Innovative Dosimetric Advances in Total-Body PET/CT with PD-L1 Tracers

A new hybrid curve-fitting method for total-body dynamic PET/CT raises estimated radiation doses for a PD-L1–targeted tracer and narrows uncertainty in time–activity integration for oncologic dosimetry.

The study used total-body dynamic PET/CT in 16 patients with solid tumors to generate dense time–activity curves and compare fitting approaches. Quantitatively, the hybrid curve-fitting method produced consistently higher time-integrated activity coefficients (TIACs), increasing the effective dose estimate to 36.33 ± 6.18 µSv/MBq versus 20.47 ± 3.08 µSv/MBq with routine fitting. The hybrid approach also showed improved dosimetric accuracy versus routine bi‑exponential fitting across organs, driven by greater fit stability and reduced bias when integrating early-phase activity.

Organ-level analysis localized the highest absorbed doses to the urinary system—kidneys, bladder wall, and spleen—reflecting concentrated renal and bladder counts. The PD-L1-targeted tracer demonstrated low uptake elsewhere but marked urinary excretion that inflates bladder and renal TIACs; careful bladder ROI placement and attention to voiding timing are therefore needed to avoid dose overestimation and to improve interpretive accuracy.

Technically, the hybrid method pairs rectangular integration for early frames with exponential fits for later phases, stabilizing parameter recovery when early data are noisy. Compared with routine methods, this reduces sensitivity to frame selection and diminishes noise-driven bias in organ-dose estimates. Implementation will likely require software updates for time–activity curve fitting, adjusted frame-reconstruction settings, and standardized VOI delineation—changes to workflow rather than immediate shifts in clinical endpoints.

Reported radiation exposure remained within diagnostic ranges despite higher TIACs with the hybrid fit: effective doses were 20.47 ± 3.08 µSv/MBq (routine) and 36.33 ± 6.18 µSv/MBq (hybrid), with no unexpected safety signals reported. Limitations include a small cohort (n=16), single-tracer evaluation, and the need for external validation across tracers and centers. Next steps should focus on prospective multicenter validation, refinement of automated fitting pipelines, and assessment of any downstream impact on patient management.