Skip to main content

Table 3 Performance analysis for the Lung-RADS and ANN models (n = 234)

From: Artificial neural networks improve LDCT lung cancer screening: a comparative validation study

Scale/model Lung-RADS ANN
Cut-off Category 3 > 0.012
AUC (95% CI) 0.764 (0.705, 0.817) 0.873 (0.823, 0.913)
Classification accuracy (%) 93.16 84.62
Sensitivity (95% CI) 12.50 (0.3, 52.7) 75.00 (34.9, 96.8)
Specificity (95% CI) 96.02 (92.6, 98.2) 84.96 (79.6, 89.4)
PPV (95% CI) 10.0 (1.6, 43.7) 15.0 (9.6, 22.6)
NPV (95% CI) 96.9 (96.0, 97.6) 99.0 (96.7, 99.7)
LR+ (95% CI) 3.14 (0.5, 21.9) 4.99 (3.0, 8.3)
LR- (95% CI) 0.91 (0.7, 1.2) 0.29 (0.1, 1.0)
  1. AUC area under the curve; CI confidence interval; LR+ positive likelihood ratio; LR− negative likelihood ratio; NPV negative predictive value; PPV positive predictive value