Patterns of locoregional relapses for patients with esophageal cancer for whom exclusive chemoradiotherapy failed: A plea for dose escalation and elective nodal irradiation.

Abstract

99 Background: Nearly half of the patients (pts) with an esophageal cancer (EC) have a locoregional failure (LRF) after exclusive chemoradiation (eCRT). eCRT delivering 50Gy remains the standard of care for non operable pts. We aim to evaluate the patterns of LRF with respect to planned dose and/or the incidental (unplanned) dose that covered LRF. Methods: Twenty-two pts with EC who failed locally and/or regionally in their follow-up were exclusively reviewed. All the pts have been initially treated (t0) in a curative intent with eCRT. Co-image registration of CT or PET-CT at time of failure and planning CT at t0 was used for image fusion. Each nodal failure (Nf) and each local failure of the primary tumor (Lf) has been outlined, as well as each nodal station (NS) including Nf according to the RTOG classification. Dosimetric parameters in relation with Lf, Nf and involved NS were derived from the initial dosimetric plan. Results: All the patients underwent eCRT including a 5-FU based chemotherapy regimen. Eighteen patients were treated with elective nodal irradiation (ENI) whereas 4 pts did not. The median dose delivered was 50Gy [50Gy-64Gy]. In the follow-up period, 14 pts were in complete response, 3 pts in partial response, 4 pts had a progressive disease (1pt unknown). The median delay between the start of radiotherapy and LRF was 14.3 months [4.27-48.46]. 13 pts had a Lf (included in “planned-dose”), 9 pts had a Nf, 2 pts had a Lf with Nf and 7 pts had a concomittant distant failure. Among pts with Nf, 8 failures were in-field whereas 3 pts had an out-field relapse. Re-calculated doses for NS delineated on the CT performed at t0 were significantly less important than the planned dose (see Table). Conclusions: Our results suggest that an inadequate dose to both the primary tumor and NS could explain high LRF rates observed in EC. A French randomized phase III trial (NCT 01348217) is currently testing a higher dose to the primary tumor and/or ENI with IMRT in an attempt to improve locoregional control. [Table: see text]