Novel endoscopic categorization for prediction of chemoradiotherapy response in locally advanced esophageal cancer

Abstract

Background and aim: Preoperative chemoradiotherapy (CRT) followed by esophagectomy is a well-known treatment modality for patients with locally advanced esophageal cancer (EC). This study developed an algorithm to predict pathological complete response (CR) in these patients using post-CRT endoscopic category with biopsy and validated the proposed algorithm.

Methods: A retrospective review of 141 consecutive patients who completed preoperative CRT and underwent surgical resection for locally advanced EC was performed. The post-CRT endoscopic findings of each patient were stratified into five categories.

Results: The distribution of post-CRT endoscopic categories was significantly different between the pathological CR and non-pathological CR groups (P < 0.001). About 76.8% (73/95) of patients in category 0, 1, or 2 achieved pathological CR. In contrast, 91.3% (42/46) of endoscopic categories 3 and 4 patients did not achieve pathological CR. Sensitivity of post-CRT biopsy was 11.1%. Therefore, an algorithm combining biopsy results and dichotomized post-CRT endoscopic category (category 0, 1, or 2 vs category 3 or 4) was developed. The sensitivity, specificity, and accuracy in predicting pathological CR by the proposed algorithm were 64.8%, 95.9%, and 82.8%, respectively. In the multivariate analysis, the proposed algorithm remained a significant negative factor of survival (P < 0.001).

Conclusions: Algorithm using post-CRT endoscopic category with biopsy may help identify locally advanced EC patients who achieved pathological CR after preoperative CRT. Modalities to accurately detect subepithelial remnant EC may further aid in predicting pathological CR.

Keywords: chemoradiotherapy; esophageal neoplasms; esophagectomy; regression.