Predictors of Local Control and Symptomatic Pneumonitis after Lung SBRT

Abstract

Stereotactic body radiation therapy (SBRT) is commonly used to treat primary or oligometastatic malignancies in the lung, though most of the available data describing the safety and efficacy of SBRT is for smaller tumors. The purpose of this study was to evaluate the impact of tumor size, among other factors, on local control (LC) and radiation pneumonitis (RP) in patients who received lung SBRT. This retrospective study included patients with primary and metastatic lung tumors treated with SBRT at a single institution from 2012 - 2017. Patients were excluded for prior radiation to the targeted site. All patients underwent 4D-CT simulation with respiratory motion management as needed, and received 3-5 fractions of varying doses prescribed to the PTV on a linear accelerator using a highly conformal IMRT or VMAT technique. Factors evaluated for association with LC included GTV diameter and volume, PTV diameter and volume, volumetric ratio of PTV:GTV, histology, prescription biologically effective dose (BED10), and PTV maximum dose (Dmax). Additional factors evaluated for association with grade 2 or higher (symptomatic) RP included BED3, conformity index (CI), max and mean dose to the heart and central airway, mean lung dose (MLD), and the volume of lung that received at least 2.5 Gy (V2.5), 5 Gy (V5), 13 Gy (V13), 20 Gy (V20), 30 Gy (V30), 40 Gy (V40), and 50 Gy (V50). Analysis included the Mann-Whitney U-test and binary logistic regression. Recursive partitioning analysis was used to determine the optimal dose constraint (OC) predictive of RP. A total of 125 patients with 150 irradiated sites, consisting of 104 (69.3%) primary lung tumors and 46 (30.7%) metastatic lung tumors were included in the study. The median follow-up for survival/toxicity analysis and LC analysis were 13.0 months (IQR 7.0 - 26.0) and 11.5 months (IQR 6.0 – 20.0), respectively. A large range of tumor sizes were represented with a median GTV volume of 6.02 cm3 (IQR 2.19 – 13.79). LC rates at 12 and 24 months post-treatment were 92.6% and 85.9%, respectively. LC was associated with higher BED10 (median 112.5 Gy vs. 92.8 Gy, p=0.02), PTV Dmax (62.5 Gy vs. 53.1 Gy, p<0.01), and PTV:GTV (5.8 vs. 3.5, p=0.01), and there was a trend towards improved LC with smaller tumor diameter (2.8 cm vs. 4.0 cm, p=0.08). RP was associated with higher ipsilateral lung MLD (7.3 Gy vs. 5.3 Gy, p=0.02, OC=8.6 Gy), V2.5 (38.8% vs. 32.0%, p=0.03, OC=30.8%), V5 (32.9% vs. 24.9%, p=0.02, OC=26.8%), V13 (18.8% vs. 13.4%, p=0.03, OC=21.8%), V20 (11.7% vs. 8.0%, p=0.05, OC=14.9%), and heart Dmax (14.4 Gy vs. 6.1 Gy, p=0.02, OC=31.7 Gy). A higher PTV prescription dose, greater dose heterogeneity within the target, and larger PTV margins were significantly associated with improved LC, whereas a weaker association was observed between tumor size and LC. Ipsilateral lung doses were most correlated with RP.