Sixty-six patients with squamous cell carcinoma of the esophagus were treated with combined surgery and radiotherapy, mostly in a sandwich fashion. Fourteen patients received misonidazole during preoperative radiotherapy. Prospective cumulative survival at 1, 2, and 5 years was 60%, 35%, and 17%, respectively. Original biopsy specimens and operation specimens were evaluated according to a histopathologic grading system based on an evaluation of the tumor cell population in terms of cell differentiation, structure, nuclear polymorphism, and the frequency of mitotic figures. The tumor-host relationship was also estimated by the mode of invasion, stage of invasion, vascular invasion, and degree of lymphocytic infiltration. A multivariate regression analysis according to Cox and actuarial survival were used to determine the relative contributions of the clinical and histopathologic parameters to patient outcome. The major results were as follows: (1) patients who were old (70 to 80 years) fared as well as younger patients (P = 0.9198); (2) tumor site did not influence therapeutic outcome (P = 0.1100); (3) there was an insignificant difference in survival between patients with M0 and M1 disease (P = 0.7130); (4) radical surgery gave better survival; (5) misonidazole administered preoperatively was associated with worse survival (P = 0.0147); and (6) the histopathologic malignancy grading score system was very useful for prognostication--the tumor-host score in the operative specimen was the strongest of all analyzed predictive parameters. Since palliation was excellent in all patients treated in the combined fashion, selection criteria should be wide for such a program, not the least since long-term survival is possible also for patients with very advanced tumors.