RESUMEN
BACKGROUND: Even though the perioperative chemotherapy improves the overall survival (OS) compared to surgery alone in patients with a resectable gastroesophageal adenocarcinoma (GEA), prognosis of these patients remains poor. Docetaxel (D), cisplatin (C), and 5-fluorouracil (F) regimen improves OS compared to CF among patients with advanced GEA. We evaluated the potential interest of a perioperative DCF regimen, compared to standard (S) regimens, in resectable GEA patients. METHODS: We identified 459 patients treated with preoperative DCF or S regimens. The primary endpoint was OS. Propensity scores were estimated with a logistic regression model in which all baseline covariates were included. We then used two methods to take PS into account and thus make DCF and S patients comparable. OS analyses were performed with Kaplan-Meier and Cox models in propensity score matched samples, and inverse probability of treatment weighted (IPTW) samples. RESULTS: In the propensity score matched sample, the p-value from the log rank test for OS was 0.0961, and the 3-year OS rate was 73% and 55% in DCF and S groups, respectively. The multivariate Cox regression underlined a Hazard Ratio of 0.55 (95% CI 0.27-1.13) for DCF patients compared to S patients. The results from IPTW analyses showed that DCF was significantly and independently associated with OS (HR = 0.52; 95% CI 0.40-0.69). CONCLUSIONS: In this retrospective multicenter, hypothesis-generating study, the propensity score analyses underlined encouraging results in favor of DCF compared to S regimens regarding OS. This promising result should be validated in a phase-3 trial.