Reduction of false positives using zone-specific prostate-specific antigen density for prostate MRI-based biopsy decision strategies

Abstract

Objectives: To develop and test zone-specific prostate-specific antigen density (sPSAD) combined with PI-RADS to guide prostate biopsy decision strategies (BDS).

Methods: This retrospective study included consecutive patients, who underwent prostate MRI and biopsy (01/2012-10/2018). The whole gland and transition zone (TZ) were segmented at MRI using a retrained deep learning system (DLS; nnU-Net) to calculate PSAD and sPSAD, respectively. Additionally, sPSAD and PI-RADS were combined in a BDS, and diagnostic performances to detect Grade Group ≥ 2 (GG ≥ 2) prostate cancer were compared. Patient-based cancer detection using sPSAD was assessed by bootstrapping with 1000 repetitions and reported as area under the curve (AUC). Clinical utility of the BDS was tested in the hold-out test set using decision curve analysis. Statistics included nonparametric DeLong test for AUCs and Fisher-Yates test for remaining performance metrics.

Results: A total of 1604 patients aged 67 (interquartile range, 61-73) with 48% GG ≥ 2 prevalence (774/1604) were evaluated. By employing DLS-based prostate and TZ volumes (DICE coefficients of 0.89 (95% confidence interval, 0.80-0.97) and 0.84 (0.70-0.99)), GG ≥ 2 detection using PSAD was inferior to sPSAD (AUC, 0.71 (0.68-0.74)/0.73 (0.70-0.76); p < 0.001). Combining PI-RADS with sPSAD, GG ≥ 2 detection specificity doubled from 18% (10-20%) to 43% (30-44%; p < 0.001) with similar sensitivity (93% (89-96%)/97% (94-99%); p = 0.052), when biopsies were taken in PI-RADS 4-5 and 3 only if sPSAD was ≥ 0.42 ng/mL/cc as compared to all PI-RADS 3-5 cases. Additionally, using the sPSAD-based BDS, false positives were reduced by 25% (123 (104-142)/165 (146-185); p < 0.001).

Conclusion: Using sPSAD to guide biopsy decisions in PI-RADS 3 lesions can reduce false positives at MRI while maintaining high sensitivity for GG ≥ 2 cancers.

Clinical relevance statement: Transition zone-specific prostate-specific antigen density can improve the accuracy of prostate cancer detection compared to MRI assessments alone, by lowering false-positive cases without significantly missing men with ISUP GG ≥ 2 cancers.

Key points: • Prostate biopsy decision strategies using PI-RADS at MRI are limited by a substantial proportion of false positives, not yielding grade group ≥ 2 prostate cancer. • PI-RADS combined with transition zone (TZ)-specific prostate-specific antigen density (PSAD) decreased the number of unproductive biopsies by 25% compared to PI-RADS only. • TZ-specific PSAD also improved the specificity of MRI-directed biopsies by 9% compared to the whole gland PSAD, while showing identical sensitivity.

Keywords: Clinical decision-making; Image-guided biopsy; Magnetic resonance imaging; Prostate-specific antigen density; Prostatic neoplasms.

Fig. 1

Patient flow diagram of the study sample. Fifteen men were missing a Prostate Imaging-Reporting and Data System (PI-RADS) score in the validation set (*). Thus, biopsy decision strategies (BDS) using prostate-specific antigen density (PSAD) or transition zone-specific (s)PSAD and PI-RADS were assessed in a subset of 228 men, while diagnostic performance of PSAD and sPSAD alone was assessed using the entire validation set (n = 243). A hold-out test set (n = 389) was used for testing the different BDS combining transition zone- or whole gland-based PSADs and PI-RADS scoring (TRUS transrectal ultrasound; TURP transurethral resection; PCa prostate cancer)

Fig. 2

Whole gland and transition zone segmentation at MRI using a nnU-Net. Axial T2-weighted images at three different levels of the prostate (left) and the correlating segmentation masks (right). The yellow and green masks represent the whole gland and transition zone volume, respectively. The white arrow indicates a target biopsy-proven Gleason Score 3+3 prostate cancer in the ventral transition zone

Fig. 3

Patient-based diagnostic performance in detecting clinically significant prostate cancer using PSA density in the validation and external test set. a Graph shows the receiver operating characteristic (ROC) curves for the detection of ISUP Grade Group (GG) ≥ 2 cancers using transrectal ultrasound (TRUS)- and MRI-based prostate-specific antigen density (PSAD) (green and orange, respectively) and MRI-based transition zone-specific (s)PSAD (blue). Comparator studies with plotted detection accuracy of any-grade PCa and GG ≥ 2 cancers included Boesen et al [11], Knaaplia et al [13], Falagario et al [9], and Hansen et al [8]. Data on cancer detection using a PSAD cutoff of > 0.15 was not available in Hansen et al [8]. b Graph shows the ROC curves for the detection of GG ≥ 2 cancers using TRUS- and MRI-based PSAD (green and orange, respectively) and MRI-based sPSAD (blue) in patients with an MRI-based PSAD of 0.1–0.2. c, d Graph shows the ROC curves for the detection of GG ≥ 2 cancers using MRI-based PSAD and sPSAD in the external test set (n = 692) [31]. 95%CIs are shown as transparent areas around the mean curves. 95%CI were estimated through bootstrapping. *Significantly superior performance in comparison to TRUS-based PSAD. °Significantly superior performance in comparison to MRI-based PSAD. (AUC area under the ROC curve)

Fig. 4

Proportions of clinically significant prostate cancer related to PI-RADS and PSA density risk category in men planned for prostate biopsy. Proportions of clinically significant prostate cancers (ISUP GG ≥ 2; red bars) and GG = 1 or no cancers (green bars), related to PI-RADS score categories (1–2, 3, 4–5) at whole gland PSA densities (PSAD; < 0.15 and ≥ 0.15 ng/mL/cc) as well as transition zone-specific (s)PSAD (< 0.42 and ≥ 0.42 ng/mL/cc) risk categories, respectively, in the validation and hold-out test sets. The overall accepted risk threshold of 9% for GG ≥ 2 by 2019 EAU prostate cancer guidelines, when using prostate MRI for biopsy decisions, is plotted as dashed horizontal line. The diagram illustrates that men without GG ≥ 2 are more likely to have a sPSAD below 0.42 than a PSAD below 0.15. *Fifteen of the 243 men in the validation set had to be excluded from the analysis as no PI-RADS score could be determined

Fig. 5

Decision curve analysis comparing clinical utility of different biopsy strategies for detecting clinically significant prostate cancer in men planned for prostate biopsy. Decision curve analyses simulate two scenarios: in one all the men with PI-RADS 3-5 would receive biopsy (PI-RADS 3-5, blue), and in the other none would undergo biopsy (zero on x-axis). Clinically useful biopsy decision strategies lie above these scenarios. The graph gives the expected net benefit per patient relative to biopsy none. The unit is the benefit associated with one patient having GG ≥ 2 duly undergoing biopsy. In internal and external datasets, at a 40% biopsy threshold (=2 out of 5 biopsies yield GG ≥ 2 cancer), the sPSAD-based biopsy decision strategy (BDS) had a net benefit compared to PI-RADS/UCLA 3-5 and PSAD-based BDS. Note how incorporating PSAD improved the net benefit of the MRI strategies in all datasets. *Fifteen of the 243 men in the validation set had to be excluded from the analysis as no PI-RADS score could be determined. (PI-RADS Prostate Imaging-Reporting and Data System; (s)PSAD (transition zone-specific) prostate-specific antigen density; UCLA Likert-like scoring system, similar to PI-RADS v2)