P504S is a recently described, prostate cancer-specific gene that encodes a protein involved in the beta-oxidation of branched chain fatty acids. A recent study has shown that immunohistochemical detection of P504S gene product is a sensitive and specific marker of prostatic carcinoma in formalin-fixed, paraffin-embedded tissues. We performed a detailed analysis of P504S protein expression in a large series of prostate and bladder specimens with special emphasis on staining in specific morphologic patterns of prostatic adenocarcinoma, posthormonal and radiation therapy cases, and invasive urothelial carcinoma. A total of 366 prostate needle core biopsies from 124 patients with prostate cancer, 10 biopsies from 2 patients without prostate cancer, 28 prostatectomy specimens (16 with specific morphologic patterns, 7 posthormonal therapy and 5 postradiation therapy specimens), 5 bladder specimens with invasive urothelial carcinoma, and a single transurethral resection specimen from a patient with hormonally treated prostate cancer and invasive urothelial carcinoma were stained with P504S monoclonal antibody at a 1:250 dilution using standard heat-induced epitope retrieval and avidin-biotin technique. Extent (0, no staining; 1+, 1-10% staining; 2+, 11-50% staining; 3+, > or =51% staining) and location (luminal, subluminal, and diffuse cytoplasmic) of immunoreactivity in carcinoma and benign tissues were recorded. A total of 153 of 186 biopsies (82%) with prostatic adenocarcinoma stained for P504S. Pseudohyperplastic, atrophic, ductal, and mucinous prostatic carcinomas stained similarly, as did cases treated with hormone or radiotherapy. In 81 of 377 (21%) foci of benign prostatic tissue there was staining that was almost always focal, faint, and noncircumferential. Seminal vesicles did not stain for P504S. Five of six (83%) specimens with invasive urothelial carcinoma had 2+ staining and one case had focal staining. We conclude that immunohistochemistry for P504S has potential utility in the diagnosis of prostate cancer, including those treated by hormones and radiation. Circumferential luminal to subluminal and diffuse cytoplasmic staining is the most specific staining pattern for prostatic carcinoma and is almost never associated with benign prostatic tissue. However, a negative P504S immunostain does not automatically rule out prostate cancer, as 18% of cases were negative. Additionally, occasional benign glands, high-grade prostatic intraepithelial neoplasia, atypical adenomatous hyperplasia, and urothelial carcinoma may express P504S. Therefore, we think that P504S is best used only in conjunction with strict light microscopic correlation and preferably with high molecular weight cytokeratin immunostaining.

Artificial intelligence (AI) has shown promise for diagnosing prostate cancer in biopsies. However, results have been limited to individual studies, lacking validation in multinational settings. Competitions have been shown to be accelerators for medical imaging innovations, but their impact is hindered by lack of reproducibility and independent validation. With this in mind, we organized the PANDA challenge—the largest histopathology competition to date, joined by 1,290 developers—to catalyze development of reproducible AI algorithms for Gleason grading using 10,616 digitized prostate biopsies. We validated that a diverse set of submitted algorithms reached pathologist-level performance on independent cross-continental cohorts, fully blinded to the algorithm developers. On United States and European external validation sets, the algorithms achieved agreements of 0.862 (quadratically weighted κ, 95% confidence interval (CI), 0.840–0.884) and 0.868 (95% CI, 0.835–0.900) with expert uropathologists. Successful generalization across different patient populations, laboratories and reference standards, achieved by a variety of algorithmic approaches, warrants evaluating AI-based Gleason grading in prospective clinical trials.