Optimizing acquisition times for total-body positron emission tomography/computed tomography with half-dose F-18-fluorodeoxyglucose in oncology patients

Background The present study aimed to explore the boundary of acquisition time and propose an optimized acquisition time range for total-body positron emission tomography (PET)/computed tomography (CT) oncological imaging using half-dose (1.85 MBq/kg) F-18-fluorodeoxyglucose activity based on diagnostic needs. Methods In this retrospective study based on a total-body PET system (uEXPLORER), an exploration cohort (October 2019-December 2019) of 46 oncology patients was first studied. The acquisition time for all patients was 15 min, and the acquired images were reconstructed and further split into 15-, 8-, 5-, 3-, 2-, and 1-min duration groups (abbreviated as G15, G8, G5, G3, G2, and G1). The image quality and lesion detectability of reconstructed PET images with different acquisition times were evaluated subjectively (5-point scale, lesion detection rate) and objectively (standardized uptake values, tumor-to-background ratio). In the same way, the initial optimized acquisition times were further validated in a cohort of 147 oncology patients (December 2019-June 2021) by using the Gs images (the images obtained using the 15- and 10-min acquisition times) as controls. Results In the exploration cohort, the subjective scores for G1, G2, G3, G5, and G8 images were 2.0 +/- 0.2, 2.9 +/- 0.3, 3.0 +/- 0.0, 3.9 +/- 0.2, and 4.2 +/- 0.4, respectively. Two cases in G1 were rated as 1 point. No significant difference in scores was observed between G5 and G8 (p > 0.99). In general, groups with a longer acquisition time showed lower background uptake and lesion conspicuity. Compared with G15, lesion detection rate significantly reduced to 85.3% in G1 (p < 0.05). In the validation cohort, the subjective score was 3.0 +/- 0.2 for G2, 3.0 +/- 0.1 for G3, 3.6 +/- 0.5 for G5, 4.0 +/- 0.3 for G8, and 4.4 +/- 0.5 for Gs. Only the scores between G2 and G3 were not significantly different (p > 0.99). The detection rates (204 lesions) significantly reduced to 94.1-90.2% in G3 and G2 (all p < 0.05). Conclusion A 2-min acquisition time provided acceptable performance in certain groups and specific medical situations. And protocols with acquisition times >= 5 min could provide comparable lesion detectability as regular protocols, showing better compatibility and feasibility with clinical practice.