Baseline COPD, mean lung dose, and longitudinal exhaled nitric oxide for predicting symptomatic radiation pneumonitis in esophageal squamous cell carcinoma: A prospective single-center pilot study.

120

Background: Symptomatic radiation pneumonitis (RP, ≥Grade 2) is a dose-limiting toxicity in esophageal squamous cell carcinoma (ESCC) patients undergoing thoracic radiotherapy (RT), with 20–25% incidence. Baseline COPD and mean lung dose (MLD) are established predictors, but dynamic non-invasive biomarkers during treatment are lacking. This prospective pilot study evaluated feasibility and incremental value of integrating baseline COPD, MLD, and longitudinal exhaled nitric oxide (eNO) dynamics for RP risk stratification in chemoradiotherapy-treated ESCC. Methods: Fifty-two locally advanced ESCC patients received definitive or neoadjuvant IMRT (50.4 Gy/28 fractions). Baseline COPD was defined per GOLD (FEV₁/FVC <0.70). Lung dosimetry included MLD, V5, V20, V30. eNO was measured at multiple flows (50–200 mL/s) to derive FeNO₅₀, CANO, J’awNO at baseline, weekly during RT, completion, and 4 weeks post-RT. RP (≥Grade 2) was graded within 6 months. Penalized (Firth) logistic regression used three prespecified predictors: COPD, MLD, and early sustained eNO rise (≥2 consecutive weekly increases). Bootstrap validation (1,000 resamples) estimated optimism-corrected AUC and calibration slope. Results: Median age 64 years; 79% male; 54% ever-smokers. COPD in 14 (26.9%). RP ≥Grade 2 in 11 (21.2%); Grade 3 in 2 (3.8%). MLD (OR 1.16/Gy, 95% CI 1.00–1.41), COPD (OR 1.78, 95% CI 0.86–3.85), and early sustained eNO rise (OR 2.12, 95% CI 0.97–4.92) were associated with RP. eNO rise preceded diagnosis by median 17 days. Optimism-corrected AUC: 0.63 (MLD alone), 0.69 (COPD+MLD), 0.74 (full model). Conclusions: Longitudinal eNO monitoring is feasible in ESCC and may modestly improve RP prediction when added to COPD and MLD. Findings are preliminary due to small sample, low events, and no external validation. Larger multicenter studies with independent validation and clinical utility assessment are needed.