2448 - Identification of Dosimetric and Clinical Predictors for Post-Treatment Pulmonary Hospitalizations in NSCLC

Purpose/Objective(s): Pulmonary events following radiation therapy (RT) for non-small cell lung cancer (NSCLC) can lead to morbidity and negatively affect outcomes. This study evaluates the predictive value of dose-volume parameters (mean lung dose [MLD], V5–V60) and clinical factors for post-treatment pulmonary hospitalizations. We seek to improve risk stratification and develop more comprehensive clinical decision-making tools.

Materials/Methods: Data were analyzed from a prospective statewide quality consortium. Eligible patients (2018-2024) included those with stage I-III NSCLC treated with definitive RT with available dosimetric data. Patients receiving surgery were excluded. Pulmonary hospitalizations were classified as pneumonitis, COPD, pneumonia, or other respiratory complications based on clinical diagnosis, with multiple classifications allowed for overlapping conditions. EQD2 dose distributions were calculated (a/ß = 3). Multivariable Cox proportional hazards models were used to identify predictors of hospitalization, and cumulative incidence estimates were calculated, accounting for competing risks of death. Variables were selected using a stepwise procedure, and interactions were tested for significance.

Results: Of 1407 patients, 310(19%) experienced a lung-related hospitalization. Median follow-up was 14 months. Pneumonitis, pneumonia, and COPD-related hospitalizations occurred in 34(2%), 156(10%), and 135(8%) patients, respectively. The 24-month cumulative incidence of pneumonitis, pneumonia, and COPD-related hospitalizations was 2.7%, 10.6%, and 12.2%, respectively. The cumulative incidence of any lung-related hospitalization was 24.4%. MLD was the strongest predictor of pneumonitis (HR = 1.24 per 1 Gy, p < 0.01) and overall lung-related hospitalizations (HR = 1.04 per 1 Gy, p < 0.01). The estimated 24-month cumulative incidence of lung-related hospitalization increased from 18% at MLD = 0 Gy to 35% at MLD = 20 Gy. Lung dosimetric factors are seen in table below. ECOG performance status (2 vs. 0/1), pre-existing COPD, and supplemental oxygen use were independently associated with increased hospitalization risk.

Conclusion: This analysis confirms that MLD and dose-volume metrics are key predictors of pulmonary hospitalizations in NSCLC. Clinical factors further refine risk assessment. Integrating dosimetric and clinical data can improve risk stratification and guide personalized treatment decisions.