Table 3.
Results of external validation of the Adjuvant Colon Cancer End Points (ACCENT) nomogram for overall survival (OS) and comparison of five-year OS predictions for ACCENT vs Numeracy*
| Variable | N | Observed 5-year OS (%) | Median 5-year OS prediction (%) | % Delta: predicted – observed | % Correct: predicted vs observed | ||||
|---|---|---|---|---|---|---|---|---|---|
| K-M | 95% CI | ACCENT | Numeracy | ACCENT | Numeracy | ACCENT | Numeracy | ||
| Overall | 1903 | 78.4 | 76.5 to 80.3 | 77.2 | 75.0 | −1.2 | −3.4 | 79.2 | 77.5 | 
| Age, y | |||||||||
| ≤49 | 469 | 81.8 | 78.3 to 85.5 | 79.5 | 81.0 | −2.3 | −0.8 | 82.4 | 82.0 | 
| 50–59 | 606 | 81.1 | 78.0 to 84.3 | 79.7 | 78.0 | −1.4 | −3.1 | 81.1 | 79.5 | 
| 60–69 | 525 | 77.9 | 74.4 to 81.6 | 76.2 | 76.0 | −1.7 | −1.9 | 79.5 | 76.7 | 
| ≥70 | 303 | 68.4 | 63.3 to 73.9 | 68.1 | 74.0 | −0.3 | 5.6 | 70.2 | 68.8 | 
| Sex | |||||||||
| Male | 934 | 76.9 | 74.3 to 79.7 | 75.2 | 75.0 | −1.7 | −1.9 | 77.7 | 76.8 | 
| Female | 969 | 79.7 | 77.2 to 82.3 | 78.8 | 75.0 | −0.9 | −4.7 | 80.7 | 78.3 | 
| Grade | |||||||||
| Low | 1455 | 80.6 | 78.6 to 82.7 | 79.0 | 76.0 | −1.6 | −4.6 | 80.6 | 79.8 | 
| High | 448 | 70.9 | 66.7 to 75.3 | 69.6 | 67.0 | −1.3 | −3.9 | 74.6 | 70.3 | 
| T-stage | |||||||||
| T1/T2 | 298 | 89.4 | 85.9 to 93.0 | 88.4 | 82.0 | −1.0 | −7.4 | 89.0 | 89.0 | 
| T3 | 1443 | 78.3 | 76.2 to 80.5 | 76.1 | 74.0 | −2.2 | −4.3 | 78.7 | 77.3 | 
| T4 | 162 | 57.4 | 50.1 to 65.8 | 63.7 | 59.0 | 6.3 | 1.6 | 64.9 | 58.9 | 
| Pos. nodes | |||||||||
| 1–4 | 1338 | 84.1 | 82.1 to 86.1 | 80.6 | 78.0 | −3.5 | −6.1 | 83.7 | 83.5 | 
| 5+ | 565 | 64.9 | 61.0 to 69.0 | 65.0 | 57.0 | 0.1 | −7.9 | 68.7 | 63.7 | 
* Median predictions in bold fall outside the observed 95% K-M confidence interval (CI). Italicized comparisons indicate where Adjuvant Colon Cancer End Points (ACCENT) is worse than Numeracy (though in both cases, the ACCENT predictions are within the observed 95% CI). ACCENT = Adjuvant Colon Cancer End Points; K-M = Kaplan-Meier; OS = overall survival; Pos = positive.