Purpose/Objective(s):
Post-SBRT CT frequently shows persistent opacity from radiation lung injury, complicating early response assessment and differentiation of viable tumor from post-treatment change. We tested whether baseline FDG-PET SUVmax and early changes in ADCmean and tumor size predict a viability-oriented MRI surrogate, non-visible lesion on the ADC map at 6 months (NVL6M), after lung SBRT.
Materials/Methods:
Patients treated with lung SBRT (40 Gy/4 fractions) were prospectively enrolled. CT/MRI were obtained as a same-day set at baseline and at 1 month, 3 months (±7 days), and 6 months (±14 days). MRI included DWI/ADC plus T1- and T2-weighted imaging; T2WI supported interpretation of radiation pneumonitis versus residual tumor. ADCmean was measured from a manual 2D maximal-slice ROI delineated on DWI and T2WI and applied to the ADC map. NVL was defined as inability to delineate a measurable lesion on the ADC map. All patients underwent baseline FDG-PET/CT within 2 weeks prior to SBRT. Six-month MRI was planned for all; if NVL was achieved at 3 months, 6-month MRI could be omitted at patient request, resulting in response-driven missingness. RECIST 1.1 was measured on CT and T2WI. CT surveillance was protocolized up to 3 years; this report reflects early follow-up (~1.5 years since study initiation). Comparisons used Mann–Whitney U tests (exploratory).
Results:
Twelve SBRT patients were analyzed (NVL6M=7; non-NVL6M=5). Baseline SUVmax was lower in NVL6M versus non-NVL6M (median 7.0 vs 9.0; p=0.048). ?ADC% did not clearly separate groups (Pre?1M: 47.6% vs 40.3% [available n=6/3], p=0.795; Pre?3M: 90.1% vs 69.0% [n=5/3], p=0.571). ?LD% trended toward greater reduction in NVL6M (Pre?1M: -57.7% vs -18.3% [n=7/5], p=0.268; Pre?3M: -54.2% vs -38.5% [n=6/4], p=0.571). Among three small cell carcinoma cases, two developed mediastinal nodal relapse despite primary-lesion NVL at 3 months.
Conclusion:
Lower baseline FDG-PET SUVmax was associated with NVL6M after lung SBRT, suggesting baseline metabolic activity may influence MRI-based lesion non-visibility. Early ?ADC% was not a robust discriminator in this small cohort and was limited by response-driven missingness as lesions became non-measurable. NVL should be interpreted as a local surrogate and does not exclude nodal relapse, particularly in small cell carcinoma. Larger cohorts with longer follow-up will assess whether NVL6M improves early post-SBRT response assessment beyond CT-based measurements and predicts =12-month clinical endpoints.