Quantification of the change in SUV due to the presence of defective detector blocks in clinical FDG PET imaging

Objectives Clinical PET systems based on block detector designs suffer occasional block detector failures, which can result in patient scan cancellations. In this study we examine the effects of defective detectors on measurements of SUV_max through imaging of a torso phantom.

Methods A Data Spectrum anthropomorphic torso phantom was imaged in a normally functioning Siemens Biograph 16 HiRez PET/CT scanner using a whole-body imaging protocol. The torso volume had an activity concentration defined as a SUV of 1.0. Six spherical and shell lesions with different activity concentrations were placed in the phantom. 11 one-block and 3 two-block defect cases were created by zeroing lines of response in the sinograms. The images were reconstructed using the standard clinical OSEM protocol. SUV_max was measured for each lesion and defect case on a Siemens Leonardo Workstation, with the SUV_max for the normal case (i.e. no defect present) serving as the true value for comparison.

Results Our preliminary results show that the average decrease in SUV_max over all lesions was 2% for one-block and 6% for the two-block defect cases. The largest observed change in SUV_max was 6% for a one-block defect and 8% for a two-block defect.

Conclusions In the presence of one or two defective block detectors, the observed change in SUV_max was in no case large enough to affect the clinical diagnosis of a lesion. Thus for these cases we recommend proceeding with routine patient scanning, with the caveat of making the reading physicians aware of the defect being present.

Research Support Winnipeg Health Sciences Foundation