Bold warning: MRI alone is not ready to replace biopsies in active surveillance for prostate cancer.
In plain terms, current prostate MRI still isn’t accurate enough to drop biopsies during ongoing monitoring. The negative predictive value (NPV) of MRI—how confidently it can say there’s no higher-grade cancer—stays below the 90% threshold typically desired to safely skip biopsy. Across nearly 2,000 patients, MRI’s NPV was 75%–77% for predicting grade group 2 or higher at confirmatory and surveillance biopsies, and 79% for those with grade group 1 at diagnostic biopsy. This margin of error is particularly concerning for Black patients and for cases where PI-RADS scoring may not clearly reclassify risk.
Until improvements are broadly implemented, routine confirmatory biopsies should remain part of active surveillance. This is not a minor caveat: it’s about safeguarding against missing clinically meaningful cancers that could otherwise go undetected if MRI were used alone.
Experts acknowledge that there’s no definitive MRI-only follow-up standard yet. Opinions vary on what NPV should be deemed acceptable for ruling out further investigation; many clinicians aim for 90% or higher, especially when using a test to avoid additional procedures. Some researchers note that the challenges aren’t solely about MRI technology itself but also about how consistently PI-RADS is applied and interpreted. Artificial intelligence tools are progressing rapidly, with initiatives like PI-CAI showing strong performance in early assessments, sometimes surpassing radiologists in initial benchmarks.
A randomized trial would be ideal to settle the debate, but it would take years to yield meaningful outcomes. In the meantime, integrating new imaging tools within a rigorous biomarker framework—such as REMARK—appears to be the practical path forward. For now, biopsy remains the reference standard to validate MRI’s accuracy.
Current guidelines reflect this cautious stance. While some recommendations suggest MRI could eventually replace confirmatory biopsy, the evidence isn’t robust enough yet to support a broad shift away from biopsy in active surveillance. A notable single-center study reported lower MRI-based NPVs, particularly in patients with higher PSA density, underscoring that MRI alone may miss significant disease.
In the study examining 1,901 U.S. veterans with grade group 1–2 at diagnostic biopsy (2013–2023), all had MRI within 180 days before confirmatory and/or surveillance biopsy. The investigation tracked MRI’s NPV using PI-RADS ≤2 as negative and PI-RADS ≥3 as positive. Over time, pre-biopsy MRI use increased substantially, but MRI’s NPV for confirmatory or surveillance biopsy remained under 80% overall. Grade group 2 at diagnostic biopsy lowered NPV further, with particular vulnerability in higher PSA density subgroups. Sensitivity analyses adjusting PI-RADS thresholds slightly improved NPV but did not reach the levels needed to forgo biopsy reliably.
The researchers concluded that high-quality multiparametric MRI can aid targeted biopsy decisions and improve diagnostic accuracy, yet relying on MRI as a biopsy substitute in active surveillance would risk underdiagnosing clinically significant cancers—though such cancers may not be immediately life-threatening. The takeaway is clear: until MRI proves consistently reliable enough, biopsy continues to be essential in active surveillance to ensure accurate risk assessment and timely intervention when needed.
Would you like a version tailored for patients, clinicians, or policymakers, with specific examples and practical takeaways for decision-making?
