Table 3.
Comparison of performance in stroke volume estimation between the deep learning model and FloTrac algorithm.
| Measure | Deep learning model, mean (SD) | FloTrac, mean (SD) | Statistical test | ||
|
|
Count, n | P value | |||
| Error (mL) | |||||
|
|
Overall (n=56) | –7.9 (20.7) | –8.4 (22.1) | 158725 | <.001 |
|
|
Cardiac surgery (n=16) | –1.7 (15.2) | 2.7 (18.3) | 65260 | <.001 |
| Liver transplantation (n=40) | –9.0 (21.3) | –10.3 (22.2) | 93465 | <.001 | |
| Absolute error (mL) | |||||
|
|
Overall (n=56) | 16.5 (14.8) | 18.3 (15.1) | 158725 | <.001 |
|
|
Cardiac surgery (n=16) | 11.1 (10.5) | 14.3 (11.8) | 65260 | <.001 |
|
|
Liver transplantation (n=40) | 17.4 (15.3) | 19.0 (15.4) | 93465 | <.001 |
| Percentage error (%) | |||||
|
|
Overall (n=56) | –4.4 (26.9) | –5.6 (28.6) | 158725 | <.001 |
|
|
Cardiac surgery (n=16) | 1.8 (29.9) | 9.5 (34.9) | 65260 | <.001 |
|
|
Liver transplantation (n=40) | –5.5 (26.2) | –8.3 (26.5) | 93465 | <.001 |
| Absolute percentage error (%) | |||||
|
|
Overall (n=56) | 20.3 (18.3) | 22.5 (18.5) | 158725 | <.001 |
|
|
Cardiac surgery (n=16) | 19.3 (22.9) | 25.4 (25.7) | 65260 | <.001 |
|
|
Liver transplantation (n=40) | 20.4 (17.4) | 22.0 (16.9) | 93465 | <.001 |