Basal cell-specific anti-cytokeratin antibody (34βE12) decorates the basal cells of benign prostatic epithelium by standard immunohistochemical techniques, whereas adenocarcinoma of the prostate lacks immunoreactivity with this antibody. We reviewed our experience with this antibody to determine its utility in the diagnosis of adenocarcinoma of the prostate as well as its pattern of usage at a tertiary medical center. In all, 7,242 prostate specimens from 5.262 men were seen at Johns Hopkins Hospital between 1/89 and 4/93. Immunostaining for basal cell-specific cytokeratin (34βE12) was originally used for diagnostic purposes in 289 questionable areas from 228 cases; 45% of these cases were seen in consultation. The distribution of cases using 34βE12 was 52% needle biopsies, 32% transurethral prostatic resections (TURPs), 13% radical prostatectomies, and 3% open enucleations. These procedures constituted 2.8% of all needle biopsies, 7.2% of all TURPs, 1.7% of all radical prostatectomies, and 3.5% of all enucleations seen during this time period. For this study the hematoxylin and eosin stain was reviewed without knowledge of the original diagnosis, a diagnosis was favored, the 34βE12 stain was examined, and a final diagnosis was determined. The 34βE12 stain established (14%), confirmed (58%), or changed (2%) our favored diagnoses, while 18% remained or became equivocal. The 34βE12 stain was of no use in 8% of the cases, yet we felt we were still able to render a final diagnosis even without the help of the stain. The differential diagnoses in the questionable foci using 34βE12 were cancer versus focus of atypical glands (44%), adenosis (39%), prostatic intraepithelial neoplasia (PIN) (8%), basal cell hyperplasia (5%), and atrophy (4%). However, 34βE12 was used in only 15–20% of all cases of adenosis and basal cell hyperplasia and in < 2% of PIN and atrophy cases seen during this time. Reasons for equivocal results were loss of suspicious glands on cut downs used for staining (49%), too few glands to rely on negative staining (23%), technical problems (15%), limited number of positive staining glands in a small focus (7%), and cautery artifact (6%). Although equivocal cases tended to have fewer negative stained glands than cases diagnosed with cancer, there was no minimum number of negative stained glands required to establish a diagnosis of cancer. From these data we conclude that 34βE12 staining is a useful tool in confirming, establishing, or changing the diagnosis in questionable foci seen in the everyday practice of surgical pathology. We feel that it is most useful in the workup of foci of atypical glands on needle biopsy and in differentiating low-grade adenocarcinoma from adenosis on TURP. Staining with 34βE12 is a cost-effective means to work up a needle biopsy as compared with ultrasound-guided re-biopsy.