Patterns of locoregional failure in stage III non-small cell lung cancer treated with definitive chemoradiation therapy

Purpose: Chemoradiation therapy (CRT) is the core treatment of locally advanced non-small cell lung cancer (LA-NSCLC), but potential toxicities limit radiation therapy dose. These toxicities, plus the advent of increasingly conformal radiation therapy, have prioritized target definition and the use of involved-field radiation therapy (IFRT). Published data largely focus on regional rather than local failure patterns. We report our pattern-of-failure experience treating patients with LA-NSCLC with definitive CRT, focusing on both local and regional recurrences with detailed dosimetric analyses of failure location. Methods and materials: Patients treated between December 2004-2010 were included. Imaging scans from date of failure were fused with the RT-planning CT scan, and recurrent nodes were contoured to determine if the recurrence was in a previously irradiated region, defined as involved nodal recurrence (INR) versus elective nodal recurrence (ENR). Local failures were contoured and identified as in-field, marginal, or out-of-field based on dose received. Actuarial overall survival (OS) and progression-free survival (PFS) were calculated, and the cumulative incidences of local, regional, locoregional, and distant recurrence (CILR, CIRR, CILRR, CIDR) were determined with death as a competing risk. Results: One hundred five patients were included with a median survival of 21.8 months. The 3-year OS and PFS were 36% and 22%, respectively. The 3 year CILRR, CILR, CIRR, CIDR were 41%, 38%, 40%, and 58%, respectively. Thirty patients failed regionally, but only 7 patients developed an ENR with no concurrent local failure or INR, and only 1 of these patients did not develop distant metastases within 1 month of recurrence. A total of 21 patients (20%) developed an ENR with or without other areas of recurrence. Conclusions: Elective regional recurrences rarely occurred as the sole site of failure, despite the use of IFRT. Moreover, the pattern of local failure was entirely in-field. These data strongly support field design focusing on gross nodal and primary disease.