External Validation of Dose Patterns and Dosimetric Predictors for Radiation-Induced Esophagitis In Non-Small Cell...

Area of Science:

• Radiation oncology
• Medical physics
• Clinical oncology

Background:

• Radiation-induced esophagitis (RE) is a significant complication of chemoradiotherapy for non-small cell lung cancer (NSCLC).
• Accurate prediction of severe RE (grade ≥ 2, RE2+) is crucial for treatment planning and toxicity reduction.

Purpose of the Study:

• To externally validate dose patterns and dosimetric predictors of RE2+ in an independent NSCLC cohort.
• To assess the clinical utility of voxel-based analysis (VBA) and specific dosimetric parameters for predicting RE2+.

Main Methods:

• Analysis of patient data from the Radiation Therapy Oncology Group-0617 trial, including concurrent chemotherapy and varying radiation doses.
• Voxel-based analysis (VBA) to identify spatial dose-response relationships for RE2+.
• Generalized linear models and nonparametric permutation tests were used to assess significance and external validity of predictors.

Main Results:

• Voxel-based analysis identified the upper-middle thoracic esophageal (UME) segment as a critical region for RE2+ risk.
• The relative esophageal volume receiving ≥ 55 Gy (V55Gy) and UME mean dose (Dmean) were confirmed as significant predictors.
• UME-Dmean demonstrated superior risk prediction accuracy and calibration compared to V55Gy.

Conclusions:

• Voxel-based analysis findings for RE prediction were successfully validated in an independent NSCLC cohort.
• The upper-middle esophagus is a critical region for RE2+ risk, and UME-Dmean is a robust dosimetric predictor.
• These results support the clinical implementation of VBA for optimizing radiation therapy in NSCLC patients.