Intensity-modulated versus 3-dimensional conformal radiotherapy in the definitive treatment of esophageal cancer: comparison of outcomes and acute toxicity

Matthias Felix Haefner, Kristin Lang, Vivek Verma, Stefan Alexander Koerber, Lorenz Uhlmann, Juergen Debus, Florian Sterzing

Though the vast majority of seminal trials for locally advanced esophageal cancer (EC) utilized three-dimensional conformal radiotherapy (3DCRT), the advanced and highly conformal technology known as intensity-modulated radiotherapy (IMRT) can decrease doses to critical cardiopulmonary organs. To date, there have been no studies comparing both modalities as part of definitive chemoradiation (dCRT) for EC. Herein, we investigated local control and survival and evaluated clinical factors associated with these endpoints between cohorts. We retrospectively analyzed 93 patients (3DCRT n = 49, IMRT n = 44) who received dCRT at our institution between 2000 and 2012 with the histologic diagnosis of nonmetastatic EC, a Karnofsky performance status of ≥70, curative treatment intent, and receipt of concomitant CRT. Patients were excluded if receiving <50 Gy. Kaplan-Meier analysis was used to evaluate the endpoints of local relapse rate (LR), progression-free survival (PFS), and overall survival (OS). Cox proportional hazards modeling addressed factors associated with outcomes with univariate and multivariate approaches. Rates of acute toxicities and basic dosimetric parameters were compared between 3DCRT and IMRT patients. Mean follow-up was 34.7 months. The 3-year LR was 28.6% in the 3DCRT group and 22.7% in the IMRT group (p = 0.620). Median PFS were 13.8 and 16.6 months, respectively (p = 0.448). Median OS were 18.4 and 42.0 months, respectively (p = 0.198). On univariate analysis, only cumulative radiation dose was associated with superior LR (hazard ratio (HR) 0.736; 95% confidence interval (CI) 0.635 - 0.916, p = 0.004). Factors clearly affecting survival were not observed. When comparing 3DCRT- versus IMRT-based dCRT, no survival benefits were observed. However, we found a lower local recurrence rate in the IMRT group potentially owing to dose-escalation. Prospective data are needed to verify the presented results herein.