Effect of adaptive statistical iterative reconstruction (ASiR) on dosimetry using a PET/CT unit

Abstract

Positron Emission Tomography (PET) - Computed Tomography (CT) is a sensitive medical imaging modality for oncology. This combined technique provides a high quality of images with valuable diagnostic information. CT provides high-resolution anatomy depiction, while PET provides important information on the metabolic activities of various organs. It has been widely used for diseases staging, treatment planning, treatment following up and detecting diseases recurrence. Adaptive Statistical Iterative Reconstruction (ASiR) is an advanced radiation modulation technique. It is a partially iterative technique that repeats back projection in many times. ASiR has been proven to reduce CT radiation dose, evident by reduction in weighted Computed Tomography Dose Index (CTDI) and dose-length product (DLP). The objective of this study was to evaluate the feasibility of ASiR on CT dosimetry modulation on a PET/CT unit. At the PET/CT unit of the University of Hong Kong, the system was upgraded to GE VCT Discovery 610 (Technology) ASiR in October 2015. We evaluated the CT dosimetry before and after the system upgrade. As a result, the images reconstructed before the upgrade reflected reconstructed images without applying ASiR, while those after the system upgrade represented images reconstructed by ASiR. Two main examination protocols in CT and PET/CT were chosen for evaluation. For CT protocols, multi-phased CT abdomen (4 phases) and CT thorax were chosen, and for PET/CT, contrast and non-contrast protocols were chosen. Comparison of the dosimetry was made using Mann-Whitney U test, and statistical significance was assumed when p<0.05. A significant dose reduction (p<0.01) was observed in CT Abdomen examinations after applying ASiR. The mean DLP decreased from (744±181), (844±84), (1083±160), and (1020±120) to (625±170), (698±125), (856±224), and (798±154) in the different phases, namely non-contrast, arterial, portovenous, and delayed phases, respectively. In addition, the mean DLP in CT Thorax reduced significantly (p<0.01) from (435±192) to (345±47) in non-contrast CT thorax after the system upgrade, but the mean DLP in portovenous phases did not show significant difference (p=0.331) from (469±143) to (476±94). In PET/CT, a significant dose rise (p<0.01) in both the contrast and non-contrast scan protocols were noted. The mean DLP increased from (950±442) to (1030±344), and (449±264) to (528±187) in contrast and non-contrast PET/CT scans, respectively, after the system upgrade. However, the mean DLP in delayed scans decreased significantly (p<0.01) after the system upgraded. The study results suggested that ASiR reduced CT dosimetry effectively but not with PET/CT protocols. The lack of radiation dose reduction in PET/CT acquisition was due to the inability to apply ASiR during the PET/CT protocols. Furthermore, the current PET/CT acquisition used a higher range of automated X-ray tube current (mA), which would explain the paradoxical dosimetry rise after the system upgrade, in which the reason is unclear at present and we are investigating this issue with the vendor and engineers.