Methods. Hematoxylin and eosin-stained tumor samples from 138 patients with SCC of the esophagus who underwent potentially curative resection (no residual tumor or distant metastases) were evaluated for the following histologic parameters: degree of keratinization, nuclear polymorphism, pattern of invasion, mitotic activity, and inflammatory response. The prognostic impact of these parameters was analyzed by univariate and multivariate survival analyses.

Results. In the univariate analysis, the inflammatory response (P = 0.0006), pattern of invasion (P = 0.0011), and nuclear polymorphism (P = 0.0161) were the only parameters that correlated with survival. However, in a multivariate survival analysis including these parameters, only pattern of invasion (P = 0.0010) and inflammatory response (P = 0.0076) were prognostically significant. Based on these results, a new prognostic score system was defined that correlated significantly with survival in the univariate survival analysis (P = 0.0002), In contrast, the WHO histologic grade was not prognostically significant. In the multivariate Cox regression analysis, the new prognostic score system proved to be an independent prognostic parameter (P = 0.0062), ranking next to pT classification (P = 0.0001) and pN classification (P = 0.0014).

Conclusions. For SCC of the esophagus, histologic grading based on pattern of invasion and inflammatory response had an independent prognostic impact, whereas the grading system proposed by the WHO had no significant prognostic value.