CAPRA and Epstein answer different questions

The CAPRA + Epstein prostate risk panel displays two separately source-locked results. It does not create a combined score, treatment recommendation, or consensus category.

What each component answers

| Component | Source-specific purpose | Important boundary | |---|---|---| | Original UCSF-CAPRA | Pretreatment recurrence-risk association in the original model | It is not CAPRA-S and does not determine treatment | | Johns Hopkins Epstein cT1c criteria | Historical prediction of a pathology endpoint at prostatectomy | It is not modern active-surveillance eligibility |

The former embedded NCCN v4.2024 branch was removed because its simplified local rules were not supported as a reproducible source-locked implementation. The current panel does not implement, reproduce, or infer NCCN prostate risk groups.

The Epstein source distinction

The 1994 Johns Hopkins JAMA publication derived a model for tumor extent in nonpalpable cT1c prostate cancer. The familiar four failure thresholds were operationalized prospectively in 1997:

• PSA density 0.15 or greater
• Gleason score 7 or greater
• 3 or more involved cores
• 50% or more involvement of any core

The verified interpreter therefore meets the historical operational criteria only when PSA density is below 0.15, there is no Gleason pattern 4 or 5, fewer than 3 cores are involved, and maximum core involvement is below 50%. Exact equality fails. The route does not present this simplified checklist as the literal full 1994 derivation.

Scope and limitations

This interpreter addresses a historical pathology endpoint: prediction of pathologically insignificant disease in the prostatectomy specimen.

The Epstein interpretation is limited to newly diagnosed acinar adenocarcinoma, cT1c, pretreatment assessment, comparable systematic needle-biopsy sampling, and contemporaneous PSA density. Targeted-only or otherwise noncomparable sampling, other stages or histologies, and treated contexts are outside its historical scope.

Meeting the criteria does not make an automatic surveillance candidate. Not meeting them does not diagnose clinically significant cancer or mean treatment is needed. Contemporary MRI, biopsy technique, molecular testing, comorbidity, preferences, and current guidelines require separate clinical assessment.

Use the source-specific tools

• Johns Hopkins Epstein T1c Pretreatment Criteria
• Original UCSF-CAPRA Score
• CAPRA + Epstein Prostate Risk Panel

Sources

1. Epstein JI, Walsh PC, Carmichael M, Brendler CB. JAMA. 1994;271(5):368-374. JAMA abstract.
2. Carter HB, Sauvageot J, Walsh PC, Epstein JI. J Urol. 1997;157(6):2206-2209. PubMed PMID 9146616.
3. Lee MC, et al. BJU Int. 2011;108(4):518-525. PubMed PMID 21320276.

This content explains software implementation and source scope. It is not prospective clinical validation, independent clinical validation, or medical advice.