Negative Predictive Value of a Prostate MRI in Black Men: Implications for Biopsy Decision-Making

Sarah Sandlow, Samuel Carbunaru, Zequn Sun, Bernice Ofori, Y. C. Liu, Eric V. Li, Courtney M.p. Hollowell, Patricia Vidal, Kent T. Perry, Hiten D. Patel, Ryan W. Dobbs, David D. Casalino, Cecil G. Jawahar, Andy Wood, Channa Amarasekera, Jeanne M. Horowitz, Frank H. Miller, Emrah Gumusgoz, Ximing J. Yang, Bonnie Choy, Frances Manosca, Ulaş Bağcı, Edward M. Schaeffer, Peter H. Gann, Dustin D. French, Ashley E. Ross, Shilajit Kundu, Adam B. MurphyThe Journal of Urology 2025 · DOI

TL;DR

This study reveals that standard prostate MRI (mpMRI) protocols, particularly the PIRADS scoring system, are less effective at ruling out clinically significant prostate cancer (csPCa) in Black men compared to non-Black men. The findings suggest that adjusting PSA density (PSAD) thresholds is crucial for safe biopsy deferral in Black men, potentially preventing unnecessary biopsies while maintaining high cancer detection rates.

What Problem Does This Paper Address?

Prostate cancer (PCa) is a significant health concern, and while PSA screening is common, its low specificity leads to many unnecessary biopsies and overdiagnosis of low-risk cancers. Multiparametric MRI (mpMRI) with the Prostate Imaging Reporting and Data System (PIRADS) has emerged as a valuable tool to improve risk assessment and guide biopsy decisions, aiming to reduce unnecessary procedures. However, the effectiveness of these tools can vary across different populations. Black men, who have a higher prevalence of PCa and often more aggressive disease, have not been as extensively studied in the context of mpMRI's diagnostic accuracy. This paper addresses the critical gap in understanding whether standard mpMRI and PIRADS scoring perform equally well in Black men, and if not, how to adapt these tools to ensure equitable and effective PCa screening and management for this demographic.

Methods

The researchers conducted a comparative effectiveness study combining data from multiple cohorts: three IRB-approved prospective research cohorts and one retrospective clinical cohort. The study included men referred for outpatient urology clinics due to abnormal PSA or prostate examination, who subsequently underwent mpMRI and MRI-informed prostate biopsy. MRIs were scored using PIRADS version 2.0 or 2.1.

Key methodological aspects include:

  • Combined Cohorts: Merging data from different studies to increase statistical power.
  • Race Stratification: Explicitly comparing outcomes between Black and non-Black men.
  • PIRADS Scoring: Utilizing the standardized PIRADS system for MRI interpretation.
  • PSA Density (PSAD) Analysis: Investigating the role of PSAD, a measure of PSA concentration relative to prostate volume, in refining biopsy decisions.
  • Exploratory Post Hoc Analyses: Examining specific PSAD thresholds to achieve a target sensitivity (e.g., >90%) for biopsy deferral in Black men.

Datasets

The study utilized data from a combination of sources:

  • PHI Validation Study: (2017-2020) Biopsy-naïve Black and non-Hispanic White men with elevated PSA.
  • Nomogram Study: (2018-2021) Black men, biopsy-naïve or with prior negative biopsy, and elevated PSA.
  • PHI/MRI Study: (2021-present) Black/Hispanic biopsy-naïve men with elevated PSA or abnormal DRE.
  • Clinical Cohort: (2017-2021) Biopsy-naïve men with elevated PSA or abnormal DRE.

These cohorts were combined to form the final study population, which included 286 Black men and 965 non-Black men with PSA < 15.0 ng/mL. The paper does not introduce a new benchmark dataset but rather analyzes existing clinical and research data.

Key Insights and Ideas

  • Racial Disparities in mpMRI Performance: The most striking insight is that mpMRI, as interpreted by PIRADS, exhibits lower sensitivity and negative predictive value (NPV) in Black men compared to non-Black men. This means a negative MRI is less reliable in ruling out clinically significant PCa in Black individuals.
  • Higher Risk at Lower PIRADS Scores: Black men showed a higher risk of clinically significant PCa even at lower PIRADS scores (1-4), indicating that the current PIRADS scoring system may not be adequately calibrated for this population.
  • PSAD as a Crucial Modifier: The study strongly suggests that PSAD is a critical factor in refining biopsy decisions, especially for Black men with negative or equivocal MRIs. The commonly used PSAD threshold of 0.15 ng/mL/cm³ may be too high for safely deferring biopsies in this group.
  • Need for Race-Specific Thresholds: The findings underscore the necessity of developing race-specific thresholds for PSAD to ensure accurate biopsy recommendations and avoid missing cancers in Black men.

Key Results

  • Lower NPV and Sensitivity: In the combined cohort, the NPV for PIRADS ≥ 3 was 77.1% for Black men versus 87.6% for non-Black men (P = .03). Sensitivity was also lower in Black men (90.7%) compared to non-Black men (96.3%, P = .004).
  • Higher csPCa Rates: Black men had higher rates of clinically significant PCa (GG 2-5) across PIRADS scores 1-4, even when stratified by PSAD levels. For instance, with PSAD < 0.15 and PIRADS 1-2, 13.9% of Black men had csPCa versus 8.6% of non-Black men.
  • Revised PSAD Thresholds: Exploratory analyses indicated that a PSAD threshold of < 0.09 for PIRADS 1-2 lesions in Black men could achieve >90% sensitivity and NPV for detecting csPCa. For PIRADS 3 lesions, a PSAD < 0.07 further improved sensitivity and NPV.
  • Impact on Biopsy Deferral: The study suggests that current strategies based on PSAD < 0.15 might lead to missed csPCa in Black men. Lowering PSAD thresholds could help safely defer biopsies in more Black men with negative MRIs, though it may reduce specificity.

Limitations and Future Work

The authors acknowledge several limitations:

  • Potential Selection Bias: Differences in research trial populations and the use of a single academic medical center's clinical cohort could introduce bias.
  • Radiology Concordance: Relying on radiology concordance conferences rather than centralized reviews might introduce variability in PIRADS scoring.
  • Limited Analysis of Other Factors: Factors like family history were not extensively analyzed.
  • Generalizability: While efforts were made to combine diverse cohorts, the generalizability to all Black men, particularly those with different insurance statuses or access to care, needs further validation.

Future work suggested by the authors includes:

  • External Validation: Externally validating these findings in multicenter prospective cohorts is crucial.
  • Prospective Studies for PSAD Thresholds: Conducting prospective studies specifically powered to identify optimal PSAD thresholds for Black men with PIRADS 1-3 lesions.
  • Investigating MRI-Invisible csPCa: Elucidating the reasons for increased MRI-invisible csPCa in Black men.
  • Exploring Other Ethnic Groups: Investigating mpMRI accuracy in other ethnic groups with lower PCa prevalence, such as Asians.
  • Advanced Imaging Analysis: Further validating the hypothesis that apparent diffusion coefficient-weighted and T2-weighted imaging may be less accurate in Black men.

Summary updated

Abstract

PIRADS < 3 has a lower NPV and sensitivity in Black men. For negative prostate MRIs, PSAD ≥ 0.09 may be a better threshold for safe biopsy deferral in Black men to maintain a ≥ 90% sensitivity.
BibTeX
@article{sandlow2025negative,
  author = {Sarah Sandlow and Samuel Carbunaru and Zequn Sun and Bernice Ofori and Y. C. Liu and Eric V. Li and Courtney M.P. Hollowell and Patricia Vidal and Kent T. Perry and Hiten D. Patel and Ryan W. Dobbs and David D. Casalino and Cecil G. Jawahar and Andy Wood and Channa Amarasekera and Jeanne M. Horowitz and Frank H. Miller and Emrah Gumusgoz and Ximing J. Yang and Bonnie Choy and Frances Manosca and Ulaş Bağcı and Edward M. Schaeffer and Peter H. Gann and Dustin D. French and Ashley E. Ross and Shilajit Kundu and Adam B. Murphy},
  title = {{Negative Predictive Value of a Prostate MRI in Black Men: Implications for Biopsy Decision-Making}},
  journal = {The Journal of Urology},
  year = {2025},
  doi = {10.1097/ju.0000000000004498},
}
Figure