A Comparison of Trimodality Therapy Versus Definitive Concurrent Chemoradiation in Patients With Stage IIIA Non–small Cell Lung Cancer

Abstract

The optimal treatment paradigm for pts with stage IIIA NSCLC with mediastinal (N2) involvement at diagnosis (dx) has not been established. Two common approaches include TM with preoperative chemotherapy, surgical resection and post-operative radiation vs. CRT. At our institution, the preferred approach for resectable pts is TM. Pts who are not surgical candidates undergo CRT. We retrospectively analyzed 267 pts treated with TM (N=135) or CRT (N=132) from 01/2004-07/2016. Extensive mediastinal involvement (EMI) was defined as > 2 mediastinal or hilar nodal stations involved at dx. Estimates for the endpoints of overall survival (OS), progression-free survival (PFS), distant metastasis free survival (DMFS) and local-regional control (LRC) were generated from the time of dx using the Kaplan-Meier method and compared using the log-rank test. LRC was assessed independently of distant control. Cox proportional hazards models were used for univariate (UVA) and multivariable analyses (MVA). The median follow-up time was 23.9 months for all pts and 46.2 months for survivors. Age (median 65 years), race (83% white), sex (43% male), and KPS (median 80) were similar between the 2 groups. More pts treated with CRT were current smokers (32% vs. 8%, p<0.01) with larger tumors (4.5 cm vs. 3.7 cm, p<0.01), higher T-stage (34% with T3 or T4 vs. 18%, p <0.01), squamous cell carcinoma histology (37.1% vs. 12.7%, p<0.01), subcarinal nodal involvement (52% vs. 33%, p<0.01) and EMI (31% vs. 14%, p<0.01). Pts in the TM group had R0 (79%), R1/2 (21%) resections. The median radiation dose was 54 Gy (range: 50-70 Gy) in the TM group and 64.8 Gy (range: 54-74 Gy) in the CRT group. Pts treated with TM had superior OS (2 yr: 69% vs. 46%, p<0.01), PFS (2 yr: 44% vs. 29%, p=0.04), and LRC (2 yr: 76% vs. 56%, p<0.01), but similar DMFS (2 yr: 46% vs. 41%, p=0.55). In patients with subcarinal involvement, TM no longer resulted in superior OS (64% vs. 56%, p=0.70), PFS (38% vs. 34%, =0.72), or LRC (75% vs. 58%, p=0.15). Similarly, in pts with EMI involvement, TM no longer resulted in superior OS (p=0.85), PFS (p=0.75) or LRC (p=0.92). On UVA, age, adenocarcinoma histology (AH) and TM were predictive of LRC. On MVA, age (1.03; 1.01-1.05) and AH (0.48; 0.30-0.80) were predictive of LRC. On MVA (HR; 95%CI), age (1.02; 1.01-1.04), AH (0.62; 0.42-0.91) and TM (0.64; 0.45-0.91) were predictive of OS. There were no predictive factors on MVA for DMFS or PFS. Sex, race, KPS, T-stage, tumor size, smoking status, nodal location, EMI and subcarinal involvement were not predictive. The superior OS and PFS in the TM pts may be attributable to a more fit patient population with less extensive disease and without progression after induction chemotherapy. TM and CRT resulted in similar rates of DMFS. CRT can be considered in pts with subcarinal disease or EMI to spare the morbidity of surgery. Prospective trials to control for pt and disease related confounders are needed to determine the best approach for pts with N2 Stage IIIA NSCLC.