Novel reference gene microseminoprotein‐beta based prostate cancer antigen 3 urine test: Enhanced diagnostic performance for prostate cancer and feasibility of non‐digital rectal examination urine sampling

Abstract Background The diagnosis of prostate cancer (PCa) is limited by low specificity and invasiveness. This study aims to establish a highly accurate, noninvasive reverse transcription‐quantitative polymerase chain reaction (RT‐qPCR)‐based urinary prostate cancer antigen 3 ( PCA3 ) assay system and explore the feasibility of using first‐morning urine instead of post‐digital rectal examination (DRE) samples. Methods Three prospective multicenter cohort studies were conducted on 643 Chinese men. Urine PCA3 expression was quantified by RT‐qPCR, using either the novel reference gene microseminoprotein‐beta ( MSMB ) or conventional kallikrein‐related peptidase 3 ( KLK3 , encoding prostate‐specific antigen, PSA). Diagnostic performance was evaluated through receiver operating characteristic (ROC) analysis in Cohort I ( n = 176), with validation in Cohort II ( n = 350). In Cohort III ( n = 117), the diagnostic accuracy of PCA3 / MSMB test was compared between DRE and non‐DRE urine samples. Results The PCA3 / MSMB RT‐qPCR system showed high precision and stability. In Cohort I, the PCA3 / MSMB test demonstrated superior diagnostic performance compared with the PCA3 / KLK3 test, with a significantly higher area under the receiver operating characteristic curve (AUC) (0.869 vs. 0.830, p = 0.0468), higher sensitivity (82.1% vs. 79.0%), and higher specificity (77.8% vs. 74.1%). Cohort II confirmed these results, yielding a sensitivity of 82.63% and a specificity of 77.50%. In Cohort III, no significant difference in diagnostic performance was observed between non‐DRE and DRE urine samples (AUC 0.859 vs. 0.850, p = 0.6262), suggesting the potential for simplified sample collection. Conclusion The PCA3 / MSMB urine test is a promising noninvasive diagnostic tool for PCa; it can reduce unnecessary biopsies and improve detection accuracy using non‐DRE urine samples.