2485 - Inter-fractional and Intra-fractional Metabolic Dynamics in Biology-Guided Radiotherapy: A Site-Specific Analysis of Lung and Bone Lesions
Presenter(s)
Y. Huang1,2, Q. Xu1, W. T. Watkins1, A. Amini1, T. M. Williams1, A. Liu1, and C. Han1; 1Department of Radiation Oncology, City of Hope National Medical Center, Duarte, CA, 2Jiangxi Cancer Hospital, Nanchang, China
Purpose/Objective(s): Biology-guided radiotherapy (BgRT) uses real-time PET signal to target metabolically active tumor subvolumes. However, the stability of the biological signal across different timescales and its variation by tumor site remain poorly characterized. This study comprehensively quantifies inter-fractional and intra-fractional changes in FDG uptake, metabolic tumor volume (MTV), and BgRT delivery eligibility parameters (activity concentration [AC], normalized target signal [NTS]) in patients receiving multi-fraction BgRT for lung and bone lesions, with site-specific subgroup analyses.
Materials/Methods: Fifteen patients with FDG-avid lesions (8 lung, 7 bone) treated with multi-fraction BgRT (24–30 Gy in 2–5 fractions) were retrospectively analyzed. Each patient received a functional modelling (FM) PET scan on the BgRT machine for treatment planning. Daily pre-treatment PET scans (prescan) were acquired after FDG injection. For intra-fraction analysis, sequential PET acquired during beam delivery (four pass) were compared to prescan PET. MTV was defined as the volume within GTV with =50% of SUVmax. AC and NTS were compared between the first fraction (Fx1) and FM PET, as well as between Fx1 and the last fraction. Statistical significance was assessed using Wilcoxon signed-rank tests.
Results: In the overall cohort (34 fractions), SUVmax decreased significantly from Fx1 to subsequent fractions (5.98 ± 2.35 vs. 5.19 ± 1.87; p < 0.001), as did SUVmean (3.44 ± 1.24 vs. 3.11 ± 1.24; p = 0.038). Site-specific analysis revealed marked reductions in lung lesions (SUVmax: 6.70 ± 2.70 vs. 5.05 ± 1.44, p < 0.001; SUVmean: 3.48 ± 1.29 vs. 2.73 ± 0.73, p < 0.001), whereas no significant changes were observed in bone lesions. Intra-fractionally, four pass PET demonstrated significantly higher SUVmax (overall: 5.22 ± 1.75 vs. 6.77 ± 2.42, p < 0.001) and SUVmean (overall: 3.08 ± 0.90 vs. 3.86 ± 1.28, p < 0.001) compared with prescan. MTV remained stable both in inter-fractional comparisons and between prescan and FM PET but decreased significantly from prescan to four-pass PET scans (10.22 ± 5.70 vs. 9.66 ± 5.13 cm³, p < 0.001). No statistically significant changes were observed in AC or NTS between Fx1 prescan and FM PET (p = 0.191 and 0.125, respectively). However, AC decreased significantly from Fx1 to the last fraction (18.92 ± 10.54kBq/mL vs. 14.21 ± 6.60 kBq/mL, p = 0.029), while the change in NTS remained insignificant. In the lung subgroup, NTS showed a significant decrease (11.36 ± 6.13 vs. 7.88 ± 3.03, p = 0.036). Among all patients evaluated, one patient was treated without PET guidance in the last fraction due to low PET statistics in the prescan.
Conclusion: PET imaging during BgRT reveals complex metabolic dynamics with marked site-specific differences. Lung lesions exhibit rapid inter-fraction declines in FDG avidity, while bone lesions remain stable. Intra-fraction variations in SUV indicate dynamic tracer uptake during delivery, supporting the rationale for real-time biological guidance.