Pulmonary dose-volume predictors of radiation pneumonitis following stereotactic body radiation therapy.

Abstract

Radiation pneumonitis (RP) may be severe after stereotactic body radiation therapy. Our purpose was to identify pulmonary and cardiac dosimetric parameters that predicted for post-stereotactic body radiation therapy grade ≥2 RP. A total of 335 patients with ≥3 months' follow-up were included. Normal pulmonary volume was total lungs minus gross tumor volume. Pulmonary maximum dose, mean lung dose (MLD), and the percent of lung receiving ≥x Gy for 5 to 50 Gy in 5-Gy increments were collected. Cardiac maximum dose, mean dose, volume of lung receiving ≥0.1 Gy (V0.1), V0.25 to V1, and V2.5 to V12.5 were recorded. Multivariable logistic regression with manual backward stepwise elimination was used to identify the best dosimetric predictors of toxicity. Optimal dose-volume cutoffs were isolated with recursive partitioning analysis (RPA). The grade ≥2 RP rate was 18.8%. Pulmonary V5 to V50, MLD, and cardiac V0.1 to V2.5 were significantly associated with toxicity on univariate analysis. On multivariable logistic regression, V10 was the strongest dosimetric predictor of grade ≥2 RP (odds ratio, 1.052; 95% confidence interval, 1.014-1.092; P = .007). RPA identified a 21.6% risk of grade ≥2 RP with V10 ≥6.14% (vs 3.8% with <6.14). MLD was the most significant predictor of grade ≥3 RP (odds ratio, 1.002; 95% confidence interval, 1.000-1.003; P = .031). RPA identified a 25.0% risk of grade ≥3 RP with MLD ≥7.84 Gy (vs 8.0% when <7.84 Gy). With a grade ≥2 RP rate of 18.8%, lung V10 was the best predictor of grade ≥2 toxicity. MLD was the best predictor of grade ≥3 RP.