes) perfectly coincided with the phantom scribe marks, indicating correct alignment of the CT lasers, within ± 1 mm, as recommended in IAEA HSS No.1 (IAEA HHS 1, 2009). 95 University of Ghana http://ugspace.ug.edu.gh Table 5.6: CT laser alignment test results x-plane y-plane z-plane CT Laser Deviation (mm) 0.0 0.0 0.0 [from phantom scribe marks] Results (Pass / Fail )    Bed Position Accuracy: The patient bed position accuracy test was performed to ensure accurate and reproducible motion of the bed in horizontal and vertical directions under scanner control. The test produced a good level of accuracy in table positioning for all directions of movement. Table positioning in  y ,  y ,  z and  z directions were all within the recommended tolerance level of ± 1 mm. The results gave indication of accurate bed movements under manual and computer control in clinical situations of patient set-up during imaging. Table 5.7: Table (bed) position accuracy test results Manual Set-Up Computer Controlled Set-Up Table Displacement ( ) /cm 1.0 10.0 50.0 100.0 1.0 10.0 50.0 100.0 Measured ( ) /cm 1.0 10.0 50.0 100.1 1.0 10.0 50.0 99.9 Deviation (  ) /cm 0.0 0.0 0.0 +0.1 0.0 0.0 0.0 -0.1 Table Displacement ( ) /cm -1.0 -10.0 -50.0 -100.0 -1.0 -10.0 -50.0 -100.0 Measured ( ) /cm -1.0 -10.0 -50.0 -99.9 -1.0 -10.0 -50.0 -100.0 Deviation (  ) /cm 0.0 0.0 0.0 +0.1 0.0 0.0 0.0 0.0 Table Displacement ( ) /cm 1.0 10.0 20.0 30.0 1.0 10.0 20.0 30.0 Measured ( ) /cm 1.0 10.0 20.0 30.1 1.0 10.0 20.0 30.0 Deviation (  ) /cm 0.0 0.0 0.0 +0.1 0.0 0.0 0.0 0.0 Table Displacement ( ) /cm -1.0 -10.0 -20.0 -30.0 -1.0 -10.0 -20.0 -30.0 Measured ( ) /cm -1.0 -10.0 -20.0 -30.0 -1.0 -10.0 -20.0 -30.1 Deviation (  ) /cm 0.0 0.0 0.0 0.0 0.0 0.0 0.0 -0.1 96 Bed Movement y-plane z-plane -y +y -z +z University of Ghana http://ugspace.ug.edu.gh CT Display Width: Performance of this test was to ensure that volume of a patient or organ being measured and displayed is equivalent to that selected on the CT scanner console. As shown in Table 5.8, the measured widths of 4 cm and 20.5 cm were compared with image displayed widths of 4 cm and 20.4 cm, producing deviations of 0.0 cm and 0.1 cm respectively. Deviation for the orthogonal section was equal to the set tolerance level of ± 1 mm. The results therefore gave credibility to volume measurements made with the CT scanner on patients under scanning. Table 5.8: CT Display width test results Measured width Image Displayed Deviation (cm) m (cm) width d (cm) m d Circular plane 4.0 4.0 0.0 Octagonal plane 20.5 20.4 0.1 CT Image Uniformity: Uniformity of CT image was assessed to ensure uniform distribution of CT numbers over an acquired homogeneous image. Results from the CT image uniformity test in Table 5.9 indicates that mean CT numbers in the four peripheral ROIs deviated from the central mean to within recommended tolerance level of ± 5 HU. This indicates a good level of uniformity in the CT images. By visual inspection of the acquired image (Figure 4.10), there was no observed artifact which could affect the uniformity. Noise level in the image was estimated as the standard deviation of the CT numbers divided by the average of the CT numbers in each area. The top and left regions of the image produced relatively higher noise levels compared with the central and bottom regions. These noise levels were however not enough to cause offset non uniformity in the image. 97 University of Ghana http://ugspace.ug.edu.gh Table 5.9: CT uniformity and image noise test results Region of Interest ROI-3 ROI-4 ROI-5 ROI-6 ROI-7 (Central) (Top) (Right) (Bottom) (Left) Minimum -57.0 -47.0 -47.0 -42.0 -36.0 Maximum 63.0 39.0 41.0 41.0 38.0 Mean ( x ) -2.0 -0.5 -0.7 -0.9 0.5 Deviation from central  2.0  (2.0) 0.5 (2.0)  0.7  (2.0)  0.9  (2.0) 0.5  (2.0) mean x  x ; x  2.0  0.0 1.5 1.3 1.1  2.5 Noise SD  x 0.84 3.38 2.41 1.88 3.38 2 x  xm  SD  1.688 N 1 Computed Tomographic Dose Index: Dose index (CTDI) assessment was performed to ensure appropriateness of radiation doses delivered in CT examinations. Results for the head and body CT tests are presented in Table 5.10(a) and Table 5.10(b) respectively. Table 5.10(a): CTDI test results for head phantom at 120 kVp and 150 mAs Charge Exposure CTDI100 CT console Q1 (nC) Q2 (nC) Qave (nC) X (C/kg) X (rad) (rad) CTDIvol (mGy) Central (C) 0.3080 0.3080 0.3080 1.90E-03 5.7338 3.5836 42.40 Periphery (P1) 0.3780 0.3790 0.3785 2.33E-03 7.0462 4.4039 42.40 Periphery (P2) 0.3810 0.3820 0.3815 2.35E-03 7.1020 4.4388 42.40 Periphery (P3) 0.3840 0.3840 0.3840 2.36E-03 7.1486 4.4679 42.40 Periphery (P4) 0.3830 0.3830 0.3830 2.36E-03 7.1300 4.4562 42.40 Estimated CT console CTDI100c (rad) CTDI100p (rad) CTDIw (mGy) CTDIvol (mGy) CTDIvol (mGy) 3.5836 4.4417 41.5566 44.30 42.40 Percentage Deviation 4.49 % 98 CT Number (HU) University of Ghana http://ugspace.ug.edu.gh Scan protocols for the head phantom study were set at brain scan CT techniques of 120 kVp and 150 mAs to mimic clinical conditions for an adult patient. The estimated dose from measurements taken in this study was 44.3 mGy and the corresponding console displayed dose was 42.4 mGy. A deviation of 4.5 % was realized between the estimated and console displayed doses. Estimated dose from measurements taken in the body phantom was 20.08 mGy, and this was at pelvic scan CT techniques of 120 kVp and 100 mAs. The console displayed dose for the pelvic examination was 19.49 mGy, hence deviation of 3.1 % was realized between the estimated and console doses. Table 5.10(b): CTDI test results for body phantom at 120 kVp and 100 mAs Charge Exposure CTDI100 CT console Q1 (nC) Q2 (nC) Qave (nC) X (C/kg) X (rad) (rad) CTDIvol (mGy) Central (C) 0.1330 0.1340 0.1335 8.22E-04 2.4853 1.5533 19.49 Periphery (P1) 0.1770 0.1760 0.1765 1.09E-03 3.2857 2.0536 19.49 Periphery (P2) 0.1770 0.1770 0.1770 1.09E-03 3.2951 2.0594 19.49 Periphery (P3) 0.1750 0.1750 0.1750 1.08E-03 3.2578 2.0361 19.49 Periphery (P4) 0.1760 0.1760 0.1760 1.08E-03 3.2764 2.0478 19.49 Estimated CT console CTDI100c (rad) CTDI100p (rad) CTDIw (mGy) CTDIvol (mGy) CTDIvol (mGy) 1.5533 2.0492 18.8391 20.08 19.49 Percentage Deviation 3.05 % The estimated deviations for the head and body phantom studies signify a very good measure of dose output from the CT scanner, in comparison with the tolerance level of ± 20% specified by the Human Health Series No. 1 (IAEA HHS 1, 2009). The results gave credibility to displayed console doses as during CT examinations. 99 University of Ghana http://ugspace.ug.edu.gh PET-CT Image Uniformity: Uniformity estimate for PET-CT image was performed to assess the system’s response to a homogeneous activity distribution. Table 5.11(a) and Table 5.11(b) present results for the image uniformity test. The uniformity in PET-CT image was assessed by estimating the quantitative index of non-uniformity and coefficient of uniformity variation in the 1010 mm2 ROIs across the mid-transaxial slice of the cylindrical phantom. Table 5.11(a): PET-CT image uniformity test results Region of Counts of Activity ( Ck ) Non Uniformity (% NU) Max NU (%) Interest 100 MaxCk AveCk  100 AveCk MinCk Max Ave Min AveCk  AveCk  ROI 1 107 88.240 75 21.260 15.005 21.260 ROI 2 100 89.277 76 12.011 14.872 14.872 ROI 3 105 87.685 66 19.747 24.731 24.731 ROI 4 99 88.373 77 12.025 12.869 12.869 ROI 5 100 87.610 78 14.142 10.969 14.142 ROI 6 104 89.630 75 16.033 16.323 16.323 ROI 7 101 88.514 78 14.106 11.878 14.106 ROI 8 101 89.710 78 12.585 13.053 13.053 … … … … … … … ROI 155 119 92.535 59 28.600 36.240 36.240 … … … … … … … ROI 177 107 87.744 72 21.946 17.943 21.946 Maximum NU (%) 36.240 From the results, maximum non-uniformity was estimated as 36.24% in ROI 155 (with red indication in Figure 4.12b) and the coefficient of variation was 5.12%. The coefficient of variation gives the extent of uniformity variation over the entire axial image slice. This quantitative assessment of uniformity provides the true state of the activity distribution in the homogeneous medium, which would otherwise be difficult to predict qualitatively. Reference non-uniformity for this test was not available at the time of this study and hence measured non- 100 University of Ghana http://ugspace.ug.edu.gh uniformity could not be compared as HHS 1 recommends. Tolerance for this uniformity test was set as %NUmeasured < 1.05%NUreference (IAEA HHS 1, 2009). Table 5.11(b): Coefficient of variation for non-uniformity Counts of Activity ( C )  2Region of Interest k Ck  AveCk  ROI 1 107 1.545 ROI 2 100 33.143 ROI 3 105 0.573 ROI 4 99 45.657 ROI 5 100 33.143 ROI 6 104 3.087 ROI 7 101 22.629 ROI 8 101 22.629 … … … ROI 155 119 175.377 … … … ROI 177 107 1.545 AveC  105.757 k  5154.554 1 2 SD  C  AveC 177 1 k k  SD  5.4112  CV 100  SD AVE C  CV  5.117% k PET-CT Image Registration Accuracy: With the fusion of PET and CT images assuming correct registration, it was important to ensure that the two studies are perfectly registered for a given patient study. Quantitative assessment of the accuracy of PET and CT image registration was performed in this study by using lead bricks to assume a standard 80 kg patient weight. Results from the study indicated that the two images were registered to displacements of less than 1 mm in x, y and z directions. PET and CT images of the spheres containing the F-18 solution coincided to a high degree, similar to the edges of the image quality phantom. Figure 101 University of Ghana http://ugspace.ug.edu.gh 4.13(b) in chapter 4 shows that the centres of all the spheres in the image quality phantom perfectly coincided and Table 5.12 presents measured displacements in all three directions (axial and transaxial). The edges of the phantom were also analyzed for displacements in the PET-CT image and observed to be less than 1 mm. In relation to the recommended tolerance by HHS 1 (IAEA HHS 1, 2009) of ± 1 mm, the images were established to be accurately registered. Table 5.12: PET-CT image registration accuracy Average displacement between PET and CT registered images (mm) x-direction y-direction z-direction Spheres 0.52 0.44 0.64 Edges of phantom 0.60 0.48 0.66 PET-CT Image Quality: The test was performed to produce images simulating those obtained in a total body imaging study involving both hot and cold lesions. Image quality was assessed by estimating image contrast and background variability ratios for both hot and cold spheres, as presented in Table 5.13(a) and 5.13(b) respectively. Percentage contrast between the hot spheres of diameters 1.3 cm and 2.2 cm, and the background were estimated as 49.3% and 52.6% respectively. For the cold spheres of diameters 2.8 cm and 3.7 cm, percentage contrast of 74.8% and 75.6% were estimated. Contrast for the empty sphere was estimated to be 89.0%. Between the hot spheres and the cold spheres, it was observed that for equal concentrations, the bigger the volume, the higher the contrast. The purpose of this study follows closely the recommendations of NEMA NU2-2007 to assess contrast of PET-CT images of patients (NEMA, 2007). 102 University of Ghana http://ugspace.ug.edu.gh Table 5.13(a): Percentage contrast test results Hot Cold Empty Sphere diameter 1.3 cm 2.2 cm 2.8 cm 3.7 cm 1.7 cm Counts of activity in 325 332 19 19 8 sphere (cts) ROI 1 68 73 75 82 69 ROI 2 64 65 66 72 64 ROI 3 62 62 64 64 62 ROI 4 60 60 62 62 60 ROI 5 58 58 60 61 58 ROI 6 57 57 58 58 57 ROI 7 79 82 83 86 80 ROI 8 91 92 93 94 91 ROI 9 94 94 95 96 94 ROI 10 92 93 94 95 93 ROI 11 74 80 85 93 77 ROI 12 65 68 69 71 65 Average ( CB,n ) 72.083 73.667 75.333 77.833 72.500 Percentage Contrast (%) 49.307 52.637 74.779 75.589 88.966 Background variability was estimated by drawing twelve regions of interest of sizes equal to the spheres in the phantom. The estimated percentage background variability for the ROIs of diameters 1.3 cm, 2.2 cm, 2.8 cm, 3.7 cm and 1.7 cm were 19.3%, 19.1%, 18.6%, 19.0% and 19.3% respectively. The results, as presented in Table 5.13(b), suggest an approximate background variability of 19% in the image. The background variability estimation allows assessment of the accuracy of the absolute quantification of radioactivity concentration in the uniform volume of interest inside the phantom. A 5% tolerance criterion with respect to baseline values of percentage contrast and background variability could not be estimated because baseline values were not readily available. 103 Background Counts (cts) University of Ghana http://ugspace.ug.edu.gh Table 5.13(b): Background variability test results ROI diameter 1.3 cm 2.2 cm 2.8 cm 3.7 cm 1.7 cm ROI 1 16.671 0.445 0.111 17.364 12.250 ROI 2 65.335 75.117 87.105 34.024 72.250 ROI 3 101.667 136.119 128.437 191.352 110.250 ROI 4 145.999 186.787 177.769 250.684 156.250 ROI 5 198.331 245.455 235.101 283.350 210.250 ROI 6 227.497 277.789 300.433 393.348 240.250 ROI 7 47.845 69.439 58.783 66.700 56.250 ROI 8 357.853 336.099 312.123 261.372 342.250 ROI 9 525.189 413.431 386.791 330.040 462.250 ROI 10 396.687 373.765 348.457 294.706 420.250 ROI 11 3.675 40.107 93.451 230.038 20.250 ROI 12 50.169 32.115 40.107 46.690 56.250 1 K  CB,n,k CB,n  2 K 1 N %100 k1 19.336 19.139 18.639 18.976 19.324  CB,n PET-CT System Resolution: Resolution test was performed to assess the smallest tumour size the PET-CT system could pick up during patient imaging. The resolution test is based on the NEMA NU2-2007 spatial resolution protocol (NEMA, 2007). Results for the resolution test are presented in Table 5.14. Full width at half maximum estimate for the least resolved rods (in sector 1) signified the resolution of the PET-CT system, and hence the size of the lesion that can be detected. Resolution of 0.5 ± 0.01 cm was estimated for the system. This indicates that tumours with sizes of about 5 mm could be picked up by the PET-CT system during imaging. Tolerance for this study was set according to HHS 1 (IAEA HHS 1, 2009) as Rmeasured < 1.05Rexpected, however, there was no reference resolution value available to compare the measured value with. 104 Difference between ROI counts and average counts in the background University of Ghana http://ugspace.ug.edu.gh Table 5.14: PET-CT system resolution test results FWHM (cm) Sector 1 Sector 2 Sector 3 Sector 4 Sector 5 Sector 6 Rod 1 0.52 0.61 0.70 0.79 0.90 0.99 Rod 2 0.49 0.60 0.69 0.79 0.89 1.00 Rod 3 0.50 0.59 0.69 0.81 0.90 1.00 Rod 4 0.51 0.59 0.70 0.80 0.91 1.02 Rod 5 0.50 0.61 0.71 0.80 0.89 1.01 Rod 6 0.50 0.61 0.71 0.81 0.90 1.00 Rod 7 0.49 0.61 0.72 0.82 0.90 0.99 Rod 8 0.51 0.62 0.71 0.82 0.92 1.00 Average 0.50 0.61 0.70 0.81 0.90 1.00 St Dev 0.01 0.01 0.01 0.01 0.01 0.01 5.2 ACQUIRED PROSTATIC IMAGES The prostatic images acquired on the ultrasound and PET-CT systems (with needle inserts) are presented in Figure 5.1 and Figure 5.2 respectively. The four needles were inserted into the prostate capsule to serve as points of reference for the fusing procedure. The ultrasound system produced an image of the prostate capsule which is well visualized and very well defined, however, the implanted needles were not well defined due to the appearance of streak artifacts around it. The artifacts were produced as a result of reflection of greater percentage of the sound waves as they encountered the metallic objects. In such situation, performing brachytherapy treatment planning with ultrasound images alone could possibly introduce some degree of errors. The planning system for prostate brachytherapy is dependent on shapes of implanted seed sources and assumes isotropic distribution of radiation dose around them. Hence a number of researchers have proposed the need for the addition of other imaging modalities which could 105 University of Ghana http://ugspace.ug.edu.gh produce images of clearly defined seed sources to improve outcomes of the treatment planning process. Fig. 5.1: Ultrasound image of prostate phantom On the PET-CT system, an image of the prostate capsule was produced but not clearly defined as it appeared in the ultrasound image, however, the implanted needles were very well visualized and defined. Prostate gland has density similar to that of its surrounding tissues and hence contrast between them in an x-ray CT is poor due their similar attenuation properties. Streaking was also highly minimized in the image due to the use of high level attenuation correction reconstruction algorithm in the CT system. Fluorine-18 solution was introduced into the needles inserts in order for PET imaging to be possible since the prostate phantom in itself is not radioactive and hence cannot be imaged on a PET system. The F-18 produced hotspots in the 106 University of Ghana http://ugspace.ug.edu.gh PET-CT image, simulating uptake of activity in a prostate. At the centre of the prostate gland is seen the urethra through which urine is passed from the bladder into the penis. Fig. 5.2: PET-CT image of prostate phantom 5.3 MATLAB ALGORITHMS 5.3.1 Contrast Enhancement Algorithm The graphic user interface arising out of the contrast enhancement algorithm in Appendix A1 is shown in Figure 5.3. Contrast of an image is enhanced by first importing it into the “Original Image” space using the “Acquire Image” button. By clicking the “Modify Image” button, contrast of the image changes in respect of the parameter settings in the Low-In, Low-Out, High- 107 University of Ghana http://ugspace.ug.edu.gh In, High-Out and Gamma columns. Treatment planning in prostate brachytherapy is dependent on contrast of images used, hence the need to improve contrast before images are employed to plan treatment. Fig. 5.3: Graphic user interface for contrast enhancement algorithm 5.3.2 Image Fusion Algorithm The graphic user interface resulting from the image fusion algorithm in Appendix A2 is shown in Figure 5.4. Images are loaded into space using “Load Image A” and “Load Image B” buttons. 108 University of Ghana http://ugspace.ug.edu.gh The loaded images are fused by clicking on “Fuse Now!”, and the fused image appears in the “Fused Image” space. The fused image from ultrasound and PET-CT is observed to have a good visualization and well defined prostate capsule, urethra as well as implanted seeds. This resultant image as shown in RGB (Figure 5.5(a)) and grayscale (Figure 5.5(b)) could produce more accurate treatment planning results, leading to better treatment outcomes. Fig. 5.4: Graphic user interface for image fusion algorithm 109 University of Ghana http://ugspace.ug.edu.gh Fig. 5.5(a): Fused US-PET-CT image in RGB display 110 University of Ghana http://ugspace.ug.edu.gh Fig. 5.5(b): Fused US-PET-CT image in grayscale display 5.4 ERROR ASSESSMENT Root mean square error (RMSE) formalism was used to assess the registration error of the US-PET-CT fused image. From Figure 5.6, the RMSE was estimated to be 1.3 mm, as shown in Table 5.14. The fusion accuracy level of this study could therefore be considered appreciable in 111 University of Ghana http://ugspace.ug.edu.gh relation to studies from Holupka et al (1996) and Steggerda et al (2005), who both registered CT- US images to accuracy levels of 2 mm and less than 1 mm respectively. The estimated accuracy level is also comparable to the study of Huber J.S. et al (2011) who recorded accuracies of 2.1 ± 1.7 mm, 1.9 ± 1.6 mm and 0.6 ± 0.2 mm in the x, y and z-directions in their PET-TRUS image registration. Fig. 5.6: Coordinates for image registration error assessment 112 University of Ghana http://ugspace.ug.edu.gh Table 5.14: Image registration error assessment results Coordinates Ultrasound Image PET-CT Image Deviation 2 (US – PCT) (cm) (US) (PCT) (US – PCT) X1 1.1821 1.1712 0.0109 0.0001 X2 4.2453 4.5276 -0.2823 0.0797 Y1 5.1837 5.1537 0.0300 0.0009 Y2 9.1210 9.2106 -0.0896 0.0080 X1 0.1133 0.1211 0.0078 0.0001 X2 5.6279 5.8318 -0.2039 0.0416 Y1 4.9742 4.9634 0.0108 0.0001 Y2 12.3333 12.3913 0.0580 0.0034  0.1339  n 0.0167 RMSD (cm) 0.1294 113 Phantom Prostate Housing Capsule University of Ghana http://ugspace.ug.edu.gh CHAPTER SIX CONCLUSION AND RECOMMENDATION 6.0 INTRODUCTION Fusion of medical images between different cross-sectional modalities has become a critical step in achieving improved diagnosis and treatment of a large number of diseases. Even more important is when anatomical images are fused with physiological information to manage medical conditions such as cancers. This chapter presents summary of the findings from this research study on image fusion and its associated recommendations. 6.1 CONCLUSION The purpose of this dissertation was to enhance diagnosis and treatment of prostate cancers by developing algorithms to fuse ultrasound and PET-CT images. The principal motivation for the image fusion approach was to improve the quality of the information contained in the output image from the individual input images. While there exists a number of computer programs for image fusion, MATLAB’s image processing toolbox provides user friendly capabilities for developing fused images with reduced uncertainty, increased reliability, robust system performance, and compact representation of information. In achieving the objective of this study, several tasks were addressed, which include, performance evaluation of ultrasound and PET-CT systems, development of image contrast enhancement algorithm, development of image fusion algorithm, and assessment of image fusion accuracy. 114 University of Ghana http://ugspace.ug.edu.gh Performance evaluation tests on the ultrasound brachytherapy system included axial distance measurement, lateral distance measurement, volume measurement, depth of penetration, and axial and lateral resolution. Except for depth of penetration and resolution tests in which baseline data were not available for comparison, all other quality control tests on the ultrasound system fell within set tolerances as recommended by the American Association of Physicists in Medicine Task Group Report 128 (Pfeiffer D. et al, 2008). Evaluation of the performance of the PET-CT system was done in accordance with recommendations of IAEA Human Health Series No. 1 (IAEA HHS 1, 2009). Quality control tests performed on the system included CT laser alignment, bed position accuracy, CT display width, CT image uniformity, CT dose index, PET- CT image uniformity, PET-CT image registration accuracy, PET-CT image contrast, and PET- CT image resolution. Results for all the tests fell within recommended tolerances, except for PET-CT image uniformity, image quality and system resolution, in which baseline data were not available for comparison. The results however indicated satisfactory performance of the PET-CT system based on analysis performed. After deriving satisfactory results from the performance evaluation, the ultrasound and PET-CT systems were used to acquire prostatic images with and without needle insertions. MATLAB image enhancement algorithm was developed to enhance the quality of prostatic images before fusion. The algorithm was developed by mapping the intensity values in raw images to new values in a modified image using imadjust function. By running the algorithm in MATLAB, a graphic user interface which enables contrast adjustment was produced. The contrast enhanced images of ultrasound and PET-CT were then co-registered with developed MATLAB fusion algorithm. The fusion algorithm was developed on the theory of mutual information and rigid body transformation. By running the fusion algorithm in MATLAB, 115 University of Ghana http://ugspace.ug.edu.gh graphic user interface is produced which allows fusion to be done. Ultrasound and PET-CT images are loaded into space using the load buttons, and subsequently fused by pressing the fuse button. Fused image of ultrasound and PET-CT in this study is observed to have well defined and good visualized prostate capsule, urethra and implanted seeds, which would otherwise not be the case in either of the two images separately. The resultant image could therefore produce better treatment outcomes if employed in treatment planning of prostate cancer cases. Assessment of the image registration error produced a root mean square error estimate of 1.3 mm. This estimate indicates good level of registration accuracy as in the studies of Holupka et al (1996), Steggerda et al (2005) and Huber et al (2011). 6.2 RECOMMENDATIONS This study has demonstrated the importance of fusing ultrasound and PET-CT images to improve prostate brachytherapy treatment outcomes. The prostate phantom used in the study is compatible on ultrasound, CT and MRI, but not compatible on PET due to absence of radionuclide activity in it. With the outcome of the research presented in this dissertation, the following recommendations are made to: (i) Authority of Medical Imaging Centres Quality control and quality assurance checks on medical imaging equipment should be performed on schedule and in accordance with recommended timelines. (ii) To Research Team Further study should be carried out to validate the findings on real patient studies, by fusing ultrasound and PET-CT images of patients using the developed algorithm. 116 University of Ghana http://ugspace.ug.edu.gh (iii) To Scientific Community and Phantom Manufacturers Study should be conducted to develop a prostate phantom which is compatible on both anatomical imaging modalities (ultrasound, CT and MRI) and physiological imaging modality (PET). 117 University of Ghana http://ugspace.ug.edu.gh REFERENCES American Association of Physicists in Medicine (AAPM). The measurement, reporting and management of radiation dose in CT: Report 96 (TG 23). (2008). USA, ISSN:0271- 7344. American Cancer Society (ACS). Cancer Facts & Figures 2012. Atlanta, USA. (2012). Accessed on January 21 2013 at www.cancer.org/acs/groups/cid/documents/ webcontent/003134- pdf.pdf. American Society of Clinical Oncology, "Five Things Physicians and Patients Should Question", Choosing Wisely: an initiative of the ABIM Foundation (American Society of Clinical Oncology), http://choosingwisely.org/wp-content/uploads/2012/04/5things_12_factsheet_ Amer_Soc_Clin_Onc.pdf, retrieved August 14 2012 Andronache, A., von Siebenthal, M., Szekely, G., Cattin, P.: Non-rigid registration of multi- modal images using both mutual information and cross-correlation. Med. Imag. Anal. 12, 3–15 (2006). Ash D., Flynn A., Battermann J., de Reijke T., et al. ESTRO/EAU/EORTC recommendations on permanent seed implantation for localized prostate cancer. Radioth Oncol 2000; 57: 315- 321. Baade, PD; Youlden, DR; Krnjacki, LJ (2009 Feb). "International epidemiology of prostate cancer: geographical distribution and secular trends.". Molecular nutrition & food research 53 (2): 171–84. PMID 19101947. Behre HM, Bohmeyer J, Nieschlag E. Prostate volume in testosterone-treated and untreated hypogonadal men in comparison to age-matched controls. Clin Endocrinol 1994;40:341- 9. Beyersdorff D, Taymoorian K, Knösel T, Schnorr D, Felix R, Hamm B, Bruhn H. MRI of Prostate Cancer at 1.5 and 3.0 T: Comparison of Image Quality in Tumor Detection and Staging. AJR 2005; 185:1214-1220. Billis A, Guimaraes MS, Freitas LL, et al. The impact of the 2005 international society of urological pathology consensus conference on standard Gleason grading of prostatic carcinoma in needle biopsies. J Urol 2008;180(2):548-52; discussion 552-3. http://www.ncbi.nlm.nih.gov/pubmed/18550106. BMJ Group (8 December 2009). "Prostate cancer: How far has your cancer spread? The TNM system". London: Guardian.co.uk. Retrieved 5 September 2013 at 118 University of Ghana http://ugspace.ug.edu.gh http://www.guardian.co.uk/lifeandstyle/besttreatments/prostate-cancer-how-far-has-your- cancer-spread-the-tnm-system.. Bouchelouche K., Capala J., and Oehr P., “Positron emission tomography/computed tomography and radioimmunotherapy of prostate cancer,” Current Opinion in Oncology, vol. 21, no. 5, pp. 469–474, 2009. Buron C, Le Vu B, Cosset JM et al. Brachytherapy versus prostatectomy in localized prostate cancer: Results of a French multicenter prospective medico-economic study. International Journal of Radiation Oncology Biology Physics 2007;67:812–822. Bushberg JT, Seibert JA, Leidholdt EM, Boone JM. The essential physics of medical imaging, 2nd Edition, Lilppincott Williams and Wilkins, Philadelphia –USA, Pp 469 – 553. Cheuck L. Brachytherapy (Radioactive Seed Implantation Therapy) in Prostate Cancer. Chiche A., Kanikowski M., Skowronek J., et al. Correlation between treatment plan parameters and particular prognostic factors in prostate cancer treated with high-dose-rate brachytherapy (HDR-BT) as a boost. J Contemp Brachytherapy 2009; 1: 11-17. Christensson A, Laurell CB, Lilja H. Enzymatic activity of prostate-specific antigen and its reactions with extracelluar serine proteinase inhibitors. Eur J Biochem 1990;194:755-63. http://www.ncbi.nlm.nih.gov/pubmed/1702714. Cimitan M, Bortolus R, Morassut S, Canzonieri V, Garbeglio A, Baresic T, Borsatti E, Drigo A, Trovo MG. [18F] fluorocholine PET/CT imaging for the detection of recurrent prostate cancer at PSA relapse: experience in 100 consecutive patients. Eur J Nucl Med Mol Imaging 2006;33(12):1387-98. College of American Pathologists, Prostate Cancer – Prostate Adenocarcinoma. Accessed at http://www.cap.org/apps/docs/reference/myBiopsy/ProstateAdenocarcinoma.pdf on 23 September 2013. Cornud, F. et al. endorectal color Doppler sonography and endorectal Mr imaging features of nonpalpable prostate cancer: correlation with radical prostatectomy findings. AJR Am. J. Roentgenol. 175, 1161–1168 (2000). Crook J, Fleshner N, Roberts C, Pond G. Long-term urinary sequelae following 125-Iodine prostate brachytherapy. The Journal of Urology 2008;179:141–146. Crook J.M., Gomez-Iturriaga A., Wallace K., et al. Comparison of health-related quality of life 5 years after SPIRIT: Surgical Prostatectomy Versus Interstitial Radiation Intervention Trial. J Clin Oncol 2011; 29: 362-368. Daily Graphic, Press release, Ghana, 13/08/2007, Page 43. 119 University of Ghana http://ugspace.ug.edu.gh Dehdashti, F. et al. Positron tomographic assessment of androgen receptors in prostatic carcinoma. Eur. J. Nucl. Med. Mol. Imaging 32, 344–350 (2005). Demanes D.J., Rodriguez R.R., Altieri G.A. High dose rate prostate brachytherapy: the California Endocurietherapy (CET) Method. Radiother Oncol 2000; 57: 289-296. Demanes D.J., Rodriguez R.R., Schour L., Brandt D., Altieri G. High-dose-rate intensity- modulated brachytherapy with external beam radiotherapy for prostate cancer: California endocurietherapy’s 10-year results. Int. J. of Radiat. Oncol Biol and Phys 2005, 61: 1306- 1316. Dimitropoulou, P., Lophatananon, A., Easton, D., Pocock, R., Dearnaley, D. P., Guy, M., Edwards, S., O’Brien, L., Hall, A., Wilkinson, R., The UK Genetic Prostate Cancer Study Collaborators, British Association of Urological Surgeons Section of Oncology, Eeles, R. and Muir, K. R. (2009), Sexual activity and prostate cancer risk in men diagnosed at a younger age. BJU International, 103: 178–185. doi: 10.1111/j.1464-410X.2008.08030.x Djulbegovic M, Beyth RJ, Neuberger MM, Stoffs TL, Vieweg J, Djulbegovic B, Dahm P (2010). "Screening for prostate cancer: systematic review and meta-analysis of randomised controlled trials". BMJ 341: c4543. doi:10.1136/bmj.c4543. PMC 2939952. PMID 20843937. //www.ncbi.nlm.nih.gov/pmc/articles/PMC2939952/. Ewertsen C, Nielsen MB, Torp-Pedersen S. Image fusion between ultrasonography and CT, MRI or PET/CT for image guidance and intervention – a theoretical and clinical study. Dan Med Bull 2010;57:(9) B4172. Farsad, M. et al. Detection and localization of prostate cancer: correlation of 11C-choline PET/ CT with histopathologic step-section analysis. J. Nucl. Med. 46, 1642–1649 (2005). Ferrer M., Suarez J.F., Guedea F., et al. Health-related quality of life 2 years after treatment with radical prostatectomy, prostate brachytherapy, or external beam radiotherapy in patients with clinically localized prostate cancer. Int J Radiat Oncol Biol Phys 2008; 72: 421-432. Fricke, E. et al. Positron emission tomography with 11C-acetate and 18F-FDG in prostate cancer patients. Eur. J. Nucl. Med. Mol. Imaging 30, 607–611 (2003). Fuller DB, Jin H, Koziol JA, Feng AC. CT-ultrasound fusion prostate brachytherapy: a dynamic dosimetry feedback and improvement method. A report of 54 consecutive cases. Brachytherapy 2005;4(3):207-16. Gaa J, Rummeny EJ, Seemann MD. Whole-body imaging with PET/MRI. Eur J Med Res. 2004;9: 309-312. 120 University of Ghana http://ugspace.ug.edu.gh Galalae R.M., Kovacs G., Schultze J., et al. Long term outcome after elective irradiation of the pelvic lymphatics and local dose escalation using high-dose-rate brachytherapy for locally advanced prostate cancer. Int J Radiat Oncol Biol Phys 2002; 52: 81-90. Gang H., Image fusion, image registration and radiometric normalization for high resolution image processing. PhD thesis, University of New Brunswick, Canada (2007). Garbaulet A, Potter R, Mazeron JJ et al. The GEC ESTRO Handbook of Brachytherapy. Brussels 2002; chapter 20, pp. 473-480. Giles, G.G., Severi, G., English, D.R., McCredie, M.R.E., Borland, R., Boyle, P. and Hopper, J.L. (2003), Sexual factors and prostate cancer. BJU International, 92: 211–216. doi: 10.1046/j.1464-410X.2003.04319.x Gleason DF. Histologic grading of prostate cancer: a perspective. Human Pathol 1992;23(3):273–279. http://www.ncbi.nlm.nih.gov/pubmed/1555838. Grimm P., Billiet I., Bostwick D., et al. Comparative analysis of prostate-specific antigen free survival outcomes for patients with low, intermediate and high risk prostate cancer treatment by radical therapy. Results from the Prostate Cancer Results Study Group. BJU Int 2012; 109 Suppl 1: 22-29. Halpern E.J. Contrast-enhanced ultrasound imaging of prostate cancer, Rev Urol 2006;8(Suppl 1): S29–S37. Halpern, E.J. et al. Detection of prostate carcinoma with contrast-enhanced sonography using intermittent harmonic imaging. Cancer 104, 2373–2383 (2005). Henderson E, Milosevic MF, Haider MA, Yeung IW. Functional CT imaging of prostate cancer. Phy Med Biol. 2003;48(18):3085-3100. Hofer, C. et al. Fluorine-18-fluorodeoxyglucose positron emission tomography is useless for the detection of local recurrence after radical prostatectomy. Eur. Urol. 36, 31–35 (1999). Holupka EJ, Kaplan ID, Burdette EC, Svensson GK. Ultrasound image fusion for external beam radiotherapy for prostate cancer. Int J Radiat Oncol Biol Phys 1996;35(5):975-84. Hossny, M., Nahavandi, S., Creighton, D.: Comments on Information measure for performance of image fusion. Elect. Lett. 44, 1066–1067 (2008). Hricak, H., Choyke, P. L., Eberhardt, S. C., Leibel, S. A. & Scardino, P. T. imaging prostate cancer: a multidisciplinary perspective. Radiology 243, 2028–2053 (2007). Huber J.S., Peng Q., Moses W.W., Reutter B.W., Pouliot J., and Hsu I.C. Development of a PET-Transrectal Ultrasound Prostate Imaging System. IEEE Transactions on Nuclear Science, Vol. 58, No. 3, June 2011. 121 University of Ghana http://ugspace.ug.edu.gh International Agency for Research on Cancers (IARC) Worldwide Cancer Incidence Statistics — Prostate. JNCI Cancer Spectrum. Oxford University Press. December 19, 2001. http://web.archive.org/web/20060205235509/http://www.jncicancerspectrum.oxfordjourn als.org/cgi/statContent/cspectfstat;99. Retrieved on 5 April 2013 International Agency for Research on Cancers (IARC), Globocan 2008. Accessed at http://globocan.iarc.fr/factsheet.asp on January 22, 2013. International Atomic Energy Agency (IAEA), RAF 6044/6045 Country Report for Ghana, 2012, Page 6. International Atomic Energy Agency (IAEA). Quality assurance for PET and PET-CT systems: Human Health Series No.1. (2009). Vienna – Austria, STI/PUB/1393. International Electrotechnical Commission (IEC), Radionuclide Imaging Devices – Characteristics and Test Conditions – Part 1: Positron Emission Tomographs, IEC 61675- 1, IEC, Geneva (2008). Ivan Jambor et al, Functional Imaging of Localized Prostate Cancer Aggressiveness Using 11C- Acetate PET/CT and 1H-MR Spectroscopy. J Nucl Med 2010, doi:10.2967/jnumed. 110.078667. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global Cancer Statistics, American Cancer Society, CA Cancer J Clin 2011;61:69–90. Kaplan I, Oldenburg NE, Meskell P, Blake M, Church P, Holupka EJ. Real time MRI ultrasound image guided stereotactic prostate biopsy. Elsevier 2002;20(3):295-299. Khan, K. Crawford, D.E. and Johnson, E.L (1983) Transperineal percutaneous iridium-192 implant of the prostate. Int.J. Radial Oncol. Biol. Phys., 9,1391-5. Korporaal JG, van den Berg CAT, Jeukens CRLPN, Groenendaal G, Moman MR, Luijten P, van Vulpen M, van der Heide UA. Dynamic Contrast-enhanced CT for Prostate Cancer: Relationship between Image Noise, Voxel Size, and Repeatability. Radiology 2010;256:976-984. Kotzerke, J. et al. Carbon-11 acetate positron emission tomography can detect local recurrence of prostate cancer. Eur. J. Nucl. Med. Mol. Imaging 29, 1380–1384 (2002). Kovacs G., Potter R., Tillmann L., et al. GEC/ESTRO-EAU recommendations on temporary brachytherapy using stepping sources for localized prostate cancer. Radiother Oncol 2005; 74:137-148. Krause B. J., Souvatzoglou M., Tuncel M., et al., “The detection rate of [11C]choline-PET/CT depends on the serum PSA-value in patients with biochemical recurrence of prostate 122 University of Ghana http://ugspace.ug.edu.gh cancer,” European Journal of Nuclear Medicine and Molecular Imaging, vol. 35, no. 1, pp. 18–23, 2008. Kurhanewicz J, Swanson MG, Nelson SJ, Vigneron DB. Combined Magnetic Resonance Imaging and Spectroscopic Imaging Approach to Molecular Imaging of Prostate Cancer; J Magn Reson Imaging 2002;16(4):451-463. Kwee, S.A. et al. Use of step-section histopathology to evaluate 18F-fluorocholine PET sextant localization of prostate cancer. Mol. Imaging 7, 12–20 (2008). Landis K. Griffeth. Use of PET/CT scanning in cancer patients: technical and practical considerations. Proc (Bayl Univ Med Cent). 2005 October; 18(4): 321–330. Langen K.M, Pouliot J, Anezinos C, Aubin M, Gottschalk A.R, Hsu I-C, Lowther D, Liu YM, Shinohara K, Verhey L.J, Weinberg V, Roach M. Evaluation of ultrasound-based prostate localization for image-guided radiotherapy. Int J of Radiat Oncol, Biol, Phys, 2003;57(3):635-644. Langley SE, Laing R. Prostate brachytherapy has come of age: a review of the technique and results. BJU International 2002;89:241–249. Larson, S. M. et al. Tumor localization of 16β-18Ffluoro-5α-dihydrotestosterone versus 18F- FDG in patients with progressive, metastatic prostate cancer. J. Nucl. Med. 45, 366–373 (2004). Legg, P.A., Rosin, P.L., Marshall, D.,Morgan, J.E.: Improving accuracy and efficiency of registration by mutual information using Sturges’ histogram rule. In: Proc. Med. Image Understand. Anal., pp. 26–30 (2007). Leitzmann M.F., Platz E.A., Stampfer M.J., Willett W.C., Giovannucci E., Ejaculation Frequency and Subsequent Risk of Prostate Cancer, Journal of the American Medical Association. 2004;291(13):1578-1586. Li X., Liu Q., Wang M., Jin X., Liu Q., Yao S., Liu S., Li J., C-11 choline PET/CT imaging for differentiating malignant from benign prostate lesions. Clin Nucl Med. 2008 Oct;33(10):671-6. Lister S. (February 11, 2009). "Urine test could speed treatment of prostate cancer". London: The Sunday Times. www.timesonline.co.uk/tol/news/uk/health/article5710450.ece. Retrieved 09/08/2013. Liu, I.J., Zafar, M. B., Lai, Y.H., Segall, G.M. & Terris, M.K. Fluorodeoxyglucose positron emission tomography studies in diagnosis and staging of clinically organ-confined prostate cancer. Urology 57, 108–111 (2001) 123 University of Ghana http://ugspace.ug.edu.gh Lonescu G,Lavallee S, Demongeot J. Automated Registration of Ultrasound with CT Images: Application to Computer Assisted Prostate Radiotherapy and Orthopedics. MICCAI 1999; LNCS 1679:768-778. Makarov, D. V.; Desai, R. A.; Yu, J. B.; Sharma, R.; Abraham, N.; Albertsen, P. C.; Penson, D. F.; Gross, C. P. (2012). "The Population Level Prevalence and Correlates of Appropriate and Inappropriate Imaging to Stage Incident Prostate Cancer in the Medicare Population". The Journal of Urology 187 (1): 97–102. doi:10.1016/j.juro.2011.09.042. PMID 22088337. edit Martin T., Baltas D., Kurek R., et al. 3-D conformal HDR brachytherapy as monotherapy for localized prostate cancer. A pilot study. Strahlenther Onkol 2004; 180: 225-232. Martinez A.A., Gustafson G., Gonzalez J., et al. Dose escalation using conformal high-dose-rate brachytherapy improves outcome in unfavorable prostate cancer. Int J Radiat Oncol Biol Phys 2002; 53: 316-327. Masters Men’s Clinic, www.mastersmansclinic.com/htm. Retrieved on 25/11/2013. Mitchell H.B. Image Fusion: Theories, Techniques and Applications. Springer, Scientific Publishing, India. Pg. 3. Mohler J., Bahnson R.R., Boston B., et al. NCCN clinical practice guidelines in oncology: prostate cancer. J Natl Compr Canc Netw 2010; 8: 162-200. Nag S., Beyer D., Friedland J., et al. American Brachytherapy Society (ABS) recommendations for transperineal permanent brachytherapy of prostate cancer. Int J Radiat Oncol Biol Phys 1999; 44: 789-799. Narayana, V., Roberson, P.L, Winfield, R.J. and McLaughlin, P.W. (1997) Impact of ultrasound and computed tomography prostate volume registration on evaluation of permanent prostate implants. Int.J. Radial Oncol. Biol. P&ys.,39,341-6. Nath, R., L.L. Anderson, J.A. Meli, A.J. Olch, J.A. Stitt and J.F. Williamson, Code of practice for brachytherapy physics: Report of the AAPM Radiation Therapy Committee Task Group No. 56. Medical Physics, 1997. 24(10): p. 1557-1598. National Cancer Institute (NCI). Cancer staging. USA. Retrieved 27 November 2013. National Electrical Manufacturers Association (NEMA), Performance Measurements of Positron Emission Tomographs, NEMA Standard NU2-2007, NEMA, Washington, DC (2007). Nuñez R., Macapinlac H.A., Yeung H.W.D., et al., “Combined 18F-FDG and 11C-methionine PET scans in patients with newly progressive metastatic prostate cancer,” Journal of Nuclear Medicine, vol. 43, no. 1, pp. 46–55, 2002. 124 University of Ghana http://ugspace.ug.edu.gh O’Dowd G.J., Veltri R.W., Orozco R., Miller M.C., and Oesterling J.E., “Update on the appropriate staging evaluation for newly diagnosed prostate cancer,” Journal of Urology, vol. 158, no. 3, pp. 687–698, 1997. Oyama N, Akino H, Kanamaru H, Suzuki Y, Muramoto S, Yonekura Y, Sadato N, Yamamoto K, Okada K. 11C-Acetate PET Imaging of Prostate Cancer. J Nucl Med 2002;43(2):181- 186. Oyen R.H., van Poppel H.P., Ameye F.E., Van De Voorde W.A., Baert A.L., and Baert L.V., “Lymph node staging of localized prostatic carcinoma with CT and CT-guided fine needle aspiration biopsy: prospective study of 285 patients,” Radiology, vol. 190, no. 2, pp. 315–322, 1994. Parkin DM, Bray F, Ferlay J, Pisani P: Global Cancer Statistics, 2002. CA Cancer J Clin 2005, 55:74-108. Patel AR, Klein EA. Risk factors for prostate cancer. Nat Clin Pract Urol 2009;6(2):87-95. http://www.ncbi.nlm.nih.gov/pubmed/19198622. Pfeiffer D, Sutlief S, Feng W, Pierce HM, Kofler J. Quality assurance tests for prostate brachytherapy ultrasound systems: Report of AAPM TG 128. Med. Phys. 2008, 35 (12). Polascik T.J., Pound C.R., DeWeese T.L., et al. Comparison of radical prostatectomy and iodine 125 interstitial radiotherapy for the treatment of clinically localized prostate cancer: a 7- year biochemical (PSA) progression analysis. Urology 1998; 51: 884-889; discussion 889-90. Polascik TJ, Oesterling JE, Partin AW. Prostata specific antigen: a decade of discovery – what we have learned and where we are going. J Urol 1999;162:293-306. http://www.ncbi.nlm.nih.gov/pubmed/10411025 Porter AT, Blasko JC, Grimm PD et al. Brachytherapy for Prostate Cancer. Cancer J Clin 1995; 45: 165-178. Prostate Brachytherapy Advisory Group. Accessed on January 23, 2013 at http://www.prostatebrachytherapyinfo.net. Prostate Cancer Screening: Talking With Your Patients About Screening for Prostate Cancer. http://www.uspreventiveservicestaskforce.org/prostatecancerscreening/prostatecancerscri pt.pdf. Retrieved 2013-07-02. Pujades M.C., Camacho C., Perez-Calatayud J., et al. The use of nomograms in LDR-HDR prostate brachytherapy. J Contemp Brachytherapy 2011; 3: 121-124. 125 University of Ghana http://ugspace.ug.edu.gh Rajwade, A., Banerjee, A., Rangarajan, A.: Probability density estimation using isocontours and isosurfaces: application to information theoretic image registration. IEEE Trans. Patt. Anal. Mach. Intell. (2009). Reske S.N., Blumstein N.M., and Glatting G., “[11C]choline PET/CT imaging in occult local relapse of prostate cancer after radical prostatectomy,” European Journal of Nuclear Medicine and Molecular Imaging, vol. 35, no. 1, pp. 9–17, 2008. Reynier C, Troccaz J, Fourneret P, Dusserre A, Gay-Jeune C, Descotes JL, Bolla M, Giraud J-Y. MRI/TRUS data fusion for prostate brachytherapy: Preliminary results. Medical Physics 31, 6 (2004) 1568-75. Richter J.A., Rodrıguez M., Rioja J., et al., “Dual tracer 11C-choline and FDG-PET in the diagnosis of biochemical prostate cancer relapse after radical treatment,” Molecular Imaging and Biology. Mol Imaging Biol, 2010, 12(2):210-7. Rifkin MD, Zerhouni EA, Gatsonis CA, Quint LE, Paushter, Epstein JI, Hamper U, Walsh PC, McNiel BA. Comparison of magnetic resonance imaging and ultrasonography in staging early prostate cancer. Journal of Medicine 1990;323(10):621-626. Rinnab L., Mottaghy F.M., Simon J., et al., “[11C]choline PET/CT for targeted salvage lymph node dissection in patients with biochemical recurrence after primary curative therapy for prostate cancer: preliminary results of a prospective study,” Urologia Internationalis, vol. 81, no. 2, pp. 191–197, 2008. Rinnab L., Simon J., Hautmann R.E., et al., “[11C]choline PET/CT in prostate cancer patients with biochemical recurrence after radical prostatectomy,” World Journal of Urology, vol. 27, no. 5, pp. 619–625, 2009. Rockinger, O., Fechner, T.: Pixel-level image fusion. In: Proc. SPIE, vol. 3374, pp. 378–388 (1998). Rørvik J., Halvorsen O.J., Albrektsen G., and Haukaas S., “Lymphangiography combined with biopsy and computer tomography to detect lymph node metastases in localized prostate cancer,” Scandinavian Journal of Urology and Nephrology, vol. 32, no. 2, pp. 116–119, 1998. Scheidler, J. et al. Prostate cancer: Localization with 3D proton MR spectroscopic imaging— clinicopathologic study. Radiology 213, 473–480 (1999). Scher, B. et al. value of 11C-choline PET and PET/ CT in patients with suspected prostate cancer. Eur. J. Nucl. Med. Mol. Imaging 34, 45–53 (2007). Schiavina R., Scattoni V., Castellucci P., et al., “11C-choline positron emission tomography/computerized tomography for preoperative lymph-node staging in 126 University of Ghana http://ugspace.ug.edu.gh intermediate-risk and high-risk prostate cancer: comparison with clinical staging nomograms,” European Urology, vol. 54, no. 2, pp. 392–401, 2008. Schmid DT, John H, Zweifel R, Cservenyak T, Westera G, Goerres GW, von Schulthess GK, Hany TF. Fluorocholine PET/CT in Patients with Prostate Cancer: Initial Experience. Radiology 2005;235:623-628. Schuster, D. M. et al. initial experience with the radiotracer anti-1-amino-3-18F- fluorocyclobutan e-1-carboxylic acid with PET/CT in prostate carcinoma. J. Nucl. Med. 48, 56–63 (2007). Scott, D.W.: Multivariate Density Estimation. Wiley, Chichester (1992) Seo Y, Franc BL, Hawkins RA, Wong KH, Hasegawa BH. Progress in SPECT/CT imaging of prostate cancer. Technol Cancer Res Treat. 2006;5(4):329-36. Shreve, P. D., Grossman, H. B., Gross, M. D. & Wahl, R. L. Metastatic prostate cancer: initial findings of PET with 2-deoxy-2-[F-18] fluoro-D-glucose. Radiology 199, 751–756 (1996). Siegel R, (2011). "Cancer statistics, 2011: the impact of eliminating socioeconomic and racial disparities on premature cancer deaths.". CA Cancer J Clin 61: 212–36. doi:10.3322/caac.20121. PMID 21685461. Silverman, B.: Density Estimation for Statistical Data Analysis. Chapman and Hall, Boca Raton (1986). Skowronek J. Low-dose-rate or high-dose-rate brachytherapy in treatment of prostate cancer – between options. J Contemp Brachytherapy 2013; 5, 1: 33–41. DOI: 10.5114/jcb.2013.34342. Sobin LH, Wittekind C (eds). TNM Classification of Malignant Tumours. 7th Edition. New York: Wiley-Liss, 2009. Stefan Wachter et al. 11C-Acetate Positron Emission Tomography Imaging and Image Fusion with Computed Tomography and Magnetic Resonance Imaging in Patients with Recurrent Prostate Cancer. Journal of Clinical Oncology, 2006;24(16):2513-2519. Steggerda M, Schneider C, van HM, Zijp L, Moonen L, van der Poel H. The applicability of simultaneous TRUS-CT imaging for the evaluation of prostate seed implants. Med Phys 2005;32(7):2262-70. Suri J.S., Wilson D.L., Laxminarayan S.. Handbook of Biomedical Image Analysis, Kluwer Academic / Plenum Publishers, 2005. 127 University of Ghana http://ugspace.ug.edu.gh Sylvester J.E., Grimm P.D., Wong J., et al. Fifteen-Year Biochemical Relapse-Free Survival, Cause-Specific Survival, and Overall Survival following I-125 Prostate Brachytherapy in Clinically Localized Prostate Cancer: Seattle Experience. Int J Radiat Oncol Biol Phys 2011; 81: 376-381. Tang, J., Yang, J. C., Li, Y., Li, J. & Shi, H. Peripheral zone hypoechoic lesions of the prostate: evaluation with contrast-enhanced gray scale transrectal ultrasonography. J. Ultrasound Med. 26, 1671–1679 (2007). Taymoorian, K. et al. Transrectal broadband- Doppler sonography with intravenous contrast medium administration for prostate imaging and biopsy in men with an elevated PSA value and previous negative biopsies. Anticancer Res. 27, 4315–4320 (2007). Terris MK, McNeal JE, Stamey TA. Transrectal ultrasound imaging and ultrasound guided prostate biopsies in the detection of residual carcinoma in clinical stage A carcinoma of the prostate. J Urol. 1992 Mar;147(3 Pt 2):864-9. Thompson I., Thrasher J.B., Aus G., et al. Guidelines for the management of clinically localized prostate cancer: 2007 update. J Urol 2007; 177: 2106-2131. Toshihiko Hara, Noboru Kosaka and Hiroichi Kishi. PET Imaging of Prostate Cancer Using Carbon-11-Choline. J Nucl Med 1998;39(6):990-995. Tsao, J.: Interpolation artifacts in multimodality image registration based on maximization of mutual information. IEEE Trans. Med. Imag. 22, 854–864 (2003). Virtual Medical Centre. Prostate Cancer (Adenocarcinoma of the prostate). Accessed at http://www.virtualmedicalcentre.com/diseases/prostate-cancer-adenocarcinoma-of-the- prostate/806 on September 23, 20313. von Schulthess GK. PET-CT: principles and practices. C2I2, Summer 2004. (http://www.c2i2.org/) Walsh J.W., Amendola M.A., et al., “CT detection of pelvic and inguinal LN metastases from primary and recurrent pelvic malignant disease,” Radiology, vol. 137, pp. 157–166, 1980. Williamson J.F. “Dosimetric characteristics of the DRAXIMAGE model LS I-125 interstitial brachytherapy source design: A Monte Carlo investigation” Med. Phys. 29. 509–521 (2002). Williamson J.F. “Monte Carlo modelling of the transverse-axis dose of the Model 200 103Pd interstitial brachytherapy source” Med. Phys. 27. 643–654 (2000). Wiredu EK, Armah HB. Cancer mortality patterns in Ghana: a 10-year review of autopsies and hospital mortality. BMC Public Health 2006; 6:159-165. 128 University of Ghana http://ugspace.ug.edu.gh Wolf J.S., Cher M., Dall’Era M., Presti Jr J.C., Hricak H., and Carroll P.R., “The use and accuracy of cross-sectional imaging and fine needle aspiration cytology for detection of pelvic lymph node metastases before radical prostatectomy,” Journal of Urology, vol. 153, no. 3, pp. 993–999, 1995. Yang, J. C. et al. Contrast-enhanced gray-scale transrectal ultrasound-guided prostate biopsy in men with elevated serum prostate-specific antigen levels. Acad. Radiol. 15, 1291–1297 (2008). Yoshimoto, M. et al. Characterization of acetate metabolism in tumor cells in relation to cell proliferation: acetate metabolism in tumor cells. Nucl. Med. Biol. 28, 117–122 (2001). Yu, Y., L.L. Anderson, Z.F. Li, D.E. Mellenberg, R. Nath, M.C. Schell, F.M. Waterman, A. Wu and J.C. Blasko, Permanent prostate seed implant brachytherapy: Report of AAPM TG 64. Med Phy, 1999. 26(10): p. 2054-2076. Zhang WM, Leinonen J, Kalkkinen N, et al. Purification and characterization of different molecular forms of prostate specific antigen in human seminal plasma. Clin Chem 1995;41:1567-73. http://www.ncbi.nlm.nih.gov/pubmed/7586544. 129 University of Ghana http://ugspace.ug.edu.gh APPENDIX A: MATLAB ALGORITHMS Appendix A1: MATLAB Contrast Enhancement Algorithm function varargout = legonMedical(varargin) % LEGONMEDICAL M-file for legonMedical.fig % Begin initialization code - DO NOT EDIT gui_Singleton = 1; gui_State = struct('gui_Name', mfilename, ... 'gui_Singleton', gui_Singleton, ... 'gui_OpeningFcn', @legonMedical_OpeningFcn, ... 'gui_OutputFcn', @legonMedical_OutputFcn, ... 'gui_LayoutFcn', [] , ... 'gui_Callback', []); if nargin && ischar(varargin{1}) gui_State.gui_Callback = str2func(varargin{1}); end if nargout [varargout{1:nargout}] = gui_mainfcn(gui_State, varargin{:}); else gui_mainfcn(gui_State, varargin{:}); end % End initialization code - DO NOT EDIT % --- Executes just before legonMedical is made visible. function legonMedical_OpeningFcn(hObject, eventdata, handles, varargin) % This function has no output args, see OutputFcn. % hObject handle to figure % eventdata reserved - to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % varargin command line arguments to try3 (see VARARGIN) set(handles.original, ... 'Visible', 'off', ... 'YDir' , 'reverse'); set(handles.modified, ... 'Visible', 'off', ... 'YDir' , 'reverse' ); % Choose default command line output for try3 handles.output = hObject; % Update handles structure guidata(hObject, handles); % UIWAIT makes try3 wait for user response (see UIRESUME) % uiwait(handles.figure1); %%========================================================================= % --- Outputs from this function are returned to the command line. function varargout = legonMedical_OutputFcn(hObject, eventdata, handles) % varargout cell array for returning output args (see VARARGOUT); % hObject handle to figure % eventdata reserved - to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) 130 University of Ghana http://ugspace.ug.edu.gh % Get default command line output from handles structure varargout{1} = handles.output; function lowIn_Callback(hObject, eventdata, handles) % hObject handle to lowIn (see GCBO) % eventdata reserved - to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of lowIn as text % str2double(get(hObject,'String')) returns contents of lowIn as a double % --- Executes during object creation, after setting all properties. function lowIn_CreateFcn(hObject, eventdata, handles) % hObject handle to lowIn (see GCBO) % eventdata reserved - to be defined in a future version of MATLAB % handles empty - handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor','white'); end function lowOut_Callback(hObject, eventdata, handles) % hObject handle to lowOut (see GCBO) % eventdata reserved - to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of lowOut as text % str2double(get(hObject,'String')) returns contents of lowOut as a double % --- Executes during object creation, after setting all properties. function lowOut_CreateFcn(hObject, eventdata, handles) % hObject handle to lowOut (see GCBO) % eventdata reserved - to be defined in a future version of MATLAB % handles empty - handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor','white'); end function highIn_Callback(hObject, eventdata, handles) % hObject handle to highIn (see GCBO) % eventdata reserved - to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of highIn as text % str2double(get(hObject,'String')) returns contents of highIn as a double % --- Executes during object creation, after setting all properties. function highIn_CreateFcn(hObject, eventdata, handles) % hObject handle to highIn (see GCBO) % eventdata reserved - to be defined in a future version of MATLAB % handles empty - handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor','white'); 131 University of Ghana http://ugspace.ug.edu.gh end function highOut_Callback(hObject, eventdata, handles) % hObject handle to highOut (see GCBO) % eventdata reserved - to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of highOut as text % str2double(get(hObject,'String')) returns contents of highOut as a double % --- Executes during object creation, after setting all properties. function highOut_CreateFcn(hObject, eventdata, handles) % hObject handle to highOut (see GCBO) % eventdata reserved - to be defined in a future version of MATLAB % handles empty - handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor','white'); end function gamma_Callback(hObject, eventdata, handles) % hObject handle to gamma (see GCBO) % eventdata reserved - to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hints: get(hObject,'String') returns contents of gamma as text % str2double(get(hObject,'String')) returns contents of gamma as a double % --- Executes during object creation, after setting all properties. function gamma_CreateFcn(hObject, eventdata, handles) % hObject handle to gamma (see GCBO) % eventdata reserved - to be defined in a future version of MATLAB % handles empty - handles not created until after all CreateFcns called % Hint: edit controls usually have a white background on Windows. % See ISPC and COMPUTER. if ispc && isequal(get(hObject,'BackgroundColor'), get(0,'defaultUicontrolBackgroundColor')) set(hObject,'BackgroundColor','white'); end % --- Executes on button press in acquire. function acquire_Callback(hObject, eventdata, handles) % hObject handle to acquire (see GCBO) % eventdata reserved - to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) %%======================================================================= [filename,filepath] = uigetfile({'*.png;*.jpg;*.ttif','Image Files (*.png,*.jpg,*.ttif)';'*.*','All Files(*.*)'},'Select a file'); img1 = imshow(filename,'Parent', handles.original); img1Read = imread(filename); handles.img1Read = img1Read; guidata(hObject,handles); %%======================================================================= % --- Executes on button press in modi. function modi_Callback(hObject, eventdata, handles) % hObject handle to modi (see GCBO) 132 University of Ghana http://ugspace.ug.edu.gh % eventdata reserved - to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) lowIn = (get(handles.lowIn,'String')); lowOut = (get(handles.lowOut,'String')); highIn = (get(handles.highIn,'String')); highOut = (get(handles.highOut,'String')); gamma = (get(handles.gamma,'String')); f = handles.img1Read; dv = rgb2gray(f); if get(handles.ggray,'Value') == 0 if isempty(lowIn) || isempty(lowOut) || isempty(highIn) || isempty(highOut) || isempty(gamma) dv = imadjust(dv); imshow(dv,'Parent',handles.modified) handles.dv = dv; guidata(hObject,handles); else lowIn = str2num(get(handles.lowIn,'String')); lowOut = str2num(get(handles.lowOut,'String')); highIn = str2num(get(handles.highIn,'String')); highOut = str2num(get(handles.highOut,'String')); gamma = str2num(get(handles.gamma,'String')); if lowIn > highIn errordlg('Low-In value cannot be greater than Low-Out value','Critical') elseif lowOut > highOut errordlg('Low-In value cannot be greater than Low-Out value','Critical') else dv = imadjust(dv,[lowIn;highIn],[lowOut;highOut],gamma); imshow(dv,'Parent',handles.modified) handles.dv = dv; guidata(hObject,handles); end end elseif get(handles.ggray,'Value') == 1 dsd = str2double(handles.vv); in1 = (dsd(1)); in2 = (dsd(2)); in3 = (dsd(3)); in4 = (dsd(4)); in5 = (dsd(5)); in6 = (dsd(6)); dv = imadjust(f,[in1 in2 in3; in4 in5 in6],[]); % dv = imadjust(f,[.2 .3 0; .6 .7 1],[]); imshow(dv,'Parent',handles.modified) handles.dv = dv; guidata(hObject,handles); end % -------------------------------------------------------------------- function save_Callback(hObject, eventdata, handles) % hObject handle to save (see GCBO) % eventdata reserved - to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) fz = handles.dv; [sfile,spath] = uiputfile({'*.jpg','Jpeg,jpg file';'*.png','png file';'*.ttif','ttif file'},'Save File As'); imwrite(fz,sfile); 133 University of Ghana http://ugspace.ug.edu.gh % -------------------------------------------------------------------- function magni_Callback(hObject, eventdata, handles) % hObject handle to magni (see GCBO) % eventdata reserved - to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) f = handles.dv; figure; imshow(f) % --- Executes on button press in ggray. function ggray_Callback(hObject, eventdata, handles) % hObject handle to ggray (see GCBO) % eventdata reserved - to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) % Hint: get(hObject,'Value') returns toggle state of ggray if get(hObject,'Value') == 0 set(hObject,'String','Gray Scale Output') elseif get(hObject,'Value') == 1 set(hObject,'String','Colour Output') end % -------------------------------------------------------------------- function cclor_Callback(hObject, eventdata, handles) % hObject handle to cclor (see GCBO) % eventdata reserved - to be defined in a future version of MATLAB % handles structure with handles and user data (see GUIDATA) promt = {'Low-In(1)','Low-In(2)','Low-In(3)','High-In(1)','High-In(2)','High-In(3)'}; defAns = {'.2','.3','0','.6','.7','1'}; vv = inputdlg(promt,'COLOR DATA VARIABLE',1,defAns); handles.vv = vv; guidata(hObject,handles) ----------------------------------------------------------------------------------------------- 134 University of Ghana http://ugspace.ug.edu.gh Appendix A2: MATLAB Image Fusion Algorithm function fig = fusetool() % To open this object, just type the name of the M-file at the MATLAB % prompt. The M-file and its associated MAT-file must be on your path. load fusetool h0 = figure('Units','normalized', ... 'Color',[0.827450980392157 0.827450980392157 0.827450980392157], ... 'Colormap',mat0, ... 'CreateFcn','fusef create', ... 'Name','Image Fusion Toolbox', ... 'NumberTitle','off', ... 'PaperOrientation','landscape', ... 'PaperType','a4letter', ... 'PointerShapeCData',mat1, ... 'Position',mat2, ... 'Tag','Fig1', ... 'UserData',mat3); h1 = axes('Parent',h0, ... 'Box','on', ... 'CameraUpVector',[0 1 0], ... 'CameraUpVectorMode','manual', ... 'Color',[1 1 1], ... 'ColorOrder',mat4, ... 'FontSize',8, ... 'Position',[0.02726146220570011 0.5022156573116691 0.3779429987608426 0.4387001477104874], ... 'Tag','Axes1', ... 'XColor',[0 0 0], ... 'XTickMode','manual', ... 'YColor',[0 0 0], ... 'YTickMode','manual', ... 'ZColor',[0 0 0]); h2 = text('Parent',h1, ... 'Color',[0 0 0], ... 'HandleVisibility','off', ... 'HorizontalAlignment','center', ... 'Position',[0.4967105263157894 -0.02702702702702697 9.160254037844386], ... 'Tag','Axes1Text4', ... 'VerticalAlignment','cap'); set(get(h2,'Parent'),'XLabel',h2); h2 = text('Parent',h1, ... 'Color',[0 0 0], ... 'HandleVisibility','off', ... 'HorizontalAlignment','center', ... 'Position',[-0.02302631578947369 0.4966216216216216 9.160254037844386], ... 'Rotation',90, ... 'Tag','Axes1Text3', ... 'VerticalAlignment','baseline'); set(get(h2,'Parent'),'YLabel',h2); h2 = text('Parent',h1, ... 'Color',[0 0 0], ... 'HandleVisibility','off', ... 'HorizontalAlignment','right', ... 'Position',[-0.07236842105263158 1.131756756756757 9.160254037844386], ... 'Tag','Axes1Text2', ... 'Visible','off'); set(get(h2,'Parent'),'ZLabel',h2); h2 = text('Parent',h1, ... 'Color',[0 0 0], ... 'HandleVisibility','off', ... 'HorizontalAlignment','center', ... 135 University of Ghana http://ugspace.ug.edu.gh 'Position',[0.4967105263157894 1.02027027027027 9.160254037844386], ... 'Tag','Axes1Text1', ... 'VerticalAlignment','bottom'); set(get(h2,'Parent'),'Title',h2); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'Callback','fusef loadA', ... 'ListboxTop',0, ... 'Position',[0.5353159851301115 0.8803545051698669 0.1214374225526642 0.04874446085672082], ... 'String','Load image A', ... 'Tag','Pushbutton1'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'Callback','fusef loadB', ... 'ListboxTop',0, ... 'Position',[0.7211895910780669 0.8788774002954208 0.1201982651796778 0.05022156573116691], ... 'String','Load image B', ... 'Tag','Pushbutton1'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'Callback','fusef fusion', ... 'ListboxTop',0, ... 'Position',[0.734820322180917 0.2843426883308715 0.1201982651796778 0.04874446085672082], ... 'String','Fuse Now !', ... 'Tag','Pushbutton1'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.827450980392157 0.827450980392157 0.827450980392157], ... 'Callback','fusef zoom', ... 'ListboxTop',0, ... 'Position',[0.8884758364312267 0.5140324963072377 0.08550185873605948 0.04726735598227473], ... 'String','Zoom on', ... 'Style','checkbox', ... 'Tag','ZoomBox'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.752941 0.752941 0.752941], ... 'Callback','fusef SelDisp', ... 'ListboxTop',0, ... 'Position',[0.5043370508054522 0.1462333825701625 0.1623296158612144 0.05908419497784342], ... 'String',mat5, ... 'Style','popupmenu', ... 'Tag','FusTypMenu', ... 'Value',1); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.752941 0.752941 0.752941], ... 'ListboxTop',0, ... 'Position',[0.7657992565055761 0.09010339734121121 0.1449814126394052 0.02954209748892171], ... 'String',mat6, ... 'Style','popupmenu', ... 'Tag','SelBaseMenu', ... 'Value',3, ... 'Visible','off'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.752941 0.752941 0.752941], ... 'ListboxTop',0, ... 136 University of Ghana http://ugspace.ug.edu.gh 'Position',[0.6802973977695167 0.1757754800590842 0.04956629491945477 0.03101920236336779], ... 'String',mat7, ... 'Style','popupmenu', ... 'Tag','DecompMenu', ... 'Value',4, ... 'Visible','off'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.752941 0.752941 0.752941], ... 'Callback','fusef SelDisp', ... 'ListboxTop',0, ... 'Position',mat8, ... 'String',mat9, ... 'Style','popupmenu', ... 'Tag','CoeffMenu', ... 'Value',1, ... 'Visible','off'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'ListboxTop',0, ... 'Position',[0.6802973977695167 0.07828655834564254 0.06319702602230483 0.04135893648449039], ... 'String',mat10, ... 'Style','popupmenu', ... 'Tag','AreaMenu', ... 'Value',1, ... 'Visible','off'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.827450980392157 0.827450980392157 0.827450980392157], ... 'Callback','fusef gridonoff', ... 'ListboxTop',0, ... 'Position',[0.8884758364312267 0.570162481536189 0.08426270136307311 0.04726735598227473], ... 'String','Grid on', ... 'Style','checkbox', ... 'Tag','GridBox'); h1 = axes('Parent',h0, ... 'Box','on', ... 'CameraUpVector',[0 1 0], ... 'CameraUpVectorMode','manual', ... 'Color',[1 1 1], ... 'ColorOrder',mat11, ... 'FontSize',8, ... 'Position',[0.02853598014888337 0.01329394387001477 0.3784119106699752 0.4401772525849335], ... 'Tag','Axes2', ... 'XColor',[0 0 0], ... 'XTickMode','manual', ... 'YColor',[0 0 0], ... 'YTickMode','manual', ... 'ZColor',[0 0 0]); h2 = text('Parent',h1, ... 'ButtonDownFcn','ctlpanel SelectMoveResize', ... 'Color',[0 0 0], ... 'HandleVisibility','off', ... 'HorizontalAlignment','center', ... 'Interruptible','off', ... 'Position',[0.4967105263157895 -0.02693602693602681 9.160254037844386], ... 'Tag','Axes1Text4', ... 'VerticalAlignment','cap'); set(get(h2,'Parent'),'XLabel',h2); h2 = text('Parent',h1, ... 137 University of Ghana http://ugspace.ug.edu.gh 'ButtonDownFcn','ctlpanel SelectMoveResize', ... 'Color',[0 0 0], ... 'HandleVisibility','off', ... 'HorizontalAlignment','center', ... 'Interruptible','off', ... 'Position',[-0.02302631578947368 0.494949494949495 9.160254037844386], ... 'Rotation',90, ... 'Tag','Axes1Text3', ... 'VerticalAlignment','baseline'); set(get(h2,'Parent'),'YLabel',h2); h2 = text('Parent',h1, ... 'ButtonDownFcn','ctlpanel SelectMoveResize', ... 'Color',[0 0 0], ... 'HandleVisibility','off', ... 'HorizontalAlignment','right', ... 'Interruptible','off', ... 'Position',[-0.07894736842105263 2.242424242424242 9.160254037844386], ... 'Tag','Axes1Text2', ... 'Visible','off'); set(get(h2,'Parent'),'ZLabel',h2); h2 = text('Parent',h1, ... 'ButtonDownFcn','ctlpanel SelectMoveResize', ... 'Color',[0 0 0], ... 'HandleVisibility','off', ... 'HorizontalAlignment','center', ... 'Interruptible','off', ... 'Position',[0.4967105263157895 1.02020202020202 9.160254037844386], ... 'Tag','Axes1Text1', ... 'VerticalAlignment','bottom'); set(get(h2,'Parent'),'Title',h2); h1 = axes('Parent',h0, ... 'Box','on', ... 'CameraUpVector',[0 1 0], ... 'CameraUpVectorMode','manual', ... 'Color',[1 1 1], ... 'ColorOrder',mat12, ... 'FontSize',8, ... 'Position',[0.5012406947890818 0.3692762186115214 0.3784119106699752 0.4387001477104874], ... 'Tag','Axes3', ... 'XColor',[0 0 0], ... 'XTickLabelMode','manual', ... 'XTickMode','manual', ... 'YColor',[0 0 0], ... 'YTickLabelMode','manual', ... 'YTickMode','manual', ... 'ZColor',[0 0 0]); h2 = text('Parent',h1, ... 'Color',[0 0 0], ... 'HandleVisibility','off', ... 'HorizontalAlignment','center', ... 'Position',[0.4983606557377049 -0.02702702702702697 9.160254037844386], ... 'Tag','Text4', ... 'VerticalAlignment','cap'); set(get(h2,'Parent'),'XLabel',h2); h2 = text('Parent',h1, ... 'Color',[0 0 0], ... 'HandleVisibility','off', ... 'HorizontalAlignment','center', ... 'Position',[-0.02295081967213108 0.4966216216216217 9.160254037844386], ... 'Rotation',90, ... 'Tag','Text3', ... 'VerticalAlignment','baseline'); set(get(h2,'Parent'),'YLabel',h2); 138 University of Ghana http://ugspace.ug.edu.gh h2 = text('Parent',h1, ... 'Color',[0 0 0], ... 'HandleVisibility','off', ... 'HorizontalAlignment','right', ... 'Position',[-1.324590163934426 1.435810810810811 9.160254037844386], ... 'Tag','Text2', ... 'Visible','off'); set(get(h2,'Parent'),'ZLabel',h2); h2 = text('Parent',h1, ... 'Color',[0 0 0], ... 'HandleVisibility','off', ... 'HorizontalAlignment','center', ... 'Position',mat13, ... 'Tag','Text1', ... 'VerticalAlignment','bottom'); set(get(h2,'Parent'),'Title',h2); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.827450980392157 0.827450980392157 0.827450980392157], ... 'FontSize',12, ... 'ListboxTop',0, ... 'Position',mat14, ... 'String','Input image A', ... 'Style','text', ... 'Tag','StaticText1'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.827450980392157 0.827450980392157 0.827450980392157], ... 'FontSize',12, ... 'ListboxTop',0, ... 'Position',mat15, ... 'String','Input image B', ... 'Style','text', ... 'Tag','StaticText1'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.827450980392157 0.827450980392157 0.827450980392157], ... 'FontSize',12, ... 'ListboxTop',0, ... 'Position',[0.5775434243176179 0.8094534711964549 0.2258064516129032 0.02806499261447563], ... 'String','Fused image', ... 'Style','text', ... 'Tag','StaticText1'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.827450980392157 0.827450980392157 0.827450980392157], ... 'FontSize',12, ... 'ListboxTop',0, ... 'Position',[0.5353159851301115 0.9394387001477104 0.3060718711276332 0.02954209748892171], ... 'String','Select input image files', ... 'Style','text', ... 'Tag','StaticText1'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.827450980392157 0.827450980392157 0.827450980392157], ... 'ListboxTop',0, ... 'Position',[0.5105328376703842 0.2097488921713441 0.1437422552664188 0.02215657311669128], ... 'String','Fusion method', ... 'Style','text', ... 'Tag','StaticText1'); h1 = uicontrol('Parent',h0, ... 139 University of Ghana http://ugspace.ug.edu.gh 'Units','normalized', ... 'BackgroundColor',[0.827450980392157 0.827450980392157 0.827450980392157], ... 'ListboxTop',0, ... 'Position',mat16, ... 'String','Level', ... 'Style','text', ... 'Tag','StaticText1'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.827450980392157 0.827450980392157 0.827450980392157], ... 'ListboxTop',0, ... 'Position',[0.5105328376703842 0.1225997045790251 0.1437422552664188 0.0206794682422452], ... 'String','Highpass combination', ... 'Style','text', ... 'Tag','StaticText1'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.827450980392157 0.827450980392157 0.827450980392157], ... 'ListboxTop',0, ... 'Position',[0.7645600991325898 0.121122599704579 0.1437422552664188 0.02363367799113737], ... 'String','Lowpass combination', ... 'Style','text', ... 'Tag','StaticText1'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.827450980392157 0.827450980392157 0.827450980392157], ... 'ListboxTop',0, ... 'Position',[0.6827757125154894 0.121122599704579 0.05080545229244114 0.02363367799113737], ... 'String','Area', ... 'Style','text', ... 'Tag','StaticText1'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.827450980392157 0.827450980392157 0.827450980392157], ... 'FontSize',10, ... 'ForegroundColor',[1 1 0], ... 'ListboxTop',0, ... 'Position',[0.5043370508054522 0.02806499261447562 0.3977695167286245 0.03692762186115214], ... 'Style','text', ... 'Tag','MessText'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'Callback','fusef saveF', ... 'ListboxTop',0, ... 'Position',[0.5390334572490706 0.2843426883308715 0.1214374225526642 0.04874446085672082], ... 'String','Save fused image', ... 'Tag','Pushbutton4'); h1 = uicontrol('Parent',h0, ... 'Units','normalized', ... 'BackgroundColor',[0.827450980392157 0.827450980392157 0.827450980392157], ... 'Callback','fusef gridonoff', ... 'ListboxTop',0, ... 'Position',[0.8884758364312267 0.6277695716395864 0.0966542750929368 0.04726735598227473], ... 'String','New figure', ... 'Style','checkbox', ... 'Tag','FigBox'); if nargout > 0, fig = h0; end ------------------------------------------------------------------------ 140 University of Ghana http://ugspace.ug.edu.gh APPENDIX B: PUBLISHED AND ACCEPTED ARTICLES Journal Articles 1. F. Hasford, B. Van Wyk, T. Mabhengu, M.D.T. Vangu, A.K. Kyere, J.H. Amuasi. Determination of dose delivery accuracy in CT examinations. Journal of Radiation Research an Applied Sciences (2015), http://dx.doi.org/10.1016/j.jrras.2015.05.006. 2. F. Hasford, B. Van Wyk, T. Mabhengu, M.D.T. Vangu, A.K. Kyere, J.H. Amuasi. Effect of Radionuclide Activity Concentration on PET-CT Image Uniformity. World Journal of Nuclear Medicine. WJNM_44_15. Accepted for publication. 3. F. Hasford, J.H. Amuasi, A.K. Kyere, M.D.T. Vangu, Quantitative Assessment of Radionuclide Uptake and Positron Emission Tomography-Computed Tomography Image Contrast. World Journal of Nuclear Medicine. WJNM_78_15. Accepted for publication Conference Presentations 1. F. Hasford, J.H. Amuasi, A.K. Kyere, M.D.T. Vangu. Ultrasound and PET-CT Image Fusion for Prostate Brachytherapy Image Guidance. International Conference on Clinical PET-CT and Molecular Imaging (IPET2015), 5 – 9 October 2015, Vienna – Austria. Poster presentation 2. F. Hasford, B. Van Wyk, T. Mabhengu, M.D.T. Vangu, A.K. Kyere, J.H. Amuasi. Quantitative Assessment of PET-CT Image Uniformity. International Conference on Clinical PET-CT and Molecular Imaging (IPET2015), 5 – 9 October 2015, Vienna – Austria. Poster presentation 3. F. Hasford, J.H. Amuasi, A.K. Kyere, M.D.T. Vangu. Ultrasound and PET-CT Image Fusion for Prostate Brachytherapy Image Guidance. Maiden University of Ghana Doctoral Research Conference, 5 – 6 November 2015, Accra – Ghana. (Winner: Best Poster presentation) 141