Diagnosis of limited cancer can be challenging in prostate needle biopsies, and immunohistochemistry is commonly used in such settings. Recently, TMPRSS2:ERG gene rearrangement was found to be highly specific for and detected in approximately 50% of prostate cancer. Positive immunohistochemical staining with a novel anti-ERG antibody highly correlated with TMPRSS2:ERG gene rearrangement status. We developed a double immunohistochemical staining containing both erythroblastosis virus E26 oncogen (ERG) and basal cell marker P63 antibodies and evaluated its use in the diagnosis of limited cancer in prostate needle biopsies. A total of 77 prostate needle biopsies containing cancer occupying <1 mm of the length of only 1 core of the entire biopsy set were stained with the double stain containing ERG and P63 antibodies. ERG positivity and its staining intensity in cancerous and other noncancerous lesions were evaluated. ERG expression was detected in 42% (32 of 77) of cases, with strong, moderate, and weak staining intensity in 72%, 16%, and 12% of cases. The staining was uniform in 84% of cases and heterogeneous in 16% of cases with different staining intensities in >10% of cancerous cells. High-grade prostatic intraepithelial neoplasia was present in 17 cases, and in 5 (29%) cases ERG was positive in high-grade prostatic intraepithelial neoplasia glands, which were all immediately adjacent to or intermingled with ERG-positive cancerous glands. In 4 additional cases, positive ERG staining was found in morphologically benign glands, which were also immediately adjacent to or intermingled with ERG-positive cancerous glands. All other benign lesions distant from cancerous glands, including simple and partial atrophy, were negative for ERG. P63 was negative in all cancerous glands and positive in noncancerous lesions. The P63/ERG double immunostain combines the high sensitivity of P63 and the high specificity of ERG and may be potentially useful in the work-up of difficult prostate biopsies. The high specificity of ERG for the presence of cancer may have important implications for prostate biopsy interpretation and needs to be further validated in larger prospective studies. © 2011 Lippincott Williams & Wilkins.