Purpose: Prospective data-driven gated (DDG) PET requires advance selection of patients who may benefit from gating, and leaves DDG-PET/CT registration to chance. Retrospective DDG of both PET and CT can reduce CT radiation dose, shorten scan time, improve PET/CT registration, and correct tumor motion on all patients. The first DDG-PET/CT (DDG-PET and DDG-CT) has been prototyped for patient care in a network of seven PET/CT scanners. Here, the impacts of misregistration and motion correction on image quality and lesion detectability were assessed with the new DDG-PET/CT technique.
Methods: A low-dose cine CT was acquired in misregistered regions to enable both average CT (ACT) and DDG-CT for attenuation-correction of static and DDG-PET, respectively, yielding baseline PET/CT, PET/ACT and DDG-PET/CT. For DDG-PET, end-expiration data was derived from 50% of PET data at 30% from end-inspiration. For DDG-CT, end-expiration CT data was extracted from cine CT by lung CT number and body contour. SUVmax, SUVmean, SNR, and CNR were compared for 121 lesions from 21 consecutive patients.
Results: Improved registration with ACT did not change SNR. DDG-PET/CT showed decreased background and lesion SNR with median±σ changes of -24±10% (p<0.0001), equivalent to using ~57% of total PET counts. Changes from baseline in SUVmax and CNR were 21±41% (p<0.0001) and 6±76% (p=0.27) for PET/ACT, and 68±65% (p<0.0001) and 37±141% (p<0.0001) for DDG-PET/CT. Lesions with ambiguous detectability at baseline (CNR<10) showed the most significant CNR increases in DDG-PET/CT, with nearly all lesions reaching CNR>5.
Conclusion: Using 50% of PET data in DDG-PET leads to the expected increase in image noise. However, the increase in SUV due to improved registration and motion correction in DDG-PET/CT is more substantial than the increased image noise. Therefore, lesion detectability is still largely improved with DDG, especially for lesions with uncertain detectability (CNR<10) at baseline.
Funding Support, Disclosures, and Conflict of Interest: This research was supported in part by NIH grants 1R21CA222749-01A and R03EB030280, and a ROSI grant from UT MD Anderson Cancer Center. It was conducted at the MD Anderson Center for Advanced Biomedical Imaging in-part with equipment support from GE Healthcare. T.P. was a consultant for Bracco Medical Systems, LLC.
Not Applicable / None Entered.