OBJECTIVE

To study the prognostic value of p53, bcl-2 and c-erbB-2 immunoexpression in predicting tumor relapses in low-grade papillary bladder neoplasms.

METHODS

In all patients a complete transurethral resection of the lesion was performed. All the material was formalin-fixed and paraffin-embedded. At the immunohistochemical level, the following monoclonal antibodies were utilized: p53 (clone DO7), bcl-2 (clone 124) and c-erbB-2 (clone CB11). In order to predict tumor relapses during follow-up, a study of specificity, sensitivity and predictive positive value (PVP) and negative was designed. In univariate statistical studies, the following tests were utilized: Chi-square, Kaplan-Meier estimates and Cox logistic regression.

RESULTS

Mean follow-up was 76.6 months (38 to 168). In recurrence prediction, p53 expression showed a high specificity (99%) as well as a high PPV (96%). Regarding bcl-2 and c-erbB-2 immunoexpression, both specificity (65% and 72%) and PPV (61% and 72%) were also high, although these percentages were lower than those obtained for p53 expression. The combined analysis of p53 and bcl-2 indicated that bcl-2 immunoexpression in non-basal cells of the urothelium could be independent of p53, although the number of cases showing this particular expression pattern is not high enough to perform an accurate statistical analysis. Otherwise, histologic grade demonstrated higher sensitivity (64%) and lower specificity (40%) than the immunohistochemical markers. In univariate studies, p53 showed an intense statistical correlation with relapse-free interval (RFI) and prediction of tumor recurrences during follow-up (p < 0.001), whereas the expression of bcl-2 (p = 0.065) was nearly correlated with RFI (p = 0.065). In contrast, expression of c-erbB-2 did not show statistical correlation (p = 0.112).

CONCLUSIONS

In our study, individual and combined analysis of p53 and bcl-2 immunoexpression have demonstrated to be useful in predicting tumor recurrences and RFI in low-grade bladder lesions.