[Cisplatin, 5-FU and preoperative radiotherapy in esophageal epidermoid cancer. Multicenter phase II FFCD 8804 study]

The aim of this study was to analyze the efficacy and tolerance of preoperative radiochemotherapy in esophageal squamous cell carcinoma. Patterns of recurrence and prognostic factors were also studied.

This multicentric phase II trial included patients deemed operable. Preoperative treatment associated 5-FU 800 mg/m2/d by continuous infusion, cisplatin 20 mg/m2/d and radiotherapy 3 Gy/d D 1-5 and D 22-26. Resection was planned 50 to 60 days after the beginning of therapy.

Seventeen centers accrued 96 patients, mean age 55.4 years. According to UICC 1978 classification: stage I 13%, stage II 53% and stage III 30%. Mean follow-up was 73 months. Pre-operative treatment was delivered at full dose in half of the patients. Ten percent of the patients did not receive the second cycle. Toxicity reached grade 3 in 23% and grade 4 in 7% of the patients. Two preoperative deaths occurred. Curative resection was performed in 82% of the patients. Operative mortality was 9%. tumors were sterilized on the operative specimen in 20% of the patients and microscopic remnants were observed in 13%. Preoperative clinical work-up overestimated histologic response in 10% of the cases underestimated it in 29%. After 2 years, a recurrence was observed in 56% of the patients: loco-regional in 30%, metastases in 19% and both in 7%. Median survival was 17 months and survival rates were 58% at 1 year and 25% at 5 years. Four prognostic factors influenced survival in multivariate analysis (Cox model): hematological toxicity grade 3 or 4, no complete response, circumferential extension > 2/3 and nodes visible on CT-scan. Factors positively influencing complete response were in multivariate analysis: a fungating tumor, weight loss < 8% and a full dose preoperative treatment.

In this study, preoperative treatment associating 5-FU-cisplatin and radiotherapy allowed a high resectability rate, with one third of patients achieving complete or nearly complete histologic response. A randomized study is warranted to know whether this combined treatment is better than surgery alone.