Development Of National Indication-Based Diagnostic Reference Levels And Optimisation Methods For Computed Tomography Examinations In Ghana

Abstract

ABSTRACT Background: Diagnostic reference levels (DRLs) and dose optimisation methods are crucial for effective radiation dose management. Countries utilising ionising radiation for medical purposes are required to develop and implement them, taking into consideration their clinical situations, infrastructure, population characteristics as well as social, technical and economic factors. However, in Ghana, there is no established national indication-based DRL. Main Objective: The main objective of this study was to develop national indication-based DRL values for common and prioritised indications of the adult human body for clinical application in Ghana. It was also to assess the risk of undertaking each indication-based CT examination, and also propose some steps for dose optimisation. Materials and Methods: The methodological approach recommended by the International Commission for Radiological Protection (ICRP), publication 135, for the development of DRLs, was employed. Studies on CT infrastructure and common indications as well as quality management systems (QMS) were conducted. Quality control (QC) tests were undertaken using a CT dose profiler, barracuda set, uniform water phantom and an ImageQC software v.1.43. Radiologists were mainly requested to define the basic diagnostic requirement of each indication. Dose descriptors such as volume weighted CT dose index (CTDIvol) and dose length product (DLP) of reported CT scans were retrieved from the picture archiving and communication system (PACS) of scanners, constituting 71.4% of the total CT scanners in Ghana. Overall, 3,960 data sets were collected for all the common and prioritised indications which included: cerebrovascular accident (CVA) or stroke, head trauma/injury, brain tumour/space occupying lesion (SOL), lung tumour/cancer, chest lesion with chronic kidney disease, abdominopelvic lesion, kidney stone, urothelial malignancy/CT intravenous urography (CT-IVU) and pulmonary embolism (PE).

ImageJ software version 1.52 was used to analyse the objective image qualities. Statistical Package for the Social Sciences (SPSS) version 23.0 was used to extract the DRL values for the common indications in CT examinations. Microsoft excel version 2013 was used to pictorially project the results and also develop a tool (BOTB) for dose monitoring. Lifetime Attributable Risk (LAR) of cancer incidence and mortality were estimated for various organs using a Monte Carlo-based software (National Cancer Institute Dosimetry System CT software version 2.1) and the Biological Effects of Ionising Radiation (BEIR) VII model. An anthropomorphic Alderson RANDO phantom and patients’ clinical data were used to explore an optimisation method for cerebrovascular accident (CVA) imaging. Regression analyses were further used to model equations for organ doses in CVA imaging. CT phantom PBU-60 was also used to evaluate automatic exposure control (AEC) dose impact in facilities operating without AEC systems. In all inferential analyses, a p-value of ≤ 0.05 was used to interpret the findings as statistically significant. Main Results: The various indications and their respective projected DRL values in terms of CTDIvol (mGy) and DLP (mGy.cm) were CVA/stroke (77 mGy; 1313 mGy.cm), head trauma/injury (76 mGy; 1596 mGy.cm), brain tumour/SOL (77 mGy; 2696 mGy.cm), lung tumour/cancer (12 mGy; 828 mGy.cm) and chest lesion with chronic kidney disease (13 mGy; 467 mGy.cm). Others were abdominopelvic lesion (17 mGy; 1299 mGy.cm), kidney stone (15 mGy; 731 mGy.cm), urothelial malignancy/CT-IVU (11 mGy; 1449 mGy.cm) and pulmonary embolism (14 mGy; 942 mGy.cm). The risk of PE radiation-induced breast cancer ranged from 6-115.8 people in 100,000 procedures. Moreover, CT-IVU radiation-induced colon cancer risks ranged from 53.3-66.4 people in 100,000 procedures. About 1 in 38,462 to 1 in 14,706 patients were also likely to develop ovarian cancer due to CT-IVU examinations in Ghana. A novel examination protocol was further developed in the study that could be used to scan CVA related conditions

with optimal image quality, while reducing the mean effective dose of the facilities by 23.8%, and organ doses by 32% (lens), 70.7% (spinal cord), 57.2% (thyroid) and 75.6% (oral cavity). Moreover, eight organ dose equations were developed to aid in dose management. Finally, if AEC are used in facilities operating without such systems, radiation dose levels could also be reduced by a range of 46.4-58.3% without any significant compromise on image quality. Conclusion: The projected indication-based DRL values and optimisation methods could be used to manage CT radiation dose in Ghana.