Table 2.

Summary of the performance of the simplest model (2-features) and the complex model (6 features) derived in this paper in the definitive testing dataset. We used sensitivity, specificity, positive predictive value, and negative predictive value. Values in parenthesis indicate the 95% confidence interval. Corresponding threshold values are also indicated. It is worth mentioning that the simplest model can be used, for all the trade-offs between sensitivity and specificity proposed here, by simply inspecting Fig. 4.

Threshold type	Threshold value	Sensitivity (95% C.I.)	Specificity (95% C.I.)	Positive Predictive Value (95% C.I.)	Negative Predictive Value (95% C.I.)
Simple model (2 features)
Optimal	0.50	0.87 (0.85–0.88)	0.76 (0.75–0.77)	0.78 (0.76–0.80)	0.85 (0.84–0.86)
High-Sensitivity	0.48	0.94 (0.92–0.95)	0.65 (0.64–0.66)	0.73 (0.71–0.75)	0.91 (0.90–0.92)
High-Specificity	0.52	0.67 (0.64–0.69)	0.91 (0.91–0.92)	0.89 (0.86–0.91)	0.73 (0.72–0.74)
Complex model (6 features)
Optimal	0.47	0.89 (0.87–0.90)	0.82 (0.81–0.83)	0.83 (0.81–0.85)	0.88 (0.87–0.89)
High-Sensitivity	0.40	0.92 (0.90–0.93)	0.79 (0.78–0.80)	0.82 (0.80–0.84)	0.90 (0.89–0.91)
High-Specificity	0.67	0.74 (0.72–0.77)	0.90 (0.90–0.91)	0.89 (0.86–0.91)	0.78 (0.76–0.79)