Do dysplastic proximal resection margins predict the risk of anastomotic recurrence and overall survival in patients with esophageal squamous cell carcinoma?

Abstract

Positive proximal resection margins are strongly associated with anastomotic recurrence in esophageal cancer. However, the prognostic significance of dysplastic proximal resection margins remains unclear. The aim of this study is to investigate whether the dysplastic proximal resection margin can predict anastomotic recurrence and overall survival in patients with esophageal squamous cell carcinoma. Between 2000 and 2014, patients with esophageal squamous cell carcinoma who received a nonpalliative resection and survived the perioperative period were included. Two expert pathologists independently reviewed the proximal resection margin status, which was classified as negative, dysplastic, or positive. The kappa statistic was used to test interobserver reliability. Anastomotic recurrence and overall survival served as the main outcome measures. The study cohort consisted of 469 patients (445 males and 27 females). There was an excellent interobserver agreement for negative (kappa=0.88), dysplastic (kappa=0.88), and positive (kappa=1) proximal resection margins-which were identified in 418 (89.1%), 37 (7.9%), and 14 (3.0%) patients, respectively. After a median follow-up of 21.6 months, 30 (6.4%) patients developed an anastomotic recurrence. Compared with patients with negative proximal resection margins (24/418, 5.7%), the occurrence of anastomotic recurrence was more commonly observed in those with positive proximal resection margins (3/14, 21.4%, P=0.017) but not in those with dysplastic proximal resection margins (3/37, 8.1%, P=0.56). Multivariable Cox regression analysis identified positive proximal resection margins (hazard ratio: 5.93, P=0.010) and advanced clinical stage (hazard ratio: 12.04, P=0.023) as independent risk factors for anastomotic recurrence. Dysplastic proximal resection margins were not retained in the model as an independent predictor (hazard ratio: 1.38, P=0.602). The 5-year overall survival rates of patients with negative (38.2%) and dysplastic margins (27.0%) were similar (P=0.814), and significantly higher than that observed in those with positive proximal resection margins (9.5%, P=0.015). In conclusion, dysplastic proximal resection margins can be identified in at least 7.9% of patients with esophageal squamous cell carcinoma, but neither they are associated with an increased risk of anastomotic recurrence nor they portend a poor overall survival.