Skip to main content
. 2020 Jul 7;10:1070. doi: 10.3389/fonc.2020.01070

Table 3.

The diagnostic performances of the deep learning-based CAD system and experienced radiologists for asymptomatic lesions and symptomatic lesions.

SE(%) SP(%) PLR NLR PPV(%) NPV(%) Accuracy (%) AUC
(95%CI) (95%CI) (95%CI) (95%CI) (95%CI) (95%CI) (95%CI) (95%CI)
Asymptomatic lesions CAD 93.75 83.87 5.81 0.07 75.00 96.30 87.23 0.89
(86.01–97.94) (77.12–89.28) (4.04–8.36) (0.03–0.17) (65.34–83.12) (91.57–98.79) (82.28–91.22) (0.84–0.93)
radiologists 95.00 66.45 2.83 0.08 59.38 96.26 76.17 0.81
(87.69–98.62) (58.43–73.83) (2.26–3.55) (0.03–0.20) (50.34–67.96) (90.70–98.97) (70.20–81.47) (0.75–0.86)
Symptomatic lesions CAD 80.00 61.84 2.10 0.32 79.43 62.67 73.61 0.71
(72.41–86.28) (49.98–72.75) (1.56–2.82) (0.22–0.47) (71.82–85.77) (50.73–73.57) (67.20–79.36) (0.64–0.77)
radiologists 97.14 60.53 2.46 0.05 81.93 92.00 84.26 0.79
(92.85–99.22) (48.65–71.56) (1.86–3.26) (0.02–0.13) (75.22–87.46) (80.77–97.78) (78.70–88.85) (0.73–0.84)

SE, sensitivity; SP, specificity; PLR, positive likelihood ratio; NLR, negative likelihood ratio; PPV, positive predictive value; NPV, negative predictive value; AUC, area under the receiver operator characteristics curve; 95% CI, 95% confidence interval.