Table 2 Overview of quantitative model performance.
From: Automated detection and segmentation of non-small cell lung cancer computed tomography images
Data, # of patients | Detection performance | Segmentation performance | ||||||
|---|---|---|---|---|---|---|---|---|
Lung-wise | Specificity | Sensitivity | DSC (IQR) | Ji (IQR) | H95th, mm | Surf DSC | APL, cm | |
AUC (CI) | [τ = 1.18] | |||||||
(IQR) | ||||||||
Test, 93 | 0.96 | 0.97 | 0.96 | 0.85 | 0.74 | 5 | 0.75 | 106 |
(0.94–0.98) | (0.15) | (0.22) | (0.29) | (274) | ||||
External validation, 236 | 0.98 | 0.99 | 0.97 | 0.82 | 0.70 | 9.43 | 0.63 | 306 |
(0.97–0.99) | (0.17) | (0.24) | (0.28) | (984) | ||||