Table IV. Test operating characteristics of the ICPS in the LOOCV training, test and entire-data sets.
| Test operating characteristics of the ICPSa | Mortality predictor for IC (≤ −1.67)a,b |
Survival predictor for IC (>−1.67)a,b |
||||
|---|---|---|---|---|---|---|
| LOOCV training set (n = 22) | Test set (n = 23) | Entire-data setc (n = 45) | LOOCV training set (n = 22) | Test set (n = 23) | Entire-data setc (n = 45) | |
| Percentage (95% CI) | ||||||
| Sensitivity | 83 (68–99) | 83 (68–99) | 83 (72–94) | 100 | 94 (84–100) | 97 (92–100) |
| Specificity | 100 | 94 (84–100) | 97 (92–100) | 83 (68–99) | 83 (68–99) | 83 (72–94) |
| Positive predictive value | 100 | 83 (68–99) | 91 (82–99) | 94 (84–100) | 94 (84–100) | 94 (87–100) |
| Negative predictive value | 94 (84–100) | 94 (84–100) | 94 (87–100) | 100 | 83 (68–99) | 91 (82–99) |
| Accuracy | 95 (84–100) | 91 (80–100) | 93 (86–100) | 95 (84–100) | 91 (80–100) | 93 (86–100) |
| Classification error | 5 (3–7) | 9 (7–10) | 7 (5–8) | 5 (3–7) | 9 (7–10) | 7 (5–8) |
| Ratio (95% CI) | ||||||
| Positive likelihood ratio | ∞ | 14.17 (0.00–28.42) | 27.50 (14.45–40.55) | 6.00 (3.95–8.05) | 5.65 (3.62–7.67) | 5.82 (4.38–7.26) |
| Negative likelihood ratio | 0.17 (0.01–0.32) | 0.18 (0.02–0.33) | 0.17 (0.06—0.28) | 0.00 | 0.07 (0.05–0.09) | 0.04 (0.02–0.05) |
| Prognostic odds ratio | ∞d | 80.00 (63.65–96.35)d | 160.00 (52.89–267.11)d | ∞e | 80.00 (63.65–96.35)e | 160.00 (52.89–267.11)e |
| Mean±S.E. (95% CI) | ||||||
| ROC area | 0.97 ± 0.04 (0.90–1.00)f | 0.96 ± 0.04 (0.88–1.00)f | 0.96 ± 0.02 (0.91–1.00)g | 0.97 ± 0.04 (0.90–1.00)f | 0.96 ± 0.04 (0.88–1.00)f | 0.96 ± 0.02 (0.91–1.00)g |
a The ICPS is a simplified clinical-outcome prediction model for IC based on serum five-IgG antibody-reactivity signatures in IC patients at presentation (see Fig. 4). The test operating characteristics summarized in the table are for the ICPS modeled as a mortality or survival predictor for IC in order to identify IC patients who had a poor or good clinical outcome, respectively, in the ensuing 2-month period after presentation.
b Model cutoffs were defined on the basis of ROC curve analysis on training set data (see Fig. 4D). These discrimination threshold values correspond to the highest combined sensitivity and specificity (i.e., the minimum sum of false-negative plus false-positive test results) for poor or good clinical outcomes when mortality or survival predictors, respectively, for IC were determined.
c To minimize the likely overestimation of the prognostic value of the ICPS, data derived both from LOOCV training set rather than original training set and from the test set were considered in the entire data set (see Fig. 5C).
d The odds for a positive test result in IC patients who died within 2 months relative to the odds of a positive result in those who had a favorable clinical outcome.
e The odds for a positive test result in IC patients who had a favorable clinical outcome relative to the odds of a positive result in those who died within 2 months.
f p = 0.001 as compared with an area of 0.5 for a nonuseful test. A value ranging from 0.7 to 0.8 indicates reasonable discrimination, and a value more than 0.8 represents good discrimination (51).
g p < 0.001 as compared with an area of 0.5 for a nonuseful test. See footnote f.