Pii: S1010-7940(01)00737-0

Objective: Prediction of responders to induction therapy in esophageal cancer (EC) patients is important. In this study, we evaluated the role of thoracoscopic/laparoscopic (Ts/Ls) staging in prediction of treatment response and survival in EC patients with trimodality treatment. Methods: Retrospective study of EC patients who had undergone Ts/Ls staging and received trimodality treatment at the University of Maryland Medical Center and the Baltimore Veterans Administration Hospitals from July, 1991 to December, 1999. Preoperative therapy consisted of concurrent chemotherapy (5-FU 1 cisplatinum) and radiotherapy. Results: Forty-four EC patients who underwent pretreatment Ts/Ls staging during the study period were able to complete concurrent chemoradiotherapy followed by surgical resection. There were 36 men and 8 women aged 40 to 77 (median age 62). Twenty-seven (61.4%) patients were found to have lymph node metastasis by surgical staging. Fourteen patients (31.8%) had a pathologic complete response. Patients with positive lymph nodes had a lower response rate than those with negative lymph nodes (14.8% vs. 58.8%, P ˆ 0:006). Other clinicopathologic features including gender, weight loss, clinical TNM stage, surgical T stage, and histology did not correlate with treatment response. Univariate analysis showed that weight loss and treatment response were important prognostic factors for disease-free survival (P ˆ 0:01 and P ˆ 0:02, respectively). Histology, surgical N stage and surgical TNM stage appeared to be associated with prognosis (P ˆ 0:067±0.097). Multivariate analysis revealed that only surgical N status and weight loss were signi®cant prognostic factors (P ˆ 0:05, and P ˆ 0:006, respectively). Conclusions: Surgical Ts/Ls staging provides accurate evaluation of tumor spread in EC patients. Pretreatment N status was the single most important predictor of response to induction treatment as well as a reliable prognosticator of survival. q 2001 Elsevier Science B.V. All rights reserved.