Table 3 Quantitative measurements between automated segmentation and manual annotation
Quantitative metrics | All LNs | Suspicious LNs | Largest LNs | ||||||
|---|---|---|---|---|---|---|---|---|---|
Automated segmentation | Manual annotation | P | Automated segmentation | Manual annotation | P value | Automated segmentation | Manual annotation | P value | |
Internal test dataset | |||||||||
Volume (mm3) | 172.3 [71.1, 390.3] | 197.9 [83.0, 449.2] | 0.002 | 623.0 [374.3, 1123.1] | 785.1 [517.9, 1389.3] | 0.000 | 10.3 [8.7, 13.8] | 10.8 [9.4, 15.1] | 0.000 |
Short diameter (mm) | 5.4 [3.9,7.3] | 5.7 [4.0,7.7] | 0.000 | 9.1 [7.7, 11.2] | 9.7 [8.7, 11.9] | 0.000 | 752.3 [466.4, 1548.5] | 886.2 [530.8, 1787.2] | 0.000 |
External validation dataset | |||||||||
Volume (mm3) | 103.3 [45.9, 244.90] | 130.3 [61.2, 304.2] | 0.000 | 428.6 [244.9, 688.8] | 551.0 [382.7, 870.9] | 0.000 | 212.2 [95.1, 382.6] | 266.0 [134.7, 463.2] | 0.000 |
Short diameter (mm) | 4.8 [3.4, 6.9] | 5.1 [3.6, 7.5] | 0.000 | 8.7 [7.4, 10.1] | 9.4 [8.6, 10.7] | 0.000 | 6.3 [4.7, 8.1] | 6.8 [5.0, 8.6] | 0.000 |