OBJECTIVE: Staining of prostatic basal cells for the expression of high-molecular-weight cytokeratin has been suggested as a way of distinguishing benign from malignant prostate glands. We evaluated the utility of high-molecular-weight cytokeratin in the diagnosis of malignancy in prostate specimens obtained in various ways. DESIGN: Prostate tissues obtained from needle biopsies, transurethral resections, and total prostatectomies were immunostained with monoclonal antibody 34betaE12, an antibody directed against high-molecular-weight cytokeratins. RESULTS: Antiserum to high-molecular-weight cytokeratin only stained the basal cells in normal glands in 3 (12%) of 25 specimens obtained by transurethral resection. Other antigens, such as the alternate 10-nm filament protein vimentin, were unaffected and were detected in 100% of these specimens. However, keratin antigenicity in transurethral biopsies could be restored in these specimens by antigen retrieval in a low pH citrate buffer using a microwave heat technique. Keratin staining in needle biopsies and total prostatectomies was unaffected. CONCLUSION: In summary, our results indicate the technique of transurethral resection results in a specific loss of keratin antigenicity. This limits the utility of anticytokeratin 34betaE12 in interpreting transurethral resections without the application of antigen retrieval.