Patients with Non-response to Neoadjuvant Chemoradiation for Esophageal Cancer Have No Survival Advantage over Patients Undergoing Primary Esophagectomy.

BACKGROUND

Survival for patients with locally advanced esophageal cancer remains dismal. Non-response to neoadjuvant chemoradiation (nCRT) portends worse survival. We hypothesized that patients undergoing up-front esophagectomy may have better survival than those who do not respond to nCRT.

METHODS

We identified all patients undergoing esophagectomy with a pathologic stage of II or greater at our institution between 1994 and 2015 and separated them into two groups: those who received nCRT and those undergoing up-front esophagectomy. The neoadjuvant group was further separated into patients downstaged to pathologic stage 0 or I (responders) and patients with either the same or higher pathologic stage after nCRT, or with pathologic stage II disease or greater (non-responders). Overall survival was compared between groups using Kaplan-Meier statistics. Covariate-adjusted Cox modeling was used to estimate hazard ratios (HR) for mortality associated with non-response.

RESULTS

Overall, 287 patients met inclusion criteria. Fifty-nine percent of the responders had pathologic complete response (pCR). The majority of non-responders and primary esophagectomy patients had stage II or III disease (94%). Median survival was 58.3 months in responders, 23.9 months in non-responders, and 29.1 months in primary esophagectomy patients (p < 0.01). The HR for mortality associated with non-response was 1.82 compared to response to nCRT (p < 0.01) and 1.09 compared to primary esophagectomy (p = 0.71).

CONCLUSIONS

In patients with esophageal cancer who do not respond to nCRT, neoadjuvant therapy may represent a toxic and costly treatment modality that does not improve survival and may delay potentially curative resection. Further research is needed to identify potential non-responders with advanced resectable disease and allow individual tailoring of pre-surgical decision-making.