Development and external validation of a novel nomogram to predict prostate cancer in biopsy-naïve patients with PSA <10 ng/ml and PI-RADS v2.1 = 3 lesions

Abstract

Objective: To develop and externally validate a novel nomogram in biopsy-naïve patients with prostate-specific antigen (PSA) <10 ng/ml and PI-RADS v2.1 = 3 lesions.

Methods: We retrospectively collected 307 men that underwent initial biopsy from October 2015 to January 2022 in Cohort 1 (The First Affiliated Hospital of Soochow University). External cohort (Cohort 2, Kunshan Hospital) included 109 men that met our criteria from July 2016 to June 2021. By Slicer-3D Software, the volume of all lesions was divided into two subgroups (PI-RADS v2.1 = 3a and 3b). Logistic regression analysis was performed to screen for variables and construct nomogram by analyzing clinical data from Cohort 1. Receiver operating characteristics curve analysis, calibration plot and decision curve analysis (DCA) were plotted to validate the nomogram in external cohort.

Results: A total of 70 (22.8%) patients was diagnosed with prostate cancer in Institution 1. Among them, 34 (11.1%) had clinically significant prostate cancer (csPCa). Age, prostate-specific antigen density, digital rectal examination, PI-RADS v2.1 = 3 subgroups (3a and 3b) and apparent diffusion coefficient (ADC, <750 mm² /s) were predictive factors for prostate cancer (PCa) and csPCa. High area under the curve of the nomogram was found in Cohort 1 and Cohort 2 for PCa (0.857 vs. 0.850) and for csPCa (0.896 vs. 0.893). Calibration curves showed excellent agreement between the predicted probability and actual risk for the models in internal and external validation. The DCA demonstrated net benefit of our nomogram.

Conclusion: Until now, this is the first nomogram that predicts PCa and csPCa in biopsy-naïve patients with PSA <10 ng/ml and PI-RADS v2.1 = 3 lesions. Furthermore, PI-RADS v2.1 = 3 subgroups were considered to be an independent risk factor in our model. Our nomogram may assist urologists in biopsy decision making for these so-called "double gray zone" patients.

Keywords: PI-RADS v2.1; csPCa; equivocal lesions; nomogram.

FIGURE 1

Nomogram predicting prostate cancer (A). Nomogram predicting clinically significant prostate cancer (B). Instructions: 1. Locate the five variables of patients on corresponding axis separately. 2. Then draw a vertical line of each variable upward to the “points” axis to determine how many points toward the probability of cancer. 3. Sum the points for each of the predictors and locate it on the “total points” axis. 4. Draw a line straight up to find the probability of PCa or csPCa. csPCa, clinically significant prostate cancer; PCa, prostate cancer.

FIGURE 2

Calibration curve of the nomogram. The institution 1 for PCa (A) and csPCa (B); The external validation group for PCa (C) and csPCa (D). csPCa, clinically significant prostate cancer; PCa, prostate cancer.

FIGURE 3

DCA of the nomogram. The institution 1 for PCa (A) and csPCa (B); The external validation group for PCa (C) and csPCa (D). csPCa, clinically significant prostate cancer; DCA, decision curve analyses; PCa, prostate cancer.

FIGURE 4

ROC curve comparing full model with model without PI‐RADS 3 subgroups for PCa (A) and for csPCa (B) in institution 1 and for PCa (C) and for csPCa (D) in the external cohort. csPCa, clinically significant prostate cancer; PCa, prostate cancer; ROC, receiver operating characteristic.