Evaluation of performance of computed tomography scanners in Tanzania

The performance characteristics of CT scanners were evaluated from five large hospitals using the quality criteria proposed by the EC. The performance characteristics evaluated in this study include dose related parameters, physical image quality parameters, diagnostic image criteria and patient doses. The dose related technical parameters employed by each scanner in this study were verified in terms of beam alignment, slice thickness accuracy, kV reproducibility and accuracy, output reproducibility and linearity, and half value layer (HVL) measurements. These parameters were verified using non-invasive X ray test device (Victoreen, model 4000+-SI), a pencil ionization chamber (Diados, type M3009, PTW-Freiburg) coupled to electrometer (Diados, type 11003, PTW-Freiburg), and a CT performance phantom (Victoreen, model 76-410-4130) of 21.5 cm diameter. Methods for verifications of these parameters were based on those described in the AAPM Report 39. The physical image quality performance of a CT scanner from typical scanning techniques employed by each scanner was objectively evaluated in terms of CT number accuracy and linearity, CT image noise and uniformity, and subjectively evaluated in terms of high and low contrast resolution. The above tests were evaluated using CT performance phantom described above. Methods for evaluation of these tests were based on those described in the AAPM Report 39. The image quality evaluations were performed by utilizing the defined set of diagnostic image criteria proposed by EC guidelines for general CT examination of head, chest and abdomen. Two experienced radiologists using criteria recommended by the EC subjectively evaluated the qualities of these images. Estimation of patient dose in terms of weighted CT dose index (CTDIw), and dose length product (DLP) were performed using ion chamber by employing standard phantoms and typical exposure parameters. The results of dose related and physical image quality indicate that most of the CT facilities under study were within recommended tolerance limits, suggesting that maintenance of CT although irregular was adequate. The mean percentage image quality score for general CT examinations of head, chest and abdomen were 71%, 68% and 61%, respectively. The mean values of CTDIw per examination for almost all hospitals were below the reference dose levels (RDLs), while the mean values of DLP per hospital for all except the head were above the proposed RDLs as shown. The poor performance of CT scanners in terms of high values of DLP and image criteria score might be largely attributed to the lack of optimal use of CT due to inadequate skills and lack of in house quality assurance programs. It was therefore concluded that scanning parameters should be optimized and training programmes introduced for CT personnel