The extent of residual carcinoma in resected esophageal cancer specimens after preoperative chemoradiation is a predictor of survival according to 3 groups: P0 (0% residual carcinoma), P1 (1% to 50% residual carcinoma), and P2 (>50% residual carcinoma). However, the interobserver variation and reliability of this classification has not been evaluated among different pathologists. Histologic hematoxylin and eosin-stained slides from 60 coded cases of esophageal adenocarcinomas treated with preoperative chemoradiation followed by esophagectomy were independently reviewed by 6 pathologists from 4 different institutions for extent of residual carcinoma and ypT stage. Interobserver agreement was analyzed by kappa (κ) statistics, and grading of extent of residual carcinoma and ypT stage based on individual and consensus gradings were correlated with patients' survival. The interobserver agreement was excellent for extent of residual carcinoma (κ=0.84, Kendall's W=0.92; P<0.000001) and was good for ypT stage (κ=0.71, Kendall's W=0.88; P<0.000001). Agreement was excellent for all categories of residual carcinoma: P0 (κ=0.87), P1 (κ=0.81), and P2 (κ=0.85). Extent of residual carcinoma was a significant predictor for overall survival based on individual pathologist grading (all P values <0.03), and consensus grading (13 P0, 32 P1, and 15 P2), P=0.004; independent of ypT and ypN stages in multivariate analysis (P=0.02). Our results indicate that after preoperative chemoradiation in esophageal cancer there is excellent interobserver agreement among pathologists on grading extent of residual carcinoma. The extent of residual carcinoma is a reliable and reproducible predictor of survival; this grading system may allow a novel and early means of comparing outcomes after different neoadjuvant treatment regimens.