TABLE 5.
Representative values of neural network diagnostic accuracy.
| ABR | ATR | TBR | ||||
|---|---|---|---|---|---|---|
| Mean | 95% CI | Mean | 95% CI | Mean | 95% CI | |
| Accuracy | 67.71 a , b | 64.10–71.31 | 58.85 a , c | 55.40–62.31 | 49.48 b , c | 48.25–50.71 |
| Sensitivity | 56.94 b | 53.66–60.23 | 59.72 c | 47.62–71.82 | 36.36 b , c | 32.30–40.43 |
| Specificity | 74.17 a , b | 67.89–80.44 | 59.17 a | 49.54–68.79 | 62.5 b | 60.23–64.77 |
| PPV | 69.31 a , b | 64.12–74.51 | 59.60 a , c | 56.81–62.38 | 49.09 b , c | 44.96–53.24 |
| NPV | 63.23 b | 61.11–65.35 | 60.29 c | 56.24–64.34 | 49.59 b , c | 47.24–51.94 |
| AUC | 0.717 b | MIN–MAX | 0.633 c | MIN–MAX | 0.5 b , c | MIN‐MAX |
| 0.656–.778 | .594–.671 | .5–.5 | ||||
Note: Accuracy, sensitivity, specificity, PPV, NPV, and AUC are presented at disease prevalence of 50%. AUCs of receiver operating characteristic curve were calculated by single‐point method.
Abbreviations: ABR, alpha/beta ratio; ATR, alpha/theta ratio; AUC, area under curve; CI, confidence interval; NPV, negative predictive value; PPV, positive predictive value; TBR, theta/beta ratio.
a
Significantly different in the comparison between ABR and ATR.
b
Significantly different in the comparison between ABR and TBR.
c
Significantly different in the comparison between ATR and TBR.