The CAPRA-S score: A straightforward tool for improved prediction of outcomes after radical prostatectomy

Matthew R Cooperberg, Joan F Hilton, Peter R Carroll, Matthew R Cooperberg, Joan F Hilton, Peter R Carroll

Abstract

Background: The authors previously developed and validated the Cancer of the Prostate Risk Assessment (CAPRA) score to predict prostate cancer recurrence based on pretreatment clinical data. They aimed to develop a similar postsurgical score with improved accuracy via incorporation of pathologic data.

Methods: A total of 3837 prostatectomy patients in the Cancer of the Prostate Strategic Urologic Research Endeavor (CaPSURE™) national disease registry were analyzed. Cox regression was used to determine the predictive power of preoperative prostate-specific antigen (PSA), pathologic Gleason score (pGS), surgical margins (SM), extracapsular extension (ECE), seminal vesicle invasion (SVI), and lymph node invasion (LNI). Points were assigned based on the relative weights of these variables in predicting recurrence. The new postsurgical score (CAPRA-S) was tested and compared with a commonly cited nomogram with proportional hazards analysis, concordance (c) index, calibration plots, and decision-curve analysis.

Results: Recurrence appeared in 16.8% of the men; actuarial progression-free probability at 5 years was 78.0%. The CAPRA-S was determined by adding up to 3 points for PSA, up to 3 points for pGS, 1 point each for ECE and LNI, and 2 points each for SM and SVI. The hazard ratio for each point increase in CAPRA-S score was 1.54 (95% confidence interval, 1.49-1.59), indicating a 2.4-fold increase in risk for each 2-point increase in score. The CAPRA-S c-index was 0.77, substantially higher than 0.66 for the pretreatment CAPRA score and comparable to 0.76 for the nomogram. The CAPRA-S score performed better in both calibration and decision curve analyses.

Conclusions: The CAPRA-S offers good discriminatory accuracy, calibration, and ease of calculation for clinical and research settings.

Copyright © 2011 American Cancer Society.

Figures

Figure 1
Figure 1
The CAPRA-S score. Points are assigned for each variable: up to 3 for prostate specific antigen (PSA) level in ng/ml, up to 3 for pathologic Gleason score, 2 each for positive surgical margin (SM) and seminal vesicle invasion (SVI), and 1 each for extracapsular extension (ECE) and lymph node invasion (LNI). Points are summed to yield the CAPRA-S score.
Figure 2
Figure 2
A. Biochemical progression-free probability following radical prostatectomy, stratified by CAPRA-S score. B. Biochemical progression-free probability following radical prostatectomy, stratified by grouped CAPRA-S score: 0–2 indicates relatively low-risk, 3–5 intermediate-risk, and ≥6 high-risk disease in terms of likelihood of biochemical progression.
Figure 2
Figure 2
A. Biochemical progression-free probability following radical prostatectomy, stratified by CAPRA-S score. B. Biochemical progression-free probability following radical prostatectomy, stratified by grouped CAPRA-S score: 0–2 indicates relatively low-risk, 3–5 intermediate-risk, and ≥6 high-risk disease in terms of likelihood of biochemical progression.
Figure 3
Figure 3
Decision curve analysis comparing CAPRA-S to postoperative nomogram. The solid blue line indicates the CAPRA-S score predictions, the dashed red line the nomogram predictions. Across the various threshold probabilities, CAPRA-S has greater net benefit (i.e., the net increase in the proportion of patients who would be appropriately identified for adjuvant treatment using the prediction model) than the nomogram. The gray line represents the assumption that all patients would recur, the black line that none would recur. The segment of the nomogram curve which falls below the gray line suggests suboptimal calibration—specifically, over-optimistic prediction—at low predicted levels of risk.
Figure 4
Figure 4
Calibration plot comparing observed (Kaplan-Meier estimates and 95% confidence intervals) with predicted (model-based) progression-free probabilities (PGP) at 5 years’ follow-up. The 45° reference line represents perfect calibration, and labels at each point indicate the score level assessed. Panel A: CAPRA-S, Panel B: Stephenson nomogram
Figure 4
Figure 4
Calibration plot comparing observed (Kaplan-Meier estimates and 95% confidence intervals) with predicted (model-based) progression-free probabilities (PGP) at 5 years’ follow-up. The 45° reference line represents perfect calibration, and labels at each point indicate the score level assessed. Panel A: CAPRA-S, Panel B: Stephenson nomogram

Source: PubMed

Sponzoři a spolupracovníci

Zdravotní podmínky

Drogové intervence

3
Předplatit