Feasibility of low-dose FDG for whole-body TOF PET/CT oncologic workup

Objectives Despite significant advances in PET instrumentation and image reconstruction, few efforts have evaluated if administered FDG dose can be reduced in current generation systems without a loss in overall diagnostic utility. Reduced patient dose is beneficial for minimizing radiation dose and associated risks in particular for patients with younger age and those who receive multiple follow up scans. This study evaluates truncated duration PET scans, as a surrogate for reduced dose imaging, to determine if administered activity can be revised with modern PET/CT systems.

Methods Wholebody FDG-PET/CT list-mode data were acquired on a current generation TOF PET system with fixed acquisition durations (4 min/FOV). For each patient (10 mCi injection), the list-mode data were truncated to form half and quarter duration data sets. Two reviewers evaluated 62 lesion sites in 39 PET/CT image volumes (48-135 kg). With the full duration study as the reference standard, detection and quantitative performance were evaluated.

Results Lesion detection accuracy was identical in the half duration images. In the quarter duration studies, the increase in image noise led to 7 false positives, 4 false negatives, and a reduction in detection confidence. Quantitative evaluation showed the mean value in uniform regions and primary lesions did not change significantly between the different duration studies (average change < 2+/- 4%). The noise in uniform regions increased less than theoretically expected in the half and quarter duration studies (+22%, +58%) due to regularization during reconstruction.

Conclusions Evaluation of image volumes generated from shorter acquisition durations suggests that for 10mCi injection, 4min/FOV imaging, the FDG dose can be reduced by 50% without a loss of diagnostic performance. While many institutions have translated instrumentation advances to shorter scans, the current concern of risks from radiation warrants intentionality in selecting appropriate acquisition durations and injected activity