Pretreatment CT Based Emphysema and Fibrosis Scoring of Peri-Tumoral Lung Parenchyma Predicts Risk of Radiation Induced Lung Toxicity

Abstract

Many patients receiving radiation (RT) for localized non-small cell lung cancer (NSCLC) harbor underlying pulmonary dysfunction secondary to emphysema, pulmonary fibrosis, or both. However, it is unclear whether this dysfunction leads to higher rates of radiation induced lung toxicity (RILT). We sought to assess whether baseline or change in baseline to mid-treatment spirometry based pulmonary function testing (PFTs) and/or pretreatment CT graded emphysema (EM) or pulmonary fibrosis (PF) scores were associated with RILT. We performed an institutional review board approved analysis of patients with localized NSCLC treated with non-SBRT external beam RT on 4 consecutive prospective trials from 2004-2013. The primary outcome was grade ≥2 RILT including pneumonitis and fibrosis. PFTs consisted of FEV1, FVC, FEV1/FVC, and DLCO. For baseline CT EM and PF grading, the lungs were divided into 6 zones. EM was scored from 0-4 and PF from 0-3, using existing clinical scoring schemes. Global EM and PF scores were created by summing the respective scores in each of the 6 defined zones. Regional tumor EM and PF scores were created using the highest score for each in the lung zone containing gross target volume. Multivariable analysis was performed with Cox-proportional hazard models. A total of 96 patients were available for analysis. Seventy-eight percent of patients received concurrent chemotherapy. Mean mean lung dose (MLD) was 14 Gy (2 Gy equivalent). Median FEV1/FVC and DLCO were 64 (interquartile range [IQR] 52-71) and 60 (IQR 49-74), respectively. Median global and regional tumor CT based EM scores were 6 (range 0-20) and 1 (range 0-3). Median global and regional tumor CT based PF scores were 0 (range 0-8) and 0 (range 0-3). 20 patients developed RILT with a 20 month freedom from RILT of 75%. 51 patients had mid treatment PFTs. PFTs changed minimally from baseline to mid-treatment, with no difference in changes when comparing patients with and without RILT. PFTs were not associated with RILT (all p >0.1). Global and regional tumor baseline CT EM and PF scores were individually associated with RILT when controlling for patient age, MLD, smoking status, and receipt of chemotherapy (all p<0.05). When controlling for EM and PF scores jointly, global scores no longer remained statistically significant (p>0.05), whereas regional tumor scores did (EM HR: 1.7, 95% CI: 1.1-2.8, p=0.03, PF HR: 2.2, 95% CI: 1.1-4.4, p=0.04). The severity of emphysema and fibrosis adjacent to tumor on pre-treatment CT imaging is associated with development of RILT, whereas global CT scores and PFTs are not. These findings suggest that peri-tumoral regional lung function and anatomy predict for risk of grade ≥2 RILT. This easily obtainable radiographic information combined with other clinical/dosimetric factors can help guide individualization of radiation treatment delivery.