PD40-06: Usefulness of Prostate Specific Antigen–Glycosylation Isomer for Significant Cancer Detection and Pathological Findings of Index Prostate Cancers on Magnetic Resonance Imaging–Transrectal Ultrasound Fusion Image-Guided Biopsy: A Prospective Multicenter Study

SUNAO SHOJI*, MAYURA NAKANO, Isehara, Japan, KOHEI UEMURA, Bunkyo-ku, Japan, KUMPEI TAKAHASHI, TATSUO KANO, SOICHIRO YUZURIHA, IZUMI HANADA, TAKAHIRO OGAWA, TATSUYA UMEMOTO, MASAYOSHI KAWAKAMI, MASAHIRO NITTA, MASANORI HASEGAWA, Isehara, Japan, TERUMITSU HASEBE, Hachioji, Japan, AKIRA MIYAJIMA, Isehara, Japan

Introduction: This study aimed to evaluate the ability of prostate specific antigen–glycosylation isomer (PSA-Gi) to predict the pathological findings of biopsy-proven index prostate cancer (PC).

Methods: Patients with serum PSA levels = 20 ng/mL who underwent magnetic resonance imaging (MRI)–transrectal ultrasound fusion image-guided target biopsy  were included. PSA-Gi was measured by an automated two-step Wisteria Floribunda Agglutinin lectin–anti-PSA antibody sandwich immunoassay using a highly sensitive surface plasmon field-enhanced fluorescence spectrometry system. Target biopsies were performed of cancer-suspicious lesions and analyzed by a Prostate Imaging-Reporting and Data System (PI-RADS) v.2.0.

Results: Two hundred four patients were included. Patients’ median values were age, 70 years; PSA, 6.71 ng/mL; PSA-Gi, 40.2 mU/mL; and prostate volume, 35 cc. Their categories were PI-RADS 3 (n=113), PI-RADS 4 (n=82), and PI-RADS 5 (n=9). Median PSA-Gi (65.4 mU/mL vs. 25.3 mU/mL, P<0.0001) and PSA-Gi density (PSA-GiD) (0.232 mU/mL/cc vs. 0.0578 mU/mL/cc, p<0.0001) levels significantly differed between patients with and those without detected biopsy-proven significant cancer (SC). Areas under the receiver operating characteristic (ROC) curves for PSA-Gi (AUC 0.851, 95%CI: 0.797-0.905, P<0.0001) and PSA-GiD (AUC 0.894, 95%CI: 0.846-0.942, P<0.0001) were significantly greater than non-discrimination for SC detection in all patients. Among patients with the highest PI-RADS category 3, ROC curves for PSA-Gi (AUC 0.831, P<0.0001) and PSA-GiD (AUC 0.863, P<0.0001) were significantly greater than non-discrimination for detecting clinically SC. Sensitivity, specificity, positive predictive, and negative predictive values of a PSA-GiD of 0.812 mU/mL/cc were 85%, 84%, 91.3%, and 72.7%, respectively. The highest Gleason score was positively correlated with PSA-Gi MRI-measured tumor volume density (PSA-GiTVD) (r=0.557, p<0.0001).

Conclusions: PSA-Gi, PSA-GiD, and PSA-GiTVD may predict pathological findings of biopsy-proven index PC. PSA-GiD may predict SC detection in patients with the highest PI-RADS category.

Source of Funding: None.