Neoadjuvant versus definitive chemoradiotherapy for locally advanced esophageal cancer Outcomes and patterns of failure

Randomized trials examining neoadjuvant chemoradiotherapy followed by surgical resection (nCRT-S) and definitive CRT (dCRT) for esophageal cancer (EC) patients are hampered by use of nonstandard treatment paradigms. Outcomes of nCRT-S versus dCRT in a more common patient population are lacking. We investigated local control and survival, evaluated clinical factors associated with endpoints, and assessed patterns of failure between these cohorts. We retrospectively analyzed 130 patients with locally advanced EC receiving either dCRT or nCRT-S at our institution from 2000-2012. Inclusion criteria were curatively treated nonmetastatic EC, Karnofsky performance status ae70%, and receipt of concomitant CRT. Patients were excluded if receiving < 41 Gy neoadjuvantly or < 50 Gy definitively. Kaplan-Meier analysis was used to evaluate local recurrence (LR), progression-free survival (PFS), and overall survival (OS). Univariate and multivariate Cox proportional hazards modeling addressed factors associated with outcomes. Patterns of failure were enumerated as local, regional, or distant. Mean follow-up was 34.2 months. The 3aEuroyear LR was 10.8% in the nCRT-S group and 21.5% in the dCRT group (p = 0.266). Median PFS were 15.6 and 14.9 months, respectively (p = 0.549). Median OS were 20.6 and 25.9 months, respectively (p = 0.81). On univariate and multivariate analysis, none of the investigated factors was associated with outcomes, although node-positive disease showed a trend for worse OS and PFS. Most common failures in both groups were distant (dCRT 31.2% vs. nCRT-S 21.6%) followed by local in-field recurrences (dCRT 26.9% vs. nCRT-S 10.8%). In this institutional analysis, no significant differences regarding outcomes and patterns of failure were observed between nCRT-S and dCRT.