Validation of RTOG 0813 Proximal Bronchial Tree Constraints for Pulmonary Toxicity With Stereotactic Body Radiation Therapy for Central Non-small Cell Lung Cancer

Abstract

Purpose: Clinical validation of protocol-specified dosimetric constraints for the proximal bronchial tree (PBT) is limited for central non-small cell lung cancer treated with stereotactic body radiation therapy. We sought to validate Radiation Therapy Oncology Group (RTOG) PBT constraints with a large institutional data set.

Methods and materials: Lesions ≤2 cm from the PBT treated with definitive stereotactic body radiation therapy from 2009 to 2016 were identified from a prospective registry of 1462 patients. Every PBT dose and volume combination, ranging from 0 cGy to 8000 cGy in increments of 10 cGy and volumes ranging from 0.03 cm³ to 50 cm³ in increments of 0.03 cm³, was analyzed. The sensitivity and specificity of these endpoints for identifying pulmonary toxicity were calculated. Pulmonary toxicity was classified as pneumonitis or nonpneumonitis toxicity (NPT) (fistula, stenosis, necrosis, hemoptysis, clinically significant pleural effusion). The optimal dosimetric predictor was chosen by calculation of F-score (highest sensitivity and specificity).

Results: The study included 132 patients, with 26.0-month median follow-up. Eight grade ≥2 NPT (2 grade 5) and 8 grade 2 pneumonitis toxicities were observed. The PBT dosimetric endpoint with the highest F-score for identification of grade 2 to 5 NPT was D0.03cc ≤5000 cGy and that for grade 3 to 5 NPT was D0.33cc ≤4710 cGy, with sensitivity and specificity of 87.5% and 76.6% and 100.0% and 85.7%, respectively. Applying the RTOG 0813 PBT constraints to our data set achieved a sensitivity and specificity of 33.3% and 92.1% for D4cc ≤1800 cGy and 37.5% and 92.7% for D0.03cc ≤5250 cGy for identification of grade 2 to 5 NPT. A PBT dosimetric correlation for pneumonitis toxicity could not be identified.

Conclusions: This novel dosimetric analysis validates current RTOG constraints and emphasizes high-dose, small-volume constraints as better predictors for NPT. We demonstrated that a slightly lower maximum point dose PBT constraint may be optimal for identification of NPT. Validation of these findings in a larger cohort of patients with longer follow-up is necessary.