Voxel-Wise Characterization of Post-Radiotherapy Liver Function Changes Using Longitudinal Sulfur Colloid SPECT/CT
Abstract
Purpose
Patients receiving radiotherapy (RT) for liver cancer have limited liver function reserves, which increases risk of radiation-induced liver injury, but tools to evaluate functional response post-RT are lacking. We modeled liver function changes on 99mTc sulfur colloid (SC) SPECT/CT to evaluate voxel-level radiation dose response and enable association with baseline risk factors.
Methods
Patients with liver cancer were prospectively enrolled to receive photon SBRT or proton therapy with SC SPECT/CT imaging before and 1-month post-RT. SPECT/CT images were deformably registered to planning CT. Physical dose was converted to EQD2 (alpha/beta=3). Voxel-wise functional change was quantified and sigmoidal dose–response relationships were modeled with 5Gy bins. For each patient, maximum functional loss (Rmax) and dose-response coherence (R²) were derived. Associations with baseline liver function, including Child–Pugh (CP) risk category, total liver function (TLF), and relative liver activity (RLA) were assessed using Spearman correlation and group comparisons.
Results
30 patients were analyzed. Median functional loss was 43% [IQR, 30–65%]. Greater functional loss was associated with high CP class (p=0.050), lower TLF (p=0.005), and lower RLA (p=0.024). High-risk patients showed more coherent dose-response behavior (R²=0.98 [0.95-0.99]) compared with low-risk patients (R²=0.69 [0.48-0.91]), with stratification by baseline liver function (CP p=0.021; TLF p=0.002; RLA p=0.015). No trends for SBRT vs protons were evident. Joint analysis revealed that combining RLA and TLF captures complementary baseline effects, with lower baseline liver function associated with higher (Rmax), more coherent (R2) dose-response behavior.
Conclusion
SC SPECT/CT dose-response modeling enables quantitative assessment of both the magnitude and coherence of liver functional response following radiotherapy. Baseline liver functional reserve strongly modulates dose–response behavior, with high-risk patients demonstrating more uniform functional decline, while lower-risk patients exhibit greater inter-patient heterogeneity. These findings support adaptive, function-guided RT approaches using functional liver imaging as a biomarker for personalized, liver-sparing planning.