Key Words: prostate cancer, biomarkers, biopsy, magnetic resonance imaging, liquid biopsy
Abstract
Purpose:
We developed and externally validated a nondigital rectal examination urine test to inform whether biopsy is necessary in patients undergoing active surveillance (AS). Performance and clinical consequences of testing were directly compared with multiparametric MRI (mpMRI).
Materials and Methods:
Biomarker models (MyProstateScore 2.0—Active Surveillance [MPS2-AS]) were derived to predict upgrading to Grade Group (GG) ≥ 3 and GG ≥ 2 and externally validated across 11 practices. Urine was prospectively collected, and patients underwent ≥ 12-core systematic biopsy plus targeted biopsy of Prostate Imaging Reporting and Data System (PI-RADS) ≥ 3 lesions. Diagnostic performance and clinical consequences of urinary testing to determine the need for biopsy were compared with mpMRI.
Results:
The validation cohort included 330 patients with GG1 cancer scheduled for AS biopsy. Overall, 280 (85%) patients had prebiopsy mpMRI, of which 130 (46%) were PI-RADS 1 to 2, 42 (15%) were PI-RADS 3, and 108 (39%) were PI-RADS 4 to 5. On biopsy, 31 (9.4%) patients upgraded to GG ≥ 3 and 123 (37%) to GG ≥ 2. MPS2-AS provided higher area under the receiver operating characteristic curve than mpMRI for upgrading to both GG ≥ 3 (0.82 vs 0.73) and GG ≥ 2 (0.74 vs 0.64). Clinically, prebiopsy MPS2-AS would have avoided 64% of unnecessary biopsies while failing to detect only 3.2% of GG ≥ 3 upgrades and 4.9% of GG ≥ 2 upgrades. By contrast, the use of PI-RADS ≥ 3 would have failed to detect 18% of GG ≥ 3 upgrades, 35% of GG ≥ 2 upgrades, and avoided fewer unnecessary biopsies (50%). Performance of MPS2-AS was consistent across clinically pertinent subgroups (confirmatory and surveillance biopsy, Black and non-Black patients).
Conclusions:
In a multisite AS population, urinary MPS2-AS provided highly accurate and actionable testing for GG ≥ 3 and GG ≥ 2 upgrading, meaningfully outperforming mpMRI on direct comparison. These findings suggest that noninvasive monitoring with MPS2-AS could reduce the need for scheduled biopsies and serial mpMRI.
Active surveillance (AS) is the preferred management option for low-risk prostate cancer (PCa).1-3 AS aims to reduce overtreatment of cancers that pose minimal risk of harm while monitoring for higher-grade tumors that necessitate treatment. Supported by data reflecting excellent long-term oncologic outcomes,4 acceptance of AS has increased substantially, with nearly 60% of patients with Grade Group (GG) 1 cancers now managed on AS.5,6
JU Insight
Study Need and Importance
Active surveillance (AS) is the preferred management option for patients with low-risk prostate cancer. Current monitoring tools, including PSA testing and MRI, are not reliable indicators of upgrading, resulting in the need for scheduled prostate biopsies during the course of AS. This study externally validated the association of a first-catch, nondigital rectal examination urine test (MyProstateScore 2.0—Active Surveillance [MPS2-AS]) with AS biopsy outcomes and directly compared its accuracy to MRI.
What We Found
In patients with Grade Group (GG) 1 prostate cancer pursuing AS, the use of MPS2-AS to direct the decision for biopsy would have detected 97% of GG ≥ 3 upgrades while avoiding 64% of unnecessary biopsies. By comparison, the use of Prostate Imaging Reporting and Data System ≥ 3 MRI would have detected 82% of GG ≥ 3 upgrades while avoiding fewer unnecessary biopsies (50%). For GG ≥ 3 and GG ≥ 2 upgrading, MPS2-AS provided sufficiently high sensitivity (95%-97%) and negative predictive value (92%-99%) to ensure detection of higher-grade cancers while avoiding unnecessary invasive biopsies. On direct comparison, urine-based testing would have provided favorable clinical outcomes (ie, upgrades detected, unnecessary biopsies avoided) relative to PSA-based testing and MRI (Table).
Limitations
Urinary MPS2-AS values were calculated retrospectively and were not considered in biopsy decision-making. As such, the current data provide clinical outcomes projected based on urinary testing rather than those prospectively observed. While performance was consistent across confirmatory and surveillance biopsies, the number of previous biopsies performed in the surveillance biopsy subgroup was not available. Finally, there were limited data from underrepresented populations, including Black men, meriting further study.
Interpretation for Patient Care
- In patients on active surveillance for GG1 prostate cancer, noninvasive testing with urinary MPS2-AS was more accurate than MRI and PSA-based testing for predicting higher-grade cancer on biopsy. These findings suggest that urine-based testing can reduce the need for invasive biopsies without compromising prompt detection of higher-grade cancers.
Table.
Diagnostic Accuracy and Clinical Consequences of MyProstateScore 2.0—Active Surveillance, PSA, Prostate Cancer Prevention Trial Risk Calculator, PSA Density, and Multiparametric MRI for Prediction of Upgrading in the Validation CohortThreshold Sens (%) Spec (%) NPV (%) PPV (%) Potentially missed/delayed upgrading (%) Unnecessary Bx potentially avoided per 1000 patients A. GG1 to GG ≥3 MPS2-AS 11.5% 97 64 99 22 3.2 635 PSA 3.7 97 16 98 11 3.2 164 PCPTrc 6.0% 97 17 98 11 3.2 171 PSAD 0.08 97 23 99 12 3.2 234 PSAD 0.15 65 65 95 16 35 652 mpMRI PI-RADS ≥3 82 50 96 15 18 496 PI-RADS ≥4 79 66 97 20 21 659 B. GG1 to GG ≥2 MPS2-AS 11.5% 95 34 92 46 4.9 338 PSA 2.9 95 12 80 39 4.9 116 PCPTrc 5.1% 95 13 82 39 4.9 130 PSAD 0.05 95 12 81 39 4.9 121 PSAD 0.15 54 72 72 53 46 720 mpMRI PI-RADS ≥3 65 54 71 47 35 538 PI-RADS ≥4 55 72 72 56 45 719 Abbreviations: Bx, biopsy; GG, Grade Group; mpMRI, multiparametric MRI; MPS2-AS, MyProstateScore 2.0—Active Surveillance; NPV, negative predictive value; PCPTrc, Prostate Cancer Prevention Trial Risk Calculator; PI-RADS, Prostate Imaging Reporting and Data System; PPV, positive predictive value; PSAD, PSA density; Sens, sensitivity; Spec, specificity.
AS monitoring protocols have traditionally included serial PSA testing, digital rectal examination (DRE), and, more recently, multiparametric MRI (mpMRI) to monitor for evidence of cancer upgrading.1-3 Yet, each of these modalities comes with notable limitations. PSA kinetics do not reliably predict grade reclassification on AS,7 and the limited accuracy of DRE is well described.8 While informative in the diagnostic setting, several expert centers have shown that mpMRI lacks accuracy to reliably rule out or rule in upgrading in AS patients.9-11 A recent meta-analysis revealed that mpMRI missed 41% of higher-grade cancers during surveillance.12 Based on the limitations of these tools, repeat biopsies remain a required component of AS, despite associated discomfort and increasing risk of complications over time.13
A major priority in the field is to reduce the need for invasive testing during AS while maintaining excellent oncologic outcomes.14 While existing biomarker tests have shown an association with cancer upgrading,15,16 none have provided a clear clinical approach to reliably inform the need for forthcoming surveillance biopsies. MyProstateScore 2.0 (MPS2) is an 18-gene urinary test validated to inform biopsy decision-making in patients with elevated PSA.17 Because MPS2 captures genes uniquely overexpressed in high-grade vs low-grade cancers, there is a strong biologic foundation supporting its ability to differentiate between AS patients truly harboring only low-grade disease and those most likely to harbor occult high-grade cancers.
In this study, we developed an MPS2-based non-DRE testing approach (MyProstateScore 2.0—Active Surveillance [MPS2-AS]) to inform the need for repeat biopsy during AS, externally validated this approach in a prospectively enrolled multisite cohort, and directly compared the accuracy of MPS2-AS testing with mpMRI in the AS population.
MATERIALS AND METHODS
Institutional Review Board approval was obtained from the Western Institutional Review Board-Copernicus Group (20241020) and site IRBs (UM-HUM00037879, VUMC220357), and all participants provided written informed consent. This study followed the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) Statement and Standards for Reporting of Diagnostic Accuracy (STARD) guideline.18,19
Model Development
The development cohort included 761 patients presenting for diagnostic biopsy (Supplementary Table 1, https://www.jurology.com) at the University of Michigan as previously described17 and detailed in the Supplement (https://www.jurology.com). Prebiopsy urine was prospectively collected under an IRB-approved National Cancer Institute protocol. RNA isolation, extraction, and complementary DNA synthesis were performed, as previously described.17 Biomarker amplification and measurement were performed using the QuantStudio 12K Flex Real-Time PCR System (Thermo Fisher Scientific).17 Biomarker expression was quantified by the relative cycle threshold (Ct), defined as the number of amplification cycles required for fluorescence to exceed background. All samples were run in triplicate, and mean Ct values were normalized to the prostate-specific reference gene KLK3. Normalized mean Ct values were used for analyses.
We optimized the MPS2 panel to predict upgrading from GG1 to GG ≥ 3 (MPS2-AS GG1-3 model) and GG1 to GG ≥ 2 (MPS2-AS GG1-2 model) as detailed in the Supplement (https://www.jurology.com). In brief, optimal combinations of predictive genes were identified from the MPS2 18-gene panel using recursive feature elimination and coefficient-based filtering. PSA density (PSAD) was locked into models a priori based on predictive value for high-grade cancer.20 Model outputs were calibrated to provide the percentage risk (0%-100%) of upgrading to GG ≥ 3 and GG ≥ 2. In accordance with principles of diagnostic biomarker testing from the National Comprehensive Cancer Network,21 MPS2-AS thresholds were set to achieve a sensitivity ≥ 90% and specificity ≥ 30% for higher-grade cancer. Thresholds were calibrated to equal 11.5%, consistent with MPS2 non-DRE diagnostic testing.22
The locked MPS2-AS models for upgrading to GG ≥ 3 and GG ≥ 2 incorporated 5 genes and 9 genes from the MPS2 testing panel, respectively (Supplementary Table 2, https://www.jurology.com). The MPS2-AS–based predicted probabilities of upgrading aligned closely with the observed prevalence of upgrading, indicating good calibration (Supplementary Figure 1, https://www.jurology.com).
Power Calculation
We performed a formal power analysis using prospective-specimen-collection, retrospective-blinded-evaluation (PRoBE) principles for a rule-out test using a continuous biomarker.23 Validation cohort sizes exceeded the calculated sample sizes required (Supplementary Table 3, https://www.jurology.com), confirming that the current analysis was sufficiently powered.
Prospective Multisite Validation
A total of 1381 patients enrolled in the prospective biospecimen protocol at 11 urology practices (Supplementary Figure 2, https://www.jurology.com). The current analysis included 330 patients with GG1 PCa that were undergoing confirmatory or surveillance biopsy as part of their planned clinical management (ie, the decision to pursue biopsy was made prior to and independent of enrollment). Patients taking 5-alpha reductase inhibitors were excluded. Most patients underwent prebiopsy mpMRI according to local standard, and images were interpreted at each site using Prostate Imaging Reporting and Data System (PI-RADS) v2.24 Participants provided ≤ 40 mL first-catch urine without preceding DRE using the FDA-approved Colli-Pee device, per existing MPS2 non-DRE protocol. Analyte signals were measured per MPS2 non-DRE workflow,22 and normalized mean Ct and MPS2-AS values were calculated retrospectively as in the development cohort. All patients underwent ≥ 12-core transrectal or transperineal systematic biopsy, and patients with PI-RADS ≥ 3 lesions underwent targeted sampling.
Statistical Analysis
The primary outcome was upgrading to GG ≥ 3 cancer; upgrading to GG ≥ 2 was assessed secondarily. The area under the receiver operating characteristic curve (AUROC) and 95% CIs were calculated for discriminative accuracy of MPS2-AS models, PSA, the Prostate Cancer Prevention Trial Risk Calculator (PCPTrc),25 PSA density (PSAD), and mpMRI. Statistical significance for AUROC differences between these testing approaches was determined using the DeLong test. Measures of diagnostic performance (ie, sensitivity, specificity, negative predictive value [NPV], positive predictive value) were calculated retrospectively for each modality. Continuous tests (PSA, PCPTrc, and PSAD) were assessed at the threshold values providing equivalent sensitivity to MPS2-AS, and PSAD was additionally assessed at the conventional threshold of 0.15.21 Clinicodemographic variables were measured at the time of enrollment. Prostate volume was obtained from mpMRI or most recent transrectal ultrasound in patients without mpMRI. Analyses were repeated in clinically pertinent subgroups: confirmatory and surveillance biopsy, age < 65 and ≥ 65 years, and Black and non-Black race.
Decision curve analysis (DCA) was performed to quantify the net benefit of each testing modality on the decision to perform biopsy compared with (1) biopsying all patients and (2) biopsying no patients.26 Because risk of upgrading < 5% justifies forgoing biopsy and risk > 20% justifies performing biopsy in most patients,2,27 these values defined the threshold probability range for GG ≥ 3. Because some patients with GG2 remain AS candidates,1-3 a more permissive upper threshold of 30% was used for upgrading to GG ≥ 2. Statistical analyses were performed using R version 4.3.0 and Python version 3.9.20.
We assessed independent and combined diagnostic accuracy of MPS2-AS and mpMRI in post hoc analysis of patients with both results (N = 280). Rates of upgrading associated with negative and positive results were calculated for each test and for concordant and discordant (eg, MPS2-AS negative/mpMRI positive) results. Finally, we calculated the additive diagnostic value that would have been provided by performing MPS2-AS testing after mpMRI.
RESULTS
Validation Cohort
MPS2-AS testing was performed on 338 urine samples, of which 330 yielded valid results (98% valid rate). The external validation cohort thus included 330 men with GG1 PCa presenting for confirmatory or surveillance biopsy (Table 1). The median time from diagnosis to confirmatory biopsy (n = 99) was 21 months (IQR 14-40), and the median time from prior biopsy was 25 months (IQR 16-44) for patients undergoing surveillance biopsy (n = 221). Overall, 280 (85%) patients underwent mpMRI at a median of 2.1 months (IQR 0.9-8.7) before index biopsy. Of these, 130 (46%) were classified as PI-RADS 1 to 2, 42 (15%) as PI-RADS 3, and 108 (39%) as PI-RADS 4 to 5.
Table 1.
Validation Cohort Characteristics
| Characteristic | No. (%) or median (IQR) |
| Total No. | 330 |
| Median age (IQR), y | 69 (63-73) |
| No. Black race (%) | 36 (11) |
| No. positive family history (%) | 56 (17) |
| No. suspicious DRE (n = 323) (%) | 17 (5) |
| Median PSA (IQR), ng/mL | 6 (4.4-9.2) |
| Median prostate volume (IQR) | 47 (35-66) |
| Median PSA density (IQR) | 0.12 (0.08-0.2) |
| No. underwent MRI (%) | 280 (85) |
| PI-RADS 1-2 (% of mpMRI) | 130 (46) |
| PI-RADS 3 | 42 (15) |
| PI-RADS 4 | 73 (26) |
| PI-RADS 5 | 35 (13) |
| Biopsy indication | |
| Confirmatory biopsy | 99 (30) |
| Surveillance biopsy | 221 (67) |
| Not reported | 10 (3) |
| Biopsy results | |
| Benign/Grade Group 1 | 207 (63) |
| Grade Group 2 | 92 (28) |
| Grade Group 3 | 21 (6) |
| Grade Group 4 | 8 (2) |
| Grade Group 5 | 2 (1) |
| Grade Group ≥2 | 123 (37) |
| Grade Group ≥3 | 31 (9.4) |
Abbreviations: DRE, digital rectal exam; mpMRI, multiparametric MRI; PI-RADS, Prostate Imaging Reporting and Data System.
GG ≥ 3 Upgrading: Comparative Accuracy and Clinical Consequences of Testing
On study biopsy, 31 (9.4%) patients upgraded to GG ≥ 3, of which 17 (55%) were detected on systematic biopsy only, 6 (19%) on targeted biopsy only, and 8 (26%) on both systematic and targeted biopsy (Table 2A). AUROC values (95% CI) for GG ≥ 3 were 0.68 (0.59-0.78) for PSA, 0.70 (0.60-0.79) for PCPTrc, 0.74 (0.65-0.84) for PSAD, and 0.73 (0.63-0.82) for mpMRI, as compared with 0.82 (0.76-0.88) for MPS2-AS (Supplementary Figure 3A, https://www.jurology.com). MPS2-AS provided significantly better discrimination for GG ≥ 3 than PSA (P < .001), PCPTrc (P = .012), and PSAD (P = .005), and provided clinically meaningful improvement relative to mpMRI (0.82 vs 0.73) that did not meet conventional levels of significance (P = .10; Supplementary Table 4A, https://www.jurology.com).
Table 2.
Detection of Upgrading by Systematic and Targeted Biopsy
| Upgraded, No. (% of cohort) | Detected on SBx | Detected on TBx | Detected on SBx only | Detected on TBx only | Detected on SBx and TBx |
| A. GG1 to GG ≥3 | |||||
| 31 (9.4) | 25 (81) | 14 (45) | 17 (55) | 6 (19) | 8 (26) |
| B. GG1 to GG ≥2 | |||||
| 123 (37) | 108 (88) | 49 (40) | 74 (60) | 15 (12) | 34 (28) |
Abbreviations: GG, Grade Group; SBx, systematic biopsy; TBx, targeted biopsy.
Values represent No. (%) of upgrades unless otherwise indicated.
Consistent with guidelines,21 we used a testing threshold to ensure < 10% missed/delayed diagnoses of GG ≥ 3 (ie, sensitivity ≥ 90%) and > 30% unnecessary biopsies avoided (ie, specificity > 30%). At the validated diagnostic threshold of 11.5%, MPS2-AS testing would have detected 97% of GG ≥ 3 upgrades and avoided 64% of unnecessary biopsies, as compared with avoiding 16% of unnecessary biopsies using PSA, 17% using PCPTrc, and 23% using PSAD at equal sensitivity. MPS2-AS testing was more sensitive than mpMRI (97% vs 79%-82%) and generally more specific (64% vs 50%-66%) for GG ≥ 3 cancer using either PIRADS ≥ 3 or PI-RADS ≥ 4 (Table 3A). Applied clinically, MPS2-AS would have missed/delayed detection of 3.2% of GG ≥ 3 cancers, as compared with 18% to 21% using mpMRI, while MPS2-AS also avoided more unnecessary biopsies.
Table 3.
Diagnostic Accuracy and Clinical Consequences of MyProstateScore 2.0—Active Surveillance, PSA, Prostate Cancer Prevention Trial Risk Calculator, PSA Density, and Multiparametric MRI for Prediction of Upgrading in the Validation Cohort
| Threshold | Sens (%) | Spec (%) | NPV (%) | PPV (%) | Potentially missed/delayed upgrading (%) | Unnecessary Bx potentially avoided per 1000 patients | |
| A. GG1 to GG ≥3 | |||||||
| MPS2-AS | 11.5% | 97 | 64 | 99 | 22 | 3.2 | 635 |
| PSA | 3.7 | 97 | 16 | 98 | 11 | 3.2 | 164 |
| PCPTrc | 6.0% | 97 | 17 | 98 | 11 | 3.2 | 171 |
| PSAD | 0.08 | 97 | 23 | 99 | 12 | 3.2 | 234 |
| PSAD | 0.15 | 65 | 65 | 95 | 16 | 35 | 652 |
| mpMRI | PI-RADS ≥3 | 82 | 50 | 96 | 15 | 18 | 496 |
| PI-RADS ≥4 | 79 | 66 | 97 | 20 | 21 | 659 | |
| B. GG1 to GG ≥2 | |||||||
| MPS2-AS | 11.5% | 95 | 34 | 92 | 46 | 4.9 | 338 |
| PSA | 2.9 | 95 | 12 | 80 | 39 | 4.9 | 116 |
| PCPTrc | 5.1% | 95 | 13 | 82 | 39 | 4.9 | 130 |
| PSAD | 0.05 | 95 | 12 | 81 | 39 | 4.9 | 121 |
| PSAD | 0.15 | 54 | 72 | 72 | 53 | 46 | 720 |
| mpMRI | PI-RADS ≥3 | 65 | 54 | 71 | 47 | 35 | 538 |
| PI-RADS ≥4 | 55 | 72 | 72 | 56 | 45 | 719 |
Abbreviations: Bx, biopsy; GG, Grade Group; mpMRI, multiparametric MRI; MPS2-AS, MyProstateScore 2.0—Active Surveillance; NPV, negative predictive value; PCPTrc, Prostate Cancer Prevention Trial Risk Calculator; PI-RADS, Prostate Imaging Reporting and Data System; PPV, positive predictive value; PSAD, PSA density; Sens, sensitivity; Spec, specificity.
Accuracy for GG ≥ 3 cancer remained consistent throughout pertinent clinical subgroups (Table 4A). The sensitivity of MPS2-AS was 96% to 100% across patients undergoing confirmatory and surveillance biopsies, patients < 65 and ≥ 65 years old, and Black and non-Black patients.
Table 4.
Clinical Performance Measures of MyProstateScore 2.0—Active Surveillance in Pertinent Subgroups at 11.5% Threshold
| Subgroup | No. | Upgraded | Sens (%) | Spec (%) | NPV (%) | PPV (%) |
| A. GG1 to GG ≥3 | ||||||
| Confirmatory Bx | 99 | 4 (4%) | 100 | 56 | 100 | 9 |
| Surveillance Bx | 221 | 26 (12%) | 96 | 68 | 99 | 29 |
| Age <65 y | 108 | 6 (6%) | 100 | 68 | 100 | 15 |
| Age ≥65 y | 222 | 25 (11%) | 96 | 61 | 99 | 24 |
| Non-Black race | 294 | 26 (9%) | 96 | 65 | 99 | 21 |
| Black race | 36 | 5 (14%) | 100 | 48 | 100 | 24 |
| B. GG1 to GG ≥2 | ||||||
| Confirmatory Bx | 99 | 36 (36%) | 97 | 27 | 94 | 43 |
| Surveillance Bx | 221 | 85 (39%) | 94 | 38 | 91 | 48 |
| Age <65 y | 108 | 30 (28%) | 97 | 37 | 97 | 37 |
| Age ≥65 y | 222 | 93 (42%) | 95 | 32 | 89 | 50 |
| Non-Black race | 294 | 109 (37%) | 95 | 35 | 93 | 46 |
| Black race | 36 | 14 (39%) | 93 | 27 | 86 | 45 |
Abbreviations: Bx, biopsy; GG, Grade Group; NPV, negative predictive value; PPV, positive predictive value; Sens, sensitivity; Spec, specificity.
GG ≥ 2 Upgrading: Comparative Accuracy and Clinical Consequences of Testing
Among the 330 AS patients, 123 (37%) patients upgraded to GG ≥ 2, of which 74 (60%) were detected on systematic biopsy only, 15 (12%) on targeted biopsy only, and 43 (28%) on both systematic and targeted biopsy (Table 2B). AUROC values (95% CI) for PSA, PCPTrc, PSAD, and mpMRI were 0.60 (0.53-0.66), 0.64 (0.58-0.70), 0.67 (0.61-0.73), and 0.64 (0.58-0.71), respectively, as compared with 0.74 (0.68-0.79) for MPS2-AS (Supplementary Figure 3B, https://www.jurology.com). MPS2-AS provided a statistically significant improvement in discrimination over PSA (P < .001), PCPTrc (P = .004), PSAD (P = .005), and mpMRI (P = .01; Supplementary Table 4B, https://www.jurology.com).
Clinical use of MPS2-AS to determine the need for biopsy would have detected 95% of GG ≥ 2 upgrades while avoiding 34% of unnecessary biopsies (Table 3B). The use of mpMRI to determine the need for biopsy would have missed 35% (PI-RADS ≥3) to 45% (PI-RADS ≥4) of GG ≥ 2 upgrades. MPS2-AS performance was consistent across subgroups, with 93% to 97% sensitivity for GG ≥ 2 PCa (Table 4B).
DCA
DCA was performed to determine the net clinical benefit of prebiopsy testing relative to performing biopsy in all patients or no patients. MPS2-AS provided the greatest net reduction in unnecessary biopsies without failing to detect a single additional upgrading event (Figure) and the highest net clinical benefit (Supplementary Figure 4, https://www.jurology.com) across most pertinent threshold probabilities.
Figure.
Decision curve analysis plots illustrating the net reduction in biopsies performed per 1000 patients based on prebiopsy testing with MyProstateScore 2.0—Active Surveillance (MPS2-AS; red), multiparametric MRI (mpMRI; light blue), PSA density (PSAD; gray), Prostate Cancer Prevention Trial Risk Calculator (PCPTrc; green), and PSA (blue) for the outcomes of upgrading to Grade Group (GG) ≥ 3 (A) and GG ≥ 2 (B).
MPS2-AS and mpMRI
Combined Accuracy of MPS2-AS and mpMRI
In patients who underwent mpMRI (N = 280), concordant positive MPS2-AS and mpMRI provided additive value, as 28% of patients with concordant positive tests upgraded to GG ≥ 3, compared with 23% with positive MPS2-AS alone and 15% with positive mpMRI alone (Table 5A). By contrast, concordant negative results were associated with a 1.1% risk of GG ≥ 3 upgrading, which was higher than the 0.6% risk observed with a negative MPS2-AS alone. When discordant, MPS2-AS was more strongly associated with observed outcomes. GG ≥ 3 upgrading occurred in 0 of 68 (0%) patients with a negative MPS2-AS and positive mpMRI, as compared with 11% with positive MPS2-AS and negative mpMRI. Similar associations were observed for the outcome of GG ≥ 2 upgrading (Table 5B).
Table 5.
Individual and Combined Accuracy of MyProstateScore 2.0—Active Surveillance and Multiparametric MRI for the Detection of Upgrading
| Subgroup | PI-RADS | MPS2-AS ≤11.5% | MPS2-AS >11.5% | Total |
| A. GG1 to GG ≥3 | 1-2 | 1.1% (1/93) | 11% (4/37) | 3.8% (5/130) |
| 3-5 | 0% (0/68) | 28% (23/82) | 15% (23/150) | |
| Total | 0.6% (1/161) | 23% (27/119) | 10% (28/280) | |
| B. GG1 to GG ≥2 | 1-2 | 9.1% (4/44) | 40% (34/86) | 29% (38/130) |
| 3-5 | 5.6% (1/18) | 53% (70/132) | 47% (71/150) | |
| Total | 8.1% (5/62) | 48% (104/218) | 39% (109/280) |
Abbreviations: GG, Grade Group; MPS2-AS, MyProstateScore 2.0—Active Surveillance; PI-RADS, Prostate Imaging Reporting and Data System.
AUROC values (95% CI) for GG ≥ 3 upgrading were 0.73 (0.63-0.82) for mpMRI, 0.80 (0.72-0.88) for mpMRI plus MPS2-AS, and 0.82 (0.76-0.88) for MPS2-AS (Supplementary Figure 5, https://www.jurology.com). As such, addition of MPS2-AS to mpMRI improved AUROC by 7%, but the addition of mpMRI to MPS2-AS did not improve AUROC. For GG ≥ 2 upgrading, AUROC values were 0.64 (0.58-0.71) for mpMRI, 0.74 (0.68-0.79) for MPS2-AS, and 0.75 (0.69 to 0.81) for their combined use.
Additive Value of MPS2-AS in Patients With Existing mpMRI
Because an increasing number of patients undergo serial mpMRI during AS, we assessed the diagnostic performance and clinical outcomes of MPS2-AS testing in patients with known mpMRI results (Table 6). In patients with a negative mpMRI, the use of a positive MPS2-AS to trigger biopsy would have captured 80% of GG ≥ 3 cancers that would have been missed due to a false-negative mpMRI, while still avoiding 74% of unnecessary biopsies. In patients with PI-RADS 3 mpMRI, use of MPS2-AS would have avoided 66% of unnecessary biopsies while maintaining 100% sensitivity for GG ≥ 3 cancer. While patients with PI-RADS 4 to 5 mpMRI are routinely referred for biopsy, the use of MPS2-AS would have avoided 48% of unnecessary biopsies without missing a single diagnosis of GG ≥ 3 cancer (ie, 100% sensitivity; Table 6B).
Table 6.
Diagnostic Performance Characteristics of MyProstateScore 2.0—Active Surveillance in Patients With Known Prostate Imaging Reporting and Data System Classification
| Subgroup | No. | Upgraded | MPS2-AS performance in PI-RADS strata | Potentially missed/delayed upgrading (%) | Unnecessary Bx potentially avoided per 1000 patients | |||
| Sens (%) | Spec (%) | NPV (%) | PPV (%) | |||||
| A. GG1 to GG ≥3 | ||||||||
| PI-RADS 1-2 | 130 | 5 (4%) | 80 | 74 | 99 | 11 | 20 | 736 |
| PI-RADS 3 | 42 | 1 (2%) | 100 | 66 | 100 | 7 | 0 | 659 |
| PI-RADS 4-5 | 108 | 22 (20%) | 100 | 48 | 100 | 33 | 0 | 477 |
| B. GG1 to GG ≥2 | ||||||||
| PI-RADS 1-2 | 130 | 38 (29%) | 89 | 43 | 91 | 40 | 11 | 435 |
| PI-RADS 3 | 42 | 11 (26%) | 100 | 29 | 100 | 33 | 0 | 290 |
| PI-RADS 4-5 | 108 | 60 (56%) | 98 | 17 | 89 | 60 | 1.7 | 167 |
Abbreviations: Bx, biopsy; GG, Grade Group; MPS2-AS, MyProstateScore 2.0—Active Surveillance; NPV, negative predictive value; PI-RADS, Prostate Imaging Reporting and Data System; PPV, positive predictive value; Sens, sensitivity; Spec, specificity.
DISCUSSION
Despite the oncologic safety of AS, repeat biopsies pose a major risk of nonadherence, in addition to discomfort, bleeding, and infectious complications, all of which threaten the safety of this standard-of-care management approach.13,28 There is a pressing need to reduce the number of biopsies required during AS while maximizing timely identification of high-grade cancers that merit treatment. We herein describe multisite external validation of a non-DRE urinary test to inform the need for surveillance biopsies. For the critical diagnosis of GG ≥ 3 cancer, the sensitivity (97%) and specificity (64%) of MPS2-AS indicate that its use in the AS population would avoid 64% of unnecessary AS biopsies while maintaining immediate detection of 97% of GG ≥ 3 upgrades. By contrast, even the most sensitive use of mpMRI would have failed to identify 18% of GG ≥ 3 upgrades, while avoiding fewer unnecessary biopsies (50%). Moreover, mpMRI failed to detect 35% of upgrades to GG ≥ 2, as compared with only 4.9% based on MPS2-AS. Acknowledging the well-established limitations of mpMRI in AS,9-12,29,30 these data provide an alternative monitoring approach capable of reliably detecting upgrades and ruling out a substantial proportion of unnecessary biopsies.
The current findings from academic and community centers further emphasize the potential drawbacks of mpMRI-based monitoring. Indeed, at one expert academic center, Chu et al observed an NPV of only 79.5% for GG ≥ 2 upgrading.11 These findings were corroborated in a larger study of US veterans, in which the NPV of mpMRI was again 79% for patients with GG1 disease undergoing confirmatory or surveillance biopsy.30 Our findings are consistent with these and other mpMRI data, revealing NPV of 71% for PIRADS ≥ 3 in detecting GG ≥ 2 upgrading. Combined, these data clearly demonstrate that a negative mpMRI is not sufficient to rule out the need for surveillance biopsies. By contrast, patients with a negative MPS2-AS had a 0.6% risk of GG ≥ 3 upgrading (99% NPV) and 8.1% risk of GG ≥ 2 upgrading (92% NPV), providing confident rule-out ability in this population. Notably, a negative mpMRI did not provide added NPV beyond a negative MPS2-AS. On the other hand, in patients with a positive MPS2-AS result, a positive mpMRI provided modest additive value, increasing the risk of GG ≥ 3 upgrading from 23% with a positive MPS2-AS alone to 28% with positive MPS2-AS and mpMRI—and from 48% to 53%, respectively, for GG ≥ 2. Taken together, these findings provide a practical clinical testing pathway, whereby a negative MPS2-AS alone can confidently rule-out the need for mpMRI or biopsy, and a positive MPS2-AS can be augmented by existing or repeat mpMRI to guide sampling.
At the same time, these data further emphasize that systematic sampling remains essential in patients undergoing AS biopsies.9,31 We found that systematic biopsy detected the majority of upgrades, including 81% to GG ≥ 3 and 88% to GG ≥ 2, while targeted biopsy captured an additional 19% and 12% of upgrades, respectively. Notably, the proportion of upgrades detected by targeted biopsy was near the lower range of that reported from expert centers.32,33 Although these differences could reflect varying provider experience or other factors, these data spanning academic and community centers may better reflect real-world practice. Regardless, GG ≥ 2 upgrading was captured by either targeted or systematic biopsy in 56% of patients with PI-RADS 4 to 5 lesions, consistent with detection rates in the literature.34 It is thus not expected that higher yield of targeted sampling specifically would substantially impact the observed findings. Importantly, the literature describing the lack of informative mpMRI changes during AS suggests that a prior positive mpMRI is likely sufficient to aid in targeting, limiting the need for repeat mpMRI in patients indicated for biopsy based on biomarker testing.10,12,29 Altogether, the current data add to a large body of evidence emphasizing that mpMRI is not a reliable primary monitoring tool during AS.10,12,29 A more practical approach to monitoring could be to first use a highly sensitive, objective, noninvasive test to accurately determine the need for repeat biopsy altogether, with mpMRI applied subsequently to improve the yield of necessary biopsies. The current data suggest that MPS2-AS could be highly useful in this role, and by separately providing the likelihood of upgrading to GG ≥ 2 and GG ≥ 3, biopsy decision-making can be further personalized.
The AS population—comprised of men with low-grade, low-volume cancer—has long posed a fundamental challenge to diagnostic tools seeking to detect occult higher-grade disease.35 Indeed, it stands to reason that biomarkers developed to distinguish cancer from benign tissue have had limited success in identifying higher-grade cancers among a population of patients known to harbor cancer. While existing serum-based assays have demonstrated an association with upgrading,15,16 none have been shown to provide actionable rule-in or rule-out performance to confidently guide decision-making. Tissue-based genomic classifiers have similarly had little success in predicting upgrading,36 potentially due to focal heterogeneity and the variable results obtained depending on the tumor focus submitted for testing.37,38 Freitas and colleagues recently confirmed that molecular testing of the highest volume GG1 core could not distinguish between patients with and without GG ≥ 2 cancer elsewhere in the prostate.39 While tissue-based testing remains dependent on sample obtainment and selection, urine provides a comprehensive picture of gene expression from throughout the prostate. Combined with the recent identification of genes uniquely overexpressed in higher-grade cancer relative to GG1,17 both biologic and practical factors support the observed accuracy of MPS2-AS in the surveillance population.
This study has real and perceived limitations. First, this was a multisite study spanning academic and community settings. MRI interpretation was not centralized and was thus subject to site variation. While potentially perceived as a limitation, these are the same interpretations upon which clinical decisions are made for most patients in the United States. Importantly, the observed performance of mpMRI is consistent with limited mpMRI data available from combined academic and community settings, supporting their validity.40,41 Second, study inclusion was limited to patients undergoing biopsy as deemed necessary by their treating physician. While this could yield a higher-risk cohort relative to the overall AS population, all testing modalities (ie, PSA, PCPTrc, PSAD, mpMRI) were evaluated in this same cohort, and MPS2-AS performed favorably. Moreover, a substantial proportion of patients had undergone negative mpMRI, suggesting a broad distribution of risk. Still, these data cannot inform on patients that did not undergo biopsy, just as these data do not inform the comparative performance of MPS2 and mpMRI in the diagnostic setting. Third, MPS2-AS values were calculated retrospectively, and this study was not intended to inform the impact of MPS2-AS on the decision to undergo biopsy. Fourth, the study included a limited number of Black patients, precluding generalization to underrepresented populations. Still, the proportion of Black patients enrolled was consistent with the US population, and promisingly, MPS2-AS provided 100% sensitivity and NPV for GG ≥ 3 upgrading in Black men. Notably, data were not available on the number of previous biopsies performed or time from diagnosis in patients undergoing surveillance biopsies, which can better risk-stratify among AS patients.42 Moreover, histological subtype data, which can provide prognostic information in addition to GG, were not available. Finally, as described, the positive predictive value of MPS2-AS was aided by targeted sampling in 12% to 19% of upgrades. Still, available data suggest that repeat mpMRI provides new or higher-risk targets in a minority of patients and mpMRI changes are not consistently associated with upgrading,10,12,29 such that repeat mpMRI can be avoided in most cases in which biomarker testing indicates the need for biopsy. Importantly, the current analysis uniquely characterizes the accuracy of combined biomarker and mpMRI findings, a noted limitation of the existing literature.21
CONCLUSIONS
In this multisite, external validation study among patients on active surveillance, urinary MPS2-AS demonstrated remarkable accuracy to detect upgrading to higher-grade cancer. Importantly, MPS2-AS outperformed mpMRI on direct comparison, providing higher sensitivity for GG ≥ 3 and GG ≥ 2 upgrading and higher specificity for GG ≥ 3. Taken in the context of contemporary clinical practice, these findings support a role for this noninvasive test to guide the need for scheduled biopsies during AS.
Footnotes
Recusal: Drs Tosoian, Scarpato, Moses, and Salami are on the editorial board of The Journal of Urology® and were recused from the editorial and peer review processes.
Funding/Support: This study was funded by the Michigan-Vanderbilt Early Detection Research Network Biomarker Characterization Center (U2C-CA271854). Other sources of funding not directly involved in the conduct of this study include Prostate Cancer Foundation Young Investigator Awards (Drs Tosoian, Xiao), Michigan Prostate Specialized Program of Research Excellence (P50 CA186786), National Cancer Institute Outstanding Investigator Award (Dr Chinnaiyan; R35 CA231996), Howard Hughes Medical Institute (Dr Chinnaiyan), and the American Cancer Society (Dr Chinnaiyan). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Conflict of Interest Disclosures: Dr Barocas reported being on the advisory board for Pfizer, Lantheus, Lynx Dx, and Pacific Edge. Dr Chinnaiyan reported serving on the advisory boards of Tempus, Ascentage Pharmaceuticals, Medsyn therapeutics, Esanik, and RAAPTA therapeutics. Drs Chinnaiyan and Tosoian reported being scientific advisors and equity holders to Lynx Dx. Dr Cooperberg reported being scientific advisor to Tempus, Veracyte, GlaxoSmithKline, Pfizer, Bayer, and Janssen as well as equity holding in Lynx Dx. Dr Morgan reported being on the advisory board for Cleveland Diagnostics. Dr Palapattu reported being on the advisory board for Continuity Biosciences, SynDevRx, Taurus Diagnostics, and Lynx Dx, and investment in Continuity Biosciences and ImmunityBio. Dr Ross reported being a consultant or speaker for Astellas, Blue Earth, Bayer, Janssen, Pfizer, BilliontoOne, Lantheus, Veracyte, and Boston Scientific. Dr Salami reported being on the advisory committee for Bayer Pharma. Dr Salari reported research funding support from Convergent Genomics. Dr Meyers, Mr. Moore, Dr Cao, Mr. Bala, Mr. Oh, Dr Heaton reported being employees of Lynx Dx. No other conflicts of interest were reported.
Ethics Statement: Institutional Review Board approval was received for this study (IRB Protocol No. 20241020).
Author Contributions:
Conception and design: Tosoian, Meyers, Britton, Chevli, Joseph, Moses, Sanda, Cao, Bala, Oh, Heaton, Zheng, Chinnaiyan.
Critical revision of the manuscript for scientific and factual content: Tosoian, Meyers, Britton, Chevli, Rayford, Gadzinski, Joseph, Kasraeian, Sethi, Robinson, Assani, Rings, Pierce, Scarpato, Moses, Barocas, Chang, Penson, Sanda, Ross, Cooperberg, Salari, Liss, Zhang, Singhal, Salami, Morgan, Palapattu, Wei, Cao, Bala, Heaton, Zheng, Chinnaiyan.
Data analysis and interpretation: Tosoian, Meyers, Moore, Chevli, Rayford, Joseph, Pierce, Barocas, Chang, Penson, Ross, Cooperberg, Salari, Liss, Zhang, Singhal, Salami, Morgan, Palapattu, Wei, Cao, Bala, Oh, Heaton, Zheng.
Data acquisition: Tosoian, Meyers, Moore, Chevli, Rayford, Gadzinski, Joseph, Patel, Kasraeian, Sethi, Robinson, Assani, Rings, Scarpato, Wei, Bala, Oh, Heaton.
Drafting the manuscript: Tosoian, Meyers, Moore, Britton, Chevli, Rayford, Patel, Cao, Bala, Oh, Heaton.
Statistical analysis: Tosoian, Moore, Britton, Zhang, Cao, Bala, Oh, Zheng.
Supervision: Tosoian, Meyers, Chevli, Rayford, Gadzinski, Joseph, Patel, Kasraeian, Sethi, Assani, Rings, Pierce, Scarpato, Moses, Barocas, Chang, Penson, Sanda, Ross, Salari, Liss, Singhal, Salami, Morgan, Palapattu, Wei, Heaton, Zheng, Chinnaiyan.
Other: Robinson, Cooperberg.
Data Availability: The datasets generated during and/or analyzed during the current study are not publicly available, but are available from the corresponding author on reasonable request.
REFERENCES
- 1.Eastham JA, Auffenberg GB, Barocas DA, et al. Clinically localized prostate cancer: AUA/ASTRO guideline, part I: introduction, risk assessment, staging, and risk-based management. J Urol. 2022;208(1):10-18. doi: 10.1097/JU.0000000000002757 [DOI] [PubMed] [Google Scholar]
- 2.Cornford P, van den Bergh RCN, Briers E, et al. EAU-EANM-ESTRO-ESUR-ISUP-SIOG guidelines on prostate cancer—2024 update. Part I: screening, diagnosis, and local treatment with curative intent. Eur Urol. 2024;86(2):148-163. doi: 10.1016/j.eururo.2024.03.027 [DOI] [PubMed] [Google Scholar]
- 3.Spratt DE, Srinivas S, Adra N, et al. Prostate cancer, version 3.2026, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw. 2025;23(11):469-493. doi: 10.6004/jnccn.2025.0052 [DOI] [PubMed] [Google Scholar]
- 4.Pattenden TA, Samaranayke D, Morton A, et al. Modern active surveillance in prostate cancer: a narrative review. Clin Genitourin Cancer. 2023;21(1):115-123. doi: 10.1016/j.clgc.2022.09.003 [DOI] [PubMed] [Google Scholar]
- 5.Cooperberg MR, Meeks W, Fang R, Gaylis FD, Catalona WJ, Makarov DV. Time trends and variation in the use of active surveillance for management of low-risk prostate cancer in the US. JAMA Netw Open. 2023;6(3):e231439. doi: 10.1001/jamanetworkopen.2023.1439 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Al Hussein Al Awamlh B, Barocas DA, Zhu A, et al. Use of active surveillance vs definitive treatment among men with low- and favorable intermediate-risk prostate cancer in the US between 2010 and 2018. JAMA Intern Med. 2023;183(6):608-611. doi: 10.1001/jamainternmed.2022.7100 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Ross AE, Loeb S, Landis P, et al. Prostate-specific antigen kinetics during follow-up are an unreliable trigger for intervention in a prostate cancer surveillance program. J Clin Oncol. 2010;28(17):2810-2816. doi: 10.1200/JCO.2009.25.7311 [DOI] [PubMed] [Google Scholar]
- 8.Herrera-Caceres JO, Wettstein MS, Goldberg H, et al. Utility of digital rectal examination in a population with prostate cancer treated with active surveillance. Can Urol Assoc J. 2020;14(9):E453-E457. doi: 10.5489/cuaj.6341 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Ma TM, Tosoian JJ, Schaeffer EM, et al. The role of multiparametric magnetic resonance imaging/ultrasound fusion biopsy in active surveillance. Eur Urol. 2017;71(2):174-180. doi: 10.1016/j.eururo.2016.05.021 [DOI] [PubMed] [Google Scholar]
- 10.Chesnut GT, Vertosick EA, Benfante N, et al. Role of changes in magnetic resonance imaging or clinical stage in evaluation of disease progression for men with prostate cancer on active surveillance. Eur Urol. 2020;77(4):501-507. doi: 10.1016/j.eururo.2019.12.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Chu CE, Lonergan PE, Washington SL, et al. Multiparametric magnetic resonance imaging alone is insufficient to detect grade reclassification in active surveillance for prostate cancer. Eur Urol. 2020;78(4):515-517. doi: 10.1016/j.eururo.2020.06.030 [DOI] [PubMed] [Google Scholar]
- 12.Rajwa P, Pradere B, Quhal F, et al. Reliability of serial prostate magnetic resonance imaging to detect prostate cancer progression during active surveillance: a systematic review and meta-analysis. Eur Urol. 2021;80(5):549-563. doi: 10.1016/j.eururo.2021.05.001 [DOI] [PubMed] [Google Scholar]
- 13.Ehdaie B, Vertosick E, Spaliviero M, et al. The impact of repeat biopsies on infectious complications in men with prostate cancer on active surveillance. J Urol. 2014;191(3):660-664. doi: 10.1016/j.juro.2013.08.088 [DOI] [PubMed] [Google Scholar]
- 14.Moore CM, King LE, Withington J, et al. Best current practice and research priorities in active surveillance for prostate cancer—a report of a Movember International Consensus Meeting. Eur Urol Oncol. 2023;6(2):160-182. doi: 10.1016/j.euo.2023.01.003 [DOI] [PubMed] [Google Scholar]
- 15.Hougen HY, Reis IM, Han S, et al. Evaluating 4Kscore's role in predicting progression on active surveillance for prostate cancer independently of clinical information and PIRADS score. Prostate Cancer Prostatic Dis. 2025;28(1):180-186. doi: 10.1038/s41391-024-00898-w [DOI] [PubMed] [Google Scholar]
- 16.de la Calle CM, Jing Y, Mamawala MM, et al. Baseline Prostate Health Index risk category and risk category changes during active surveillance predict grade reclassification. Urol Oncol. 2023;41(11):455 e1-455.e6. doi: 10.1016/j.urolonc.2023.08.011 [DOI] [PubMed] [Google Scholar]
- 17.Tosoian JJ, Zhang Y, Xiao L, et al. Development and validation of an 18-gene urine test for high-grade prostate cancer. JAMA Oncol. 2024;10(6):726-736. doi: 10.1001/jamaoncol.2024.0455 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Bossuyt PM, Reitsma JB, Bruns DE, et al. STARD 2015: an updated list of essential items for reporting diagnostic accuracy studies. BMJ. 2015;351:h5527. doi: 10.1136/bmj.h5527 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Moons KG, Altman DG, Reitsma JB, et al. Transparent reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): explanation and elaboration. Ann Intern Med. 2015;162(1):W1-W73. doi: 10.7326/M14-0698 [DOI] [PubMed] [Google Scholar]
- 20.Wang S, Kozarek J, Russell R, et al. Diagnostic performance of prostate-specific antigen density for detecting clinically significant prostate cancer in the era of magnetic resonance imaging: a systematic review and meta-analysis. Eur Urol Oncol. 2024;7(2):189-203. doi: 10.1016/j.euo.2023.08.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Moses KA, Sprenkle PC, Bahler C, et al. National Comprehensive Cancer Network. Prostate Cancer Early Detection (Version 2.2025). https://www.nccn.org/professionals/physician_gls/pdf/prostate_detection.pdf/ [Google Scholar]
- 22.Tosoian JJ, Zhang Y, Meyers JI, et al. Clinical validation of MyProstateScore 2.0 testing using first-catch, non-digital rectal examination urine. J Urol. 2025;213(5):581-589. doi: 10.1097/JU.0000000000004421 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Pepe MS, Feng Z, Janes H, Bossuyt PM, Potter JD. Pivotal evaluation of the accuracy of a biomarker used for classification or prediction: standards for study design. J Natl Cancer Inst. 2008;100(20):1432-1438. doi: 10.1093/jnci/djn326 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Weinreb JC, Barentsz JO, Choyke PL, et al. PI-RADS prostate imaging – reporting and data system: 2015, version 2. Eur Urol. 2016;69(1):16-40. doi: 10.1016/j.eururo.2015.08.052 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Ankerst DP, Hoefler J, Bock S, et al. Prostate Cancer Prevention Trial risk calculator 2.0 for the prediction of low- vs high-grade prostate cancer. Urology. 2014;83(6):1362-1367. doi: 10.1016/j.urology.2014.02.035 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Vickers AJ, van Calster B, Steyerberg EW. A simple, step-by-step guide to interpreting decision curve analysis. Diagn Progn Res. 2019;3:18. doi: 10.1186/s41512-019-0064-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Vickers AJ, Van Calster B, Steyerberg EW. Net benefit approaches to the evaluation of prediction models, molecular markers, and diagnostic tests. BMJ. 2016;352:i6. doi: 10.1136/bmj.i6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Beckmann K, Cahill D, Brown C, Van Hemelrijck M, Kinsella N. Understanding reasons for non-adherence to active surveillance for low-intermediate risk prostate cancer. Transl Androl Urol. 2021;10(6):2728-2736. doi: 10.21037/tau-20-1254 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Shee K, Song JJ, Cowan JE, et al. The natural history of confirmed grade group 1 prostate cancer managed with active surveillance in the modern era. Eur Urol Oncol. 2025;8(4):1094-1100. doi: 10.1016/j.euo.2025.06.005 [DOI] [PubMed] [Google Scholar]
- 30.Cooperberg MR, Bihn JR, Culnan JM, et al. Magnetic resonance imaging or confirmatory biopsy for patients with prostate cancer receiving active surveillance. JAMA Oncol. 2026;12(1):111-113. doi: 10.1001/jamaoncol.2025.4167 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Ahdoot M, Naser-Tavakolian A, Heard JR, et al. Prostate-specific antigen density to select magnetic resonance imaging–targeted vs combined biopsy for prostate cancer diagnosis: a secondary analysis of the Trio study. J Urol. 2025;213(6):693-701. doi: 10.1097/JU.0000000000004480 [DOI] [PubMed] [Google Scholar]
- 32.Kinnaird A, Yerram NK, O'Connor L, et al. Magnetic resonance imaging-guided biopsy in active surveillance of prostate cancer. J Urol. 2022;207(4):823-831. doi: 10.1097/JU.0000000000002343 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Chappidi MR, Newcomb LF, Zheng Y, et al. Magnetic resonance imaging at second surveillance biopsy after diagnosis in patients with Grade Group 1 prostate cancer in the Canary Prostate Active Surveillance Study. J Urol. 2025;214(3):251-258. doi: 10.1097/JU.0000000000004592 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Park KJ, Choi SH, Lee JS, Kim JK, Kim MH, Jeong IG. Risk stratification of prostate cancer according to PI-RADS® version 2 categories: meta-analysis for prospective studies. J Urol. 2020;204(6):1141-1149. doi: 10.1097/JU.0000000000001306 [DOI] [PubMed] [Google Scholar]
- 35.Tosoian JJ, Loeb S, Kettermann A, et al. Accuracy of PCA3 measurement in predicting short-term biopsy progression in an active surveillance program. J Urol. 2010;183(2):534-538. doi: 10.1016/j.juro.2009.10.003 [DOI] [PubMed] [Google Scholar]
- 36.Lin DW, Zheng Y, McKenney JK, et al. 17-Gene genomic prostate score test results in the Canary Prostate Active Surveillance Study (PASS) cohort. J Clin Oncol. 2020;38(14):1549-1557. doi: 10.1200/JCO.19.02267 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Salami SS, Hovelson DH, Kaplan JB, et al. Transcriptomic heterogeneity in multifocal prostate cancer. JCI Insight. 2018;3(21):e123468. doi: 10.1172/jci.insight.123468 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Wei L, Wang J, Lampert E, et al. Intratumoral and intertumoral genomic heterogeneity of multifocal localized prostate cancer impacts molecular classifications and genomic prognosticators. Eur Urol. 2017;71(2):183-192. doi: 10.1016/j.eururo.2016.07.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Freitas PFS, Khandekar A, Porto JG, et al. Can genomic classifiers in biopsy cores with Grade Group 1 cancer predict higher-grade disease elsewhere in the prostate? Results from the prospective Miami Active Surveillance Trial. J Urol. 2025;214(4):383-392. doi: 10.1097/JU.0000000000004652 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Sonn GA, Fan RE, Ghanouni P, et al. Prostate magnetic resonance imaging interpretation varies substantially across radiologists. Eur Urol Focus. 2019;5(4):592-599. doi: 10.1016/j.euf.2017.11.010 [DOI] [PubMed] [Google Scholar]
- 41.Zhu A, Srivastava A, DiBianco JM, et al. PD38-01 Negative predictive value of prostate MRI in real world practice: results from a statewide surgical collaborative. J Urol. 2023;209(Suppl 4):e993. doi: 10.1097/JU.0000000000003336.01 [DOI] [Google Scholar]
- 42.Alam R, Carter HB, Landis P, Epstein JI, Mamawala M. Conditional probability of reclassification in an active surveillance program for prostate cancer. J Urol. 2015;193(6):1950-1955. doi: 10.1016/j.juro.2014.12.091 [DOI] [PMC free article] [PubMed] [Google Scholar]