Annals of Internal MedicinePredicting Long-Term Risk for Prostate Cancer Mortality Following a Prostate-Specific Antigen Screening Test: Prognostic Model Development and External Validation

New prognostic model predicts 20-year prostate cancer death risk from single PSA test with 77.6% accuracy in 174,787-patient validation, which is 13% better than existing tools. Model integrates life expectancy factors (age, smoking, diabetes, hypertension) to guide who benefits from screening versus who faces greater non-cancer death risk.

🩺 CLINICAL CONSIDERATIONS

• Current PSA interpretation ignores competing mortality risks: men with 5-year life expectancy screened identically to healthy 55-year-olds despite negligible benefit
• Model outperformed Prostate Biopsy Collaborative Group tool by 2.7 percentage points (AUC 0.776 vs 0.749, p=0.031) in predicting 20-year mortality
• Race, family history, and PSA level predict cancer death: age, BMI, smoking, hypertension, diabetes, stroke predict other-cause mortality for net benefit calculation
• 29.5-year follow-up in screening trial plus 20-year VA validation provides unprecedented long-term data for slow-growing prostate cancers

💊 PRACTICE APPLICATIONS

• Integrate comorbidity assessment into PSA screening decisions using validated life expectancy predictors
• Counsel patients that screening benefit requires 10+ year life expectancy to outweigh harms
• Document competing mortality risks (diabetes, cardiovascular disease, smoking) when ordering PSA tests
• Consider deferring PSA screening in men with limited life expectancy regardless of age

More on Prostate Cancer