Rare reports describe high molecular weight cytokeratin (clone 34betaE12) antibody cross-reactivity in scattered prostate carcinoma (PCa) cells, yet most often not in a true basal cell distribution. There are no data specifically describing 34betaE12 reactivity in basal cells in PCa. From August 10, 1995 to May 1, 2000, a total of 3198 consult prostate needle biopsies with PCa and a 34betaE12 immunoperoxidase stain were reviewed at our institution. Thirty-six cases (1.1%), which on hematoxylin and eosin stain were unequivocal cancer, had at least focal 34betaE12 positivity in a basal cell distribution. Twenty-five had original diagnostic slides for review. All cancers were Gleason score 6. The mean number of cancer glands per case was 36.9 (10-108) with an average of 39% of glands (1-100%) showing 34betaE12 reactivity. Twenty-one cases had patchy staining in a basal cell distribution with one other case showing continuous staining. An additional case showed mainly tumor cell reactivity with rare basal cell staining. The final two cases showed a zonal staining pattern with small glands toward one side of the lesion showing basal cells [one with high grade prostatic intraepithelial neoplasia (HGPIN); one without HGPIN]. HGPIN was present in 16 of 25 (64%) cases adjacent to PCa. The mean number of HGPIN glands was 1.36 (1-6). In cases with HGPIN the mean ratio of cancer to HGPIN glands was 6.8 (0.5-13.0). In 12 cases in which the lesion was still present on deeper sectioning, we were able to confirm in nine cases the presence of basal cells using antibodies to p63, another marker for prostatic basal cells. Four of the 25 men underwent radical prostatectomy; all showed Gleason score 6 PCa. Three radical prostatectomies demonstrated 34betaE12 reactivity: two with patchy staining in a basal cell distribution and one with mainly tumor cell staining. Adjacent HGPIN was present in all three radical prostatectomy specimens. Rare lesions with the appearance of PCa show 34betaE12 staining in a basal cell distribution either from retention of basal cells by early invasive cancer or from HGPIN outpouching. The lack of adjacent HGPIN in some cases and the large ratio of small atypical glands to HGPIN glands argue against HGPIN outpouching as the sole explanation. In cases with adjacent HGPIN a comparison of the proximity and number of the small, atypical, infiltrative appearing glands to HGPIN is helpful. The diagnosis of PCa in the face of positive 34betaE12 basal cell staining should be made with extreme caution, only in the face of unequivocal cancer on the hematoxylin and eosin stain.