Table 3 Characteristics of the training, validation, and external validation datasets for binary and subtype classification
From: An interpretable deep learning model for first-trimester fetal cardiac screening
Binary Classification (n=8062) | Subtype Classification (n=4240) | |||||||
|---|---|---|---|---|---|---|---|---|
Training set (n=5254) | Validation set (n=1534) | External validation set 1 (n=833) | External validation set 2 (n=136) | External validation set 3 (n=305) | Training set (n=2372) | Validation set (n=614) | External validation set (n=1254) | |
Patients | 4253 | 1063 | 686 | 129 | 152 | 1941 | 477 | 266 |
Mean age | 31 (4; 17–48) | 30 (3; 19–45) | 30 (4; 18–43) | 31 (4; 20–45) | 32 (3; 24–45) | 30 (4; 17–48) | 30 (4; 18–44) | 31 (3; 20–45) |
Diagnosis | ||||||||
Normal | 3935 | 984 | 671 | 98 | 142 | 1629 | 407 | 206 |
Abnormal | 1319 | 550 | 162 | 38 | 163 | |||
AVSD | 142 | 31 | 86 | |||||
HV | 208 | 27 | 257 | |||||
FSV | 393 | 149 | 705 | |||||
Chamber Number | ||||||||
1 | 35 | 10 | 0 | 0 | 0 | 36 | 8 | 0 |
2 | 1047 | 212 | 152 | 21 | 118 | 510 | 160 | 589 |
3 | 403 | 220 | 17 | 15 | 43 | 178 | 34 | 262 |
4 | 3769 | 1092 | 664 | 100 | 144 | 1648 | 412 | 403 |
Flow Number | ||||||||
Single | 1234 | 373 | 161 | 30 | 131 | 635 | 190 | 867 |
Two | 4020 | 1161 | 672 | 106 | 174 | 1737 | 424 | 387 |
Flow Symmetry | ||||||||
Equal | 3935 | 984 | 671 | 98 | 142 | 1629 | 407 | 206 |
Unequal | 1319 | 550 | 162 | 38 | 163 | 743 | 207 | 1048 |
