CALGB 9380: A prospective trial of the feasibility of thoracoscopy/laparoscopy in staging esophageal cancer

Background. The staging of esophageal cancer is imprecise. Thoracoscopic/la paroscopic (TS/LS) staging has been proposed as a more accurate lymph node (LN) staging method. We report the experience of an Intergroup NCI trial (C ALGB 9380) evaluating the feasibility and accuracy of this staging modality . Patients and Methods. From February 1995 to September 1999, 134 patients we re entered in the study. This study represents the analysis of final data o n 113 patients. TS/LS was considered feasible if TS and 1 LN sampled at lea st 3 LN by LS; a confirmed positive node was found; or T4 or M1 disease was documented. If this was accomplished in more than 70% of patients, TS/LS w as believed to be feasible. Results. The LN stations most frequently sampled in the thorax (134 patient s) were levels 2 (33%), 3 (38%), 4 (40%), 7 (76%), 8 (69%), 9 (55%), and 10 (43%) and in the abdomen levels 17 (70%) and 20 (55%). The frequency of po sitive LN by level were as follows: 2 (10%), 3 (8%), 4 (10%), 7 (10%), 8 (2 5%), 9 (10%), 10 (10%), 17 (34%), and 20 (27%). Noninvasive tests (computed tomographic scan, magnetic resonance imaging, esophageal ultrasound scan) each incorrectly identified TN staging as noted by missed positive or false -negative LN or metastatic disease found at TS/LS staging in 50%, 40%, and 30% of patients, respectively. Median operating time was 210 minutes (range , 40 to 865 minutes). Median postoperative hospital stay was 3 days (range, 1 to 35 days). There were no deaths or major complications. Seventy-three percent of patients met the definition for feasibility. In 30 patients TS w as not feasible. Positive LN disease was found in 43 patients; 32 were deem ed NO. Ten patients had T4/M1 disease. Of the 32 potentially resectable NO patients, 14 patients had preoperative induction therapy; 13 patients went directly to operation with NO confirmed in 9 patients, NX in 1 and N1 in 3. Three patients were unresectable, 1 patient died, and 1 was lost to follow -up. Conclusions. In summary, the feasibility of TS/LS was confirmed. It doubled the number of positive LNs identified by conventional, noninvasive staging . The overall accuracy remains to be defined by analysis of the LN negative group in follow-up. Although the positive predictive value was high, furth er study is warranted to confirm the role of TS/LS in the staging algorithm of esophageal cancer. (C) 2001 by The Society of Thoracic Surgeons.