Background: Various biologic markers have been reported to be prognostic factors in human esophageal cancers. In the current study, we established a new tumor-grading system representing the malignancy potential of cancer cells and compared it with the clinical-stage system.
Methods: Tumor samples from 77 patients with squamous cell carcinoma of the esophagus were immunohistochemically evaluated for the expression of 10 molecules: the cell cycle-related molecules of cyclin D1, Rb, p16INK4, p27KIP1, and PCNA; the cell-cell adhesion molecules of E-cadherin, alpha-catenin, and beta-catenin; and the heat shock proteins of HSP27 and HSP70.
Results: P27KIP1, beta-catenin, and HSP70 were selected for their high hazard ratio in multivariate analysis, and the number of their disordered molecules was used to define the malignancy grade (MG). Five-year survival rates were 83%, 54%, 17%, and 0% for MG1, MG2, MG3, and MG4. The gradation of survival curves was better for MGs than for clinical stages. MGs and clinical stages showed significant correlation; however, 55% of those in higher clinical stages (stage 3 or 4) had lower MG (MG1 or 2) and showed better prognosis than others in their group (stage 3 or 4 and MG3 or 4). The proportions of shorter survival span to cancer death patients (less than 1 year) were 0%, 33%, 75%, and 100% in MG1, 2, 3, and 4, but the clinical stage was not associated with the survival span.
Conclusions: The grading of malignancy potential is clinically useful, especially for selecting patients who may show good prognosis in the advanced clinical stage and for predicting short survival span. These predictions are not possible with the clinical-stage system, which is based on the anatomic spread of cancer cells.