Performance of prostate health index and PSA density in a diverse biopsy‐naïve cohort with mpMRI for detecting significant prostate cancer

Abstract Objective To compare Prostate Health Index (PHI) and prostate‐specific antigen (PSA) density as secondary tests after multiparametric magnetic resonance imaging (mpMRI) in improving the detection accuracy of Gleason grade group (GG) 2‐5 prostate cancer (PCa) and in decreasing unnecessary biopsies in a multiethnic biopsy‐naïve population. Methods From February 2017 to February 2020, we recruited consecutive biopsy‐naïve men in participating urology clinics for elevated PSA levels. They all had a PHI score, mpMRI, and prostate biopsy. Experienced genitourinary radiologists read all mpMRI studies based on PIRADS version 2.0. Logistic regression models were used to generate receiver operating characteristic curves. Models were tested for effect modification between Race (Black vs White) and both PHI and PSA density, and Race and PIRADS to determine if race impacted their prediction accuracy. Sensitivity, specificity, and predictive values of PHI and PSA density thresholds were calculated by PIRADS scores. The primary outcome was GG2‐5 PCa, that is, Gleason score ≥3 + 4. Results The study included 143 men, of which 65 (45.5%) were self‐reported Black. Median age was 62.0 years and 55 men (38.4%) had GG2‐5 PCa. Overall, 18.1% had PIRADS 1‐2, 32.9% had PIRADS 3, and 49.0% had PIRADS 4‐5. For the binary logistic regressions, the interactions between PIRADS and Race (P = .08), Log (PHI) and Race (P = .17), and Log (PSA density) and Race (P = .42) were not statistically significant. Within PIRADS 3 lesions, a PHI ≥49 prevented unnecessary biopsies in 55% of men and missed no GG2‐5 PCa, yielding a negative predictive value of 100%. There was no reliable PHI or PSA density threshold to avoid PCa biopsies in PIRADS 1‐2 or 4‐5. Conclusions PHI and PSA density can be used after mpMRI to improve the detection of GG2‐5 PCa in a biopsy‐naïve cohort. PHI may be superior to PSA density in PIRADS 3 lesions by avoiding 55% of unnecessary biopsies. Using both PHI and PSA density in series may further increase specificity and lead to fewer unnecessary biopsies, but further larger studies are warranted to determine the optimal threshold of each biomarker.


| INTRODUC TI ON
While prostate-specific antigen (PSA) testing has led to improved detection of prostate cancer (PCa) and a reduction in PCa-specific mortality, 1 its poor specificity in detecting clinically significant Gleason grade group (GG) 2-5 PCa (ie, Gleason score ≥3 + 4) has resulted in over-detection and overtreatment of indolent PCa. 2,3 Multiparametric magnetic resonance imaging of the prostate (mpMRI) has emerged as an important tool to enhance the detection of GG2-5 PCa and guide targeted biopsies while reducing the detection of indolent PCa. 4,5 However, by itself, it is an imperfect test. The rate of detection of GG2-5 PCa varies widely, and it may miss up to 24% of clinically significant PCa. 6,7 Additionally, systematic reviews have shown that the negative predictive value (NPV) of mpMRI ranges from 63% to 98% for GG2-5 PCa, 8,9 which may lead to a high number of unnecessary prostate biopsies. 10  Given mpMRI's limitations, many have suggested using serumbased biomarkers, such as PSA density and the Prostate Health Index (PHI), to improve the detection of GG2-5 PCa and better identify men who can avoid a prostate biopsy. Both PHI and PSA density have been shown to have higher specificity for GG2-5 PCa than PSA, which could aid in preventing unnecessary biopsies. [13][14][15][16][17][18] Previous studies have demonstrated that PSA density 12,[19][20][21] and PHI 17,22 improve the PPV and NPV of mpMRI in detecting GG2-5 PCa in multivariable models and in series.
Few studies have investigated the effect that race has on these biomarkers. The cohorts previously studied have had minimal representation of Black men, a group with a higher than average incidence of PCa and greater risk of GG2-5 PCa. 23 No study to date has investigated whether PHI improves the accuracy of mpMRI in detecting GG2-5 PCa in biopsy-naïve men with adequate representation of Black men. The objective of this study was to determine the utility of PHI and PSA density in series after mpMRI in improving the accuracy for detecting GG2-5 PCa and improving the specificity in a multiethnic population. We additionally focus on the impacts of both markers in PIRADS 1-2, 3, and 4-5 lesions.

| Patient selection
From February 2017 to February 2020, 143 men who met the selection criteria were referred to participating urology clinics for elevated PSA levels and who underwent mpMRI were recruited into this prospective study. Selection criteria included men between the ages of 40 and 79 years who never had a previous prostate biopsy.
Men with a history of a previous prostate cancer were excluded from the study. Additionally, men with signs and symptoms of infection, prostatitis, or who were taking 5-alpha reductase inhibitors were excluded. Participating urologists performed a digital rectal examination (DRE) on all men at diagnostic biopsy. The DRE was classified as normal or suspicious.
These mpMRI studies were blinded and read by two highly experienced genitourinary radiologists (DC and WM) with significant experience in interpreting prostate MRI. Lesions were assigned suspicion scores from 1 to 5 based on the standardized PIRADS criteria version 2.0. 24 For our analysis, we considered PIRADS 1-2 lesions negative, PIRADS 3 lesions equivocal, and PIRADS 4-5 lesions highly suspicious.

| Prostate biopsy
Regardless of PIRADS score, all patients underwent MRI-informed prostate biopsy within 3 months of mpMRI. Patients who had PIRADS 3-5 lesions underwent MRI-transrectal ultrasound-guided PIRADS 3 lesions by avoiding 55% of unnecessary biopsies. Using both PHI and PSA density in series may further increase specificity and lead to fewer unnecessary biopsies, but further larger studies are warranted to determine the optimal threshold of each biomarker.

K E Y W O R D S
African American, biomarker, cancer detection, PIRADS 3, Prostate Health Index, prostate MRI, PSA density CARBUNARU et Al.
All PIRADS 1-2 men underwent systematic transrectal ultrasoundguided biopsy alone, with 42% of men also receiving a targeted MRI-guided biopsy. Early in the study from 2017-2018, PIRADS 1-2 lesions were also targeted for a prostate biopsy, but this practice was stopped in 2019 as the detection rates were determined to be low. All procedures were done by urologists with several years of experience with transrectal and fusion biopsy using the Invivo UroNav.

| Pathologic review
Pathological assessment of biopsy specimens was performed by and GG5: Gleason score = 9-10. 25 The primary outcome was GG2-5 PCa on biopsy.

| PHI and PSA assay
Serum PHI samples were drawn immediately before the biopsy and centrifuged to extract plasma on the day of biopsy in accordance with the manufacturer's recommendations (Beckman Coulter).
Blood was collected at least 4 days after any prostate manipulation (eg, DRE). PSA density was derived from the PSA in the PHI assay and the MRI-derived prostate volume which has been shown to be very similar to prostate volume derived from transrectal ultrasound and is available for biopsy decision making. 19

| Statistical analysis
PHI and PSA density were evaluated for their ability to increase the accuracy in predicting GG2-5 PCa beyond clinical data and PIRADS.
Logistic regression models were used to generate receiver operat- Thresholds of PHI and PSA density were generated, and the sensitivity, specificity, PPV, and NPV were compared for GG2-5 PCa detection at each threshold and stratified by PIRADS scores as PIRADS = 1-2, = 3, and = 4-5. The highest GG reported for both the MRItargeted and systematic transrectal ultrasound-guided biopsies was used as the outcome. Different PHI and PSA density thresholds were tested as a serial screening test after mpMRI to assess their impact on sensitivity, specificity, PPV, and NPV for GG2-5 PCa detection.
Assuming a baseline AUC of 0.66 for PHI alone, we had 80% power to detect a 0.10 difference in AUC between a logistic regression model with PHI alone compared with a model with PHI and PIRADS. Using an alpha = .05, there was 83% power to detect a difference in accuracy using both tools in series with n = 143 men with correlated samples. All comparisons were two-sided and P-

| RE SULTS
The study included 143 men, of which 78 identified as White and 65 as Black. Median age was 62.0 years, and 55 men (38.4%) had GG2-5 PCa. Compared with White men, Black men had statistically higher medians for BMI (28.1 vs 26.6 kg/m 2 ), PSA (7.2 vs 5.3 ng/mL), and PSA density (0.14 ng/mL/cm 3 vs 0.10 ng/mL/cm 3 ) and had greater frequencies of abnormal DRE (25.8% vs 9.0%) and GG2-5 PCa (58.5% vs 21.8%) (see Table 1). There were no statistical differences between racial groups in age, PHI scores, family history of PCa, history of BPH/LUTS, marriage rates, and smoking history.

|
The models discussed were also constructed using PSA density instead of PHI scores. The base model +Log(PSA density) yielded an AUC of 0.79, which was higher than the base model alone (P = .002, Figure 2).
The model with both PIRADS and Log(PSA density) increased the AUC to 0.85. The addition of race to the model resulted in an AUC of 0.86, which was statistically higher than the base model with Log(PSA density) (P = .04) or PIRADS alone (P = .005). The interaction between Log(PSA density) and Race (P = .42) was not statistically significant ( Figure 2).
The last models were also constructed using both PSA density and PHI scores. The base model +Log(PHI) + Log(PSA density) had an AUC of 0.83 for the detection of GG2-5 PCa, which was statistically higher than the base model +Log(PHI) alone (P = .002), but not for the Log(PSA density) model alone (P = .08). Overall, PSA density led to higher AUCs relative to PHI scores across all model iterations, yet these were not statistically significant.
Sensitivity and specificity analysis for PHI and PSA density were performed for PIRADS 1-2, 3, and 4-5 lesions. In men with a negative MRI (n = 26), that is, PIRADS 1-2 scores, a PSA density ≥0.11ng/mL/cm 3 detected 100% of the GG2-5 PCa (n = 6) but subjected 30% of men with a negative MRI to biopsy (false positives). Meanwhile, a PHI ≥45 also detected 100% of GG2-5 PCa in men with a PIRADS 1-2 lesion but would subject 60% of men with negative MRIs to an unnecessary prostate biopsy.
In PIRADS 3 men, a PSA density threshold of ≥0.07 ng/mL/cm 3 detected 88.9% of GG2-5 PCa, while avoiding 32% of unnecessary biopsies. While a PHI score of ≥49 prevented unnecessary biopsies in 55.3% of men and missed zero cases of GG2-5 PCa (see Table 3).
This cutoff had a PPV of 34.6% and NPV of 100%. A combination of both PHI and PSA density thresholds was also calculated. If only men with PHI scores ≥49 and PSA density ≥0.07 were biopsied, one GG2-5 PCa would be missed, making the sensitivity 88.9% but increasing the specificity to 71.1%. A sensitivity of 88.9% is achieved with a PHI threshold of ≥52, but the specificity (57.9%) is lower compared with the combination of both biomarkers.
In PIRADS 4-5 men, a sensitivity of 97.5% was achieved with a PHI score ≥27 and PSA density ≥.05 but they only avoided biopsies in 3.3% and 10% of men, respectively.

| D ISCUSS I ON
Our ROC analyses showed that PHI, PSA density, and PIRADS are independent predictors of GG2-5 PCa in biopsy-naïve men. Black race was an independent predictor of GG2-5 PCa but did not modify the effect of PHI, PSA density, or PIRADS scores. Our current sample Certain limitations of our study should be noted. Our sample size in this study was relatively small, recruited from tertiary and publicly funded medical centers, and included only Black and White men, which may limit the generalizability of our results to men of other races. We were also not powered for race-stratified analyses. Furthermore, we could have missed some men with GG2-5 PCa by not directly targeting all PIRADS 1-2 lesions at prostate biopsy.
However, this is consistent with the current prostate biopsy standard of care. 26 Our findings should be validated in a larger cohort with representation of other ethnic minorities.

| CON CLUS ION
PHI, PSA density, and mpMRI of the prostate are independent predictors that aid in the detection of GG2-5 PCa in a biopsy-naïve prostate biopsy cohort. PHI is particularly useful in men with equivocal PIRADS 3 lesions where it safely avoids over 50% of unnecessary biopsies and is superior to PSA density in this regard. Using both PHI and PSA density in series may further increase specificity and lead to fewer unnecessary biopsies, but further larger studies are warranted to determine the optimal threshold of each biomarker.