(OA04) Patterns of Failure After Postoperative Radiation Therapy for Locally Advanced NSCLC: Implications for Shifting Toward More Conformal Radiation Fields

Abstract

Over the past decade, postoperative radiation therapy (PORT) volumes in locally advanced NSCLC have largely shifted from whole mediastinum to conformal fields as specified in the phase III Lung Adjuvant Radiotherapy Trial (Lung ART) trial. However, limited data is available on how this transition will affect locoregional recurrence rates (LRR) and overall survival (OS). We thus compared patterns of failure, survival outcomes, and toxicity with whole mediastinum (WM) radiation therapy vs. high-risk (HR) conformal fields in the postoperative setting. In this retrospective analysis, we identified the records of 200 consecutive patients with pT1-4 N1-2 NSCLC who were treated with PORT between July 1998 and November 2014. Medical records were reviewed for demographic, tumor- and treatment-specific, and outcome data. Patients underwent surgical resection with wedge resection (n=24, 12%), lobectomy (n=148, 74%), or pneumonectomy (n=28, 14%). The predominant T and N-stage was pT2 (n = 92, 46%) and pN2 (n = 159, 79.5%). PORT was delivered to a median dose of 50 Gy in 25 fractions, and patients were stratified into two groups according to PORT field size: WM (n=130, 65%) or HR (n=70, 35%) nodal regions (involved or involved plus adjacent regions). The most common treatment modality in the WM group was 2D (n=56, 43%) whereas IMRT (n=31, 41%) was utilized most frequently in the HR group. LRR was defined as failure in either the: a) bronchial stump/anastomosis, b) treated hilum, or c) mediastinum (levels 2-9) or supraclavicular region, treated or untreated. Patients with prior distant metastatic failure were censored at that time and were not defined as having LRR. The Kaplan-Meier method was used to calculate locoregional recurrence-free (LRRFS) and OS from the end of PORT. Comparisons between patient and treatment characteristics were made using log-rank tests. Univariate and multivariable analyses were calculated using Cox proportional hazards modeling. Median follow-up among patients treated with HR fields was 40.1 months (range, <1 month-167.2 months) and for WM patients was 33.6 months (range, <1 month-199.3 months). Median OS and LRRFS were not significant between groups but trended toward improvement with HR (48.6 months for HR and 28.3 months for WM and, p=0.11; 39.1 months for HR vs. 23.0 months for WM, p=0.20, respectively). However, on multivariable analysis, field size was not associated with significant differences in LRR, LRRFS, or OS (p=0.484, 0.146, and 0.196). Grade ≥2 toxicity was lower in the HR group (OR 0.32 [95% CI 0.14-0.77], p=0.010) and higher in patients who received concurrent chemotherapy (OR 5.45 [95% CI 1.55-19.19], p=0.008). Despite transitioning to more limited fields for PORT, LRRPFS and OS were not compromised. Toxicity was lower in patients treated with HR fields. This data supports the use of more conformal fields as per Lung ART in the setting of PORT.