3639 - AI-Derived Thymic Health, Thymic Dose, and Toxicity Outcomes Among Older Adults with Lung Cancer Undergoing Chemoradiation
Presenter(s)
L. L. Thompson1,2, V. Prudente1,2, S. Bernatz1,2, S. Pai1,3, A. T. Gregg4, S. B. Shah4, J. Yoon4, N. A. Saeed2, F. K. Keane4,5, R. H. Mak1,2, and H. Aerts1,3; 1Artificial Intelligence in Medicine (AIM) Program, Mass General Brigham, Harvard Medical School, Boston, MA, 2Department of Radiation Oncology, Brigham and Women's Hospital and Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, 3Department of Radiation Oncology, Mass General Brigham/Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, 4Harvard Medical School, Boston, MA, 5Department of Radiation Oncology, Massachusetts General Hospital, Boston, MA
Purpose/Objective(s): Older adults with locally-advanced non-small cell lung cancer (LA-NSCLC) experience substantial chemoradiation-related toxicity, but predictive models incorporating immune factors remain limited. We examined associations between a validated artificial intelligence (AI)-based measure of thymic health, thymic radiation (RT) dose, and risk of treatment-related esophagitis/pneumonitis in this population.
Materials/Methods: We pooled LA-NSCLC patients aged = 65 who completed chemoradiation at our institution 1998-2014 (n=365) or in RTOG 0617 2007-2011(n=178). We collected clinical features, treatment regimens, and rates of grade 2+ (CTCAE v3.0) esophagitis/pneumonitis. AI thymic health scores were computed from pre-treatment CTs (range 0–1, higher=better health). Given T-cell reliance on thymic function and involvement in these toxicities, we grouped patients by factors expected to compromise thymic reserve. Specifically, patients were divided by their pre-radiation thymic health and thymic radiation dose into two groups: high function [n=69, defined as a pre-RT thymic health = 75th percentile [p75] & dose <p75) and low function [n=474, defined as a pre-RT thymic health p75 or dose = p75). We assessed associations between health/dose grouping and 1) esophagitis, and 2) pneumonitis using Fine-Gray regressions with consideration of death as competing risk. Models were adjusted for age, sex, stage, treatment regimen, radiation dose, planning-target-volume, organ metrics (esophagus V60 or lung V20), and other significant covariates (univariate p<0.10).
Results: Overall, 543 patients (median age 72 years, 47.7% female, 91% stage III) completed therapy. Median lung V20 was 27.9% (IQR 21.8-33.9%). Median esophagus V60 was 10.1% (IQR 0.0-25.2%). Median thymic health was 0.15 (IQR 0.02-0.28). Median mean thymic dose was 46.3 Gy (IQR 35.5-56.3 Gy). At a median follow-up of 20.1 months, rates of grade 2+ esophagitis and pneumonitis were 39.58%/11.6%, respectively. In adjusted models, the high function subgroup had increased risk of esophagitis (HR=1.86, [95%CI:1.32-2.63], p<0.001) and pneumonitis (SHR=4.33, [95% CI 1.18-15.82], p=0.027).
Conclusion: Baseline thymic health may influence risk of treatment-related side effects. Consideration of AI-based thymic metrics may improve toxicity prediction in older NSCLC patients.