Predictors of prostate specific membrane antigen (PSMA/FOLH1) expression in a genomic database.

OBJECTIVE To assess predictors of prostate-specific membrane antigen (PSMA) expression in a genomic database; positron emission tomography with PSMA-targeted radiopharmaceuticals is increasingly being utilized. METHODS The de-identified Decipher Biosciences database, which includes expression for more than 46,000 coding and noncoding genes per patient, was queried for expression of FOLH1 (PSMA). Prostate cancer patients who underwent radical prostatectomy and received the Decipher Test were included in the analysis. PSMA expression was compared to the Gleason Grade Group, Decipher risk category (a validated 22 biomarker genomic score), basal versus luminal molecular subtype, and androgen receptor activity. Multivariable regression analyses were performed. RESULTS The Decipher de-identified Decipher Biosciences database contained 16,807 men who underwent prostatectomy with the average age being 65-year old and most being Gleason Grade Group 2 (35%) or 3 (27%). Higher Grade Group was associated with higher PSMA expression except in Grade Group 5 [Grade group: 1 (0.66), 2 (0.84), 3 (0.99), 4 (1.07), 5 (0.99), P < .001]. Luminal subtype was found to have much higher PSMA expression when compared to basal (1.01 vs 0.68, P < .001). The androgen receptor activity signature demonstrated a dramatic difference between basal (0.19) and luminal (0.62) subtypes (P < .001). In the multivariable model, luminal patients, high androgen receptor activity scores, and high Grade Groups were significantly associated with higher FOLH1 percentile rank (P < .001). CONCLUSION High PSMA expression (FOLH1) was associated with high androgen receptor activity and luminal subtype. Genomic tests could aid in predicting, interpreting, and/or directing PSMA theranostics. UROLOGY 144: 117-122, 2020. (c) 2020 Elsevier Inc.