The American journal of surgical pathology

Immunohistochemical differentiation of high-grade prostate carcinoma from urothelial carcinoma.

PMID 17667550


The histologic distinction between high-grade prostate cancer and infiltrating high-grade urothelial cancer may be difficult, and has significant implications because each disease may be treated very differently (ie, hormone therapy for prostate cancer and chemotherapy for urothelial cancer). Immunohistochemistry of novel and established prostatic and urothelial markers using tissue microarrays (TMAs) were studied. Prostatic markers studied included: prostate-specific antigen (PSA), prostein (P501s), prostate-specific membrane antigen (PSMA), NKX3.1 (an androgen-related tumor suppressor gene), and proPSA (pPSA) (precursor form of PSA). "Urothelial markers" included high molecular weight cytokeratin (HMWCK), p63, thrombomodulin, and S100P (placental S100). TMAs contained 38 poorly differentiated prostate cancers [Gleason score 8 (n=2), Gleason score 9 (n=18), Gleason score 10 (n=18)] and 35 high-grade invasive urothelial carcinomas from radical prostatectomy and cystectomy specimens, respectively. Each case had 2 to 8 tissue spots (0.6-mm diameter). If all spots for a case showed negative staining, the case was called negative. The sensitivities for labeling prostate cancers were PSA (97.4%), P501S (100%), PSMA (92.1%), NKX3.1 (94.7%), and pPSA (94.7%). Because of PSA's high sensitivity on the TMA, we chose 41 additional poorly differentiated primary (N=36) and metastatic (N=5) prostate carcinomas which showed variable PSA staining at the time of diagnosis and performed immunohistochemistry on routine tissue sections. Compared to PSA, which on average showed 18.8% of cells with moderate to strong positivity, cases stained for P501S, PSMA, and NKX3.1 had on average 42.5%, 53.7%, 52.9% immunoreactivity, respectively. All prostatic markers showed excellent specificity. HMWCK, p63, thrombomodulin, and S100P showed lower sensitivities in labeling high-grade invasive urothelial cancer in the TMAs with 91.4%, 82.9%, 68.6%, and 71.4% staining, respectively. These urothelial markers were relatively specific with only a few prostate cancers showing scattered (