Table 4. Multivariate logistic regression models to predict radical prostatectomy (RP) Gleason score (GS) based on biopsy GS: (a) Prediction of upgrading from biopsy GS 5–6 to GS >6 at RP (n = 2688); (b) prediction of downgrading from biopsy GS 3 + 4 = 7 to GS 5–6 at RP (n = 943).
| a. | ||
|---|---|---|
|
| ||
| Variable | OR (95% CI) | p value |
| Age, yr | 1.057 (1.044–1.069) | <0.0001 |
| PSA, ng/ml | 1.113 (1.104–1.163) | < 0.0001 |
| Maximum % cancer/core | 1.010 (1.008–1.013) | <0.0001 |
| Pathology weight | 0.976 (0.972–0.981) | <0.0001 |
| c-index of overall model: 0.685 | ||
|
| ||
| b. | ||
|
| ||
| Variable | OR (95% CI) | p value |
|
| ||
| PSA, ng/ml | 0.930 (0.870–0.993) | 0.031 |
| Maximum % cancer/core | 0.989 (0.983–0.996) | 0.002 |
| Pathology weight | 1.010 (1.001–1.019) | 0.019 |
| c-index of overall model: 0.629 | ||
OR = odds ratio; CI = confidence interval; PSA = pro state-specific antigen; c-index = concordance index.