Modified Prostate Health Index Density Significantly Improves Clinically Significant Prostate Cancer (csPCa) Detection

Abstract

Background: Early screening of clinically significant prostate cancer (csPCa) may offer opportunities in revolutionizing the survival benefits of this lethal disease. We sought to introduce a modified prostate health index density (mPHI) model using imaging indicators and to compare its diagnostic performance for early detection of occult onset csPCa within the prostate-specific antigen (PSA) gray zone with that of PHI and PHID.

Methods and participation: Between August 2020 and January 2022, a training cohort of 278 patients (total PSA 4.0-10.0 ng/ml) who were scheduled for a prostate biopsy were prospectively recruited. PHI and PHID were compared with mPHI $\left(\frac{\text{LD}}{\text{TRD}\times \text{APD}\times \text{TPV}}\times \text{PHI}\right)$ for the diagnosis performance in identifying csPCa. Pathology outcomes from systematic prostate biopsies were considered the gold standard.

Results: This model was tested in a training cohort consisting of 73 csPCa, 14 non-clinically significant prostate cancer(non-csPCa), and 191 benign prostatic hyperplasia (BPH) samples. In the univariate analysis for the PSA gray zone cohort, for overall PCa, the AUC of mPHI (0.856) was higher than PHI (0.774) and PHID (0.835). For csPCa, the AUC of mPHI (0.859) also surpassed PHI (0.787) and PHID (0.825). For detection of csPCa, compared with lower specificities from PHI and PHID, mPHI performed the highest specificity (76.5%), by sparing 60.0% of unnecessary biopsies at the cost of missing 11 cases of csPCa. The mPHI outperformed PHI and PHID for overall PCa detection. In terms of csPCa, mPHI exceeds diagnostic performance with a better net benefit in decision curve analysis (DCA) compared with PHI or PHID.

Conclusions: We have developed a modified PHI density (mPHI) model that can sensitively distinguish early-stage csPCa patients within the PSA gray zone.

Clinical trial registration: ClinicalTrials.gov, NCT04251546.

Keywords: PSA; clinically significant prostate cancer; gray zone; mPHI; prostate health index density.

Figure 1

(A) Flow diagram of enrolled participants of the training cohort. (B) Diagnostic scores of PHI, PHID, and mPHI in clinically significant prostate cancer and benign biopsies identification. *Overall vs csPCa, ^#Benign vs csPCa. Data are shown as Mean ± SEM. ***P <0.001, ^####P <0.0001. (C) Receiver operating analysis (ROC) curve of the training cohort show the area under the ROC curve for overall PCa (left) and clinically significant prostate cancer (right) for the prediction of clinically significant prostate cancer on prostate biopsy, namely, tPSA, total PSA, p2PSA, [−2] pro-PSA, PHI, prostate health index, PHID, PHI density, mPHI, defined as (LDTRD × APD ×TPV×PHI) . (D) Decision-curve analysis showing the net benefit and diagnostic performance of each prediction model in the training cohort for overall PCa (left) and clinically significant prostate cancer (right). The modified Prostate Health Index (mPHI) score (top line) demonstrated separation from the other biomarkers across a wide range of threshold values tPSA, total PSA; p2PSA, [−2] pro-PSA; PHI, prostate health index; PHID, PHI density; mPHI, modified PHI.