Improving lesion detectability in low uptake 18 F-FDG breast cancer by optimizing PET imaging time

There have been anecdotal reports that delayed FDG-PET oncology imaging can improve the SNR for known lesions. Our goal was to evaluate if it is possible to optimize lesion detectability by adjusting the duration between FDG injection and image acquisition. Methods: We combined three components: a kinetic model of 18FFDG radiotracer uptake, a forward model of PET data acquisition, and observer models to estimate lesion detectability. Model parameters and arterial input functions were estimated from dynamic PET studies of patients with early-stage breast cancer. These were used to generate time-activity curves (TACs). For each TAC, realistic PET sinograms were simulated, and images were reconstructed using OSEM. There were 600 i.i.d. realizations for each case. Both nonprewhitening matched filter (NPWMF) and channelized Hoteling observer (CHO) statistics were used to generate receiver operating characteristic (ROC) curves, which were summarized using the area under the curve (AUC). Results: The AUC increased for several hours past the current standard post-injection imaging window of one hour. This improvement is driven by the continued accumulation of FDG in the tumor before the impact of increased noise due to radiotracer decay begins to dominate. Conclusion: These results imply that delayed PET imaging may reveal low-conspicuity lesions in patients that would have otherwise gone undetected. However, further validation of this approach is needed.