Pre-Treatment Immune-Related Markers Predict Disease Outcomes in Non-Small Cell Lung Cancer Patients Treated With Chemoradiation and Durvalumab

Abstract

Immune-related markers from the peripheral blood and radiation dose to the immune system have been associated with outcomes in locally-advanced non-small cell lung cancers (LA-NSCLC) patients treated with concurrent chemoradiation (cCRT). We assessed if the neutrophil-to-lymphocyte ratio (NLR), and risk scores from two dose models (Estimated Dose of Radiation to Immune Cells (EDRIC) and the RTOG 0617 overall survival (OS) model including cardiopulmonary dose) are also associated with outcomes in patients treated with cCRT and durvalumab.Consecutive LA-NSCLC patients treated with platinum-based cCRT and at least one dose of durvalumab between November 2017 and December 2019 were evaluated. The pre-cCRT and pre-durvalumab NLR were calculated using samples most proximal to RT and durvalumab start (NLRpre-cCRT, NLRDurvalumab). Mean body dose, mean heart dose and mean lung dose (MLD) were extracted to calculate EDRIC; atria D45, mean of the hottest 5% of the ventricle dose (MOH5), pericardium MOH55 and MLD were used to calculate OS risk scores from the 0617 model. Progression-free survival (PFS) was defined from the start of cCRT. NLRpre-cCRT, NLRDurvalumab, risk scores from the two dose models, and patient characteristics (stage, age, PD-L1, ECOG status) were associated with PFS using bootstrapped Cox proportional hazards modeling. A Bonferroni-corrected P≤0.006 was considered statistically significant.A total of 105 patients with a median follow-up of 20 months (IQR: 13-27) were included. Median age was 67 years, 60% were male, and 45% were ECOG 1. Most patients had adenocarcinoma histology (n = 64) and stage IIIB/IIIC disease (n = 71). PD-L1 expression was available for 75% of patients, of which 66% had PD-L1 expression ≥1%. Patients were treated with a median of 7 (IQR: 2-12) months of durvalumab that started a median of 7 (5-9) weeks from end of cCRT. Twenty-eight (27%) patients discontinued durvalumab due to toxicity after a median of 2 (IQR: 1-4) months. In total, 41 patients had a progression event at a median of 8 (IQR: 6-11) months. The 12- and 18-months PFS estimates were 69% (95% CI: 60-78%) and 62% (52-72%), respectively. The median NLRpre-cCRT was 3.4 (IQR: 2.2-4.7), NLRDurvalumab was 5.5 (IQR: 3.6-7.4). The median risk score for EDRIC was 5.3 (IQR: 3.9-6.6) and 0.82 (IQR: 0.62-1.1) for the 0617 model. Among the investigated variables only NLRpre-cCRT reached significance in predicting PFS (HR: 1.10, 95% CI 1.05-1.16, P = 0.0002). Further, NLRpre-cCRT was not strongly correlated with age, stage, PD-L1 expression or ECOG status (Spearman's rho range: -0.22, 0.22).This study represents the first effort focusing on immune-related markers and dose models in LA-NSCLC treated with cCRT and durvalumab and finds the pre-cCRT NLR as the best predictor of disease outcomes.