A Comparison of Prostate Cancer Detection Between Visual-Estimation (Cognitive Registration) and Image-Fusion (Software Registration) Targeted Transperineal Prostate Biopsy

Abstract

PURPOSE: We compared clinically significant prostate cancer detection by visual estimation and image fusion targeted transperineal prostate biopsy. MATERIALS AND METHODS: This multicenter study included patients with multiparametric magnetic resonance imaging lesions undergoing visual estimation or image fusion targeted transperineal biopsy (April 2017-March 2020). Propensity score matching was performed using demographics (age and ethnicity), clinical features (prostate specific antigen, prostate volume, prostate specific antigen density and digital rectal examination), multiparametric magnetic resonance imaging variables (number of lesions, PI-RADS® score, index lesion diameter, whether the lesion was diffuse and radiological T stage) and biopsy factors (number of cores, operator experience and anesthetic type). Matched groups were compared overall and by operator grade, PI-RADS score, lesion multiplicity, prostate volume and anesthetic type using targeted-only and targeted plus systematic histology. Multiple clinically significant prostate cancer thresholds were evaluated (primary: Gleason ≥3+4). RESULTS: A total of 1,071 patients with a median age of 67.3 years (IQR 61.3-72.4), median prostate specific antigen of 7.5 ng/ml (IQR 5.3-11.2) and 1,430 total lesions underwent targeted-only biopsies (visual estimation: 372 patients, 494 lesions; image fusion: 699 patients, 936 lesions). A total of 770 patients with a median age of 67.4 years (IQR 61-72.1), median prostate specific antigen of 7.1 ng/ml (IQR 5.2-10.6) and 919 total lesions underwent targeted plus systematic biopsies (visual estimation: 271 patients, 322 lesions; image fusion: 499 patients, 597 lesions). Matched comparisons demonstrated no overall difference in clinically significant prostate cancer detection between visual estimation and image fusion (primary: targeted-only 54% vs 57.4%, p=0.302; targeted plus systematic 51.2% vs 58.2%, p=0.123). Senior urologists had significantly higher detection rates using image fusion (primary: targeted-only 45.4% vs 63.7%, p=0.001; targeted plus systematic 39.4% vs 64.5%, p <0.001). CONCLUSIONS: We found no overall difference in clinically significant prostate cancer detection, although image fusion may be superior in experienced hands.