Table 5.
Number of men with high-grade* tumor classified as low, middle, or high risk from predictive models with or without genetic score
| Risk categories of the model of clinical variables only | Risk categories of the model combining clinical variables and genetic score
|
NRI | p value** | |||||
|---|---|---|---|---|---|---|---|---|
| Low (1st quartile) | Middle (2nd and 3rd quartiles) | High (4th quartile) | Total | No. (%) reclassified as higher risk | No. (%) reclassified as lower risk | |||
| Low (1st quartile) | ||||||||
| No. of men | 334 | 72 | 0 | 406 | – | – | – | – |
| No. of negative | 327 | 70 | 0 | 397 | 70 (17.63) | N/A | – | – |
| No. of positive PCa | 7 | 2 | 0 | 9 | 2 (22.22) | N/A | – | – |
| 4-yr detection rate (95% CI) | 2 (1–4) | 3 (31 to 7) | 0 | 2 (1–4) | – | – | 0.046 | 0.772 |
| Middle (2nd and 3rd quartiles) | ||||||||
| No. of men | 72 | 666 | 74 | 812 | – | – | – | – |
| No. of negative | 69 | 630 | 64 | 763 | 64 (8.39) | 69 (9.04) | – | – |
| No. of positive PCa | 3 | 36 | 10 | 49 | 10 (20.41) | 3 (6.12) | – | – |
| 4-yr detection rate (95% CI) | 4 (0–9) | 5 (4–7) | 14 (6–21) | 6 (4–8) | – | – | 0.149 | 0.047 |
| High (4th quartile) | ||||||||
| No. of men | 0 | 74 | 332 | 406 | – | – | – | – |
| No. of negative | 0 | 68 | 274 | 342 | N/A | 68 (19.88) | – | – |
| No. of positive PCa | 0 | 6 | 58 | 64 | N/A | 6 (9.38) | – | – |
| 4-yr detection rate (95% CI) | 0 | 8 (2–14) | 17 (13–22) | 16 (12–19) | – | – | 0.105 | 0.413 |
| Total | ||||||||
| No. of men | 406 | 812 | 406 | 1624 | – | – | – | – |
| No. of negative | 396 | 768 | 338 | 1502 | 134 (8.92) | 137 (9.12) | – | – |
| No. of positive PCa | 10 | 44 | 68 | 122 | 12 (9.84) | 9 (7.38) | – | – |
| 4-yr detection rate (95% CI) | 2 (1–4) | 5 (4–7) | 17 (13–20) | 8 (6–9) | – | – | 0.027 | 0.497 |
NRI = net reclassification index; N/A = not applicable; PCa = prostate cancer; CI = confidence interval.
*
High grade is defined as Gleason grade ≥7.
**
p value is for testing H0: NRI = 0 via the z-score.