This retrospective, MRI-driven active surveillance (AS) study enrolled 619 men from two academic centres in the UK and USA between 2012 and 2024. Median follow-up was 6.4 years. The study team defined the composite primary outcome—AS failure—as disease-specific mortality, metastatic disease, progression to grade group (GG) ≥4, or post-treatment biochemical recurrence. Across the cohort, 39 of 619 men (6%) met that composite endpoint during follow-up.
The investigators analysed cumulative incidence and used multivariable Fine‑Gray competing-risk models to test whether imaging and biopsy features predicted AS failure. The source reports that analyses focused on final MRI and biopsy status and on the timing of cribriform detection (baseline versus follow-up). The report is a preprint and has not yet been peer reviewed.
At baseline, cribriform architecture at grade group ≥2 was documented in 24 of 619 men (4%). On follow-up biopsy, cribriform ≥GG2 was detected in 82 of 619 men (13%). The study team noted this roughly three-fold increase in prevalence at re-biopsy and attributed the difference to baseline sampling error rather than incident emergence of cribriform morphology.
Among baseline diagnostic categories, the follow-up cribriform detection rate was comparable in men originally classified as GG1 and those with non-cribriform GG2 disease, according to the source. The authors reported that newly detected cribriform tumours occurred exclusively in lesions that were MRI-visible.
Detection of cribriform ≥GG2 on a follow-up biopsy carried a substantially higher short-term risk of AS failure. The investigators reported a three-year AS-failure cumulative incidence of 31% after the final biopsy in men with cribriform ≥GG2. By contrast, men with stable disease or with non-cribriform histology exhibited three-year failure rates in the 0–2% range.
The source also described that 44% of AS failures occurred despite a stable MRI appearance, indicating that MRI stability did not exclude subsequent adverse outcomes in a notable fraction of events. The study therefore separates the concepts of MRI progression and histologic progression—reporting that biopsy-detected cribriform disease showed stronger association with failure than imaging-based progression categories.
In Fine‑Gray competing-risk models restricted to re-biopsied men, the authors identified cribriform ≥GG2 on follow-up biopsy as the only independent predictor of AS failure. The reported subdistribution hazard ratio for cribriform ≥GG2 was 30 (95% CI 5.2–167; P < 0.001). No MRI progression category—by the classifications the investigators used—retained independent predictive value in the multivariable models.
The team also examined a subgroup of 595 men who were cribriform-negative at baseline. For these men, a second biopsy that revealed cribriform ≥GG2 was associated with a six-year AS-failure incidence of 34%, compared with 7% in those who remained cribriform-negative at repeat sampling.
Because newly detected cribriform tumours in this cohort were reported only in MRI-visible lesions, the authors propose that early repeat biopsy in patients with an MRI-visible lesion could materially improve risk stratification. The data suggest histologic sampling—specifically identification of cribriform architecture—may add prognostic discrimination beyond serial MRI interpretation alone.
The study therefore frames cribriform status on follow-up biopsy as a high-risk histologic feature in the contemporary MRI-era AS pathway. Investigators reported that MRI progression categories, as applied here, did not independently predict the composite failure outcome when cribriform status was included in the model.
The source is a retrospective analysis from two academic centres and the report itself acknowledges this design. The preprint states that all data produced in the study are available upon reasonable request to the authors. The investigators also note ethical approvals from institutional review boards at Cambridge University Hospitals and the University of California San Diego; these approvals are documented in the manuscript metadata.
Several practical questions remain unanswered in the preprint. The authors do not provide prospective validation of a specific biopsy-timing algorithm, nor do they report protocolized thresholds for when to perform early repeat biopsy beyond the descriptive observation that newly detected cribriform tumours were MRI-visible. The study team also reports that MRI progression categories were not independently predictive in their models, but the manuscript does not include external validation of MRI assessment protocols or standardized imaging criteria across the participating centres.
Clinicians and guideline panels will likely want confirmation in prospective or multi-institutional validation cohorts before altering AS schedules. The preprint has not undergone peer review; that status should temper immediate clinical translation. The authors stated competing interests as none and declared that data can be requested for further evaluation.