Table 2.
Echocardiographic measurements by AI measurement support software and reference in the different data sets
| Data set 1 | Data set 2 | Data set 3 | Data set 4 | |
|---|---|---|---|---|
| (real time) | (HUNT4 repa) | (HUNT4) | (Tromsø7) | |
| EDV AI, mL | 120.5 (38.6) | 106.7 (26.6) | 100.6 (26.7) | 96.9 (24.5) |
| ESV AI, mL | 61.2 (31.6) | 49.2 (17.0) | 46.9 (15.4) | 43.5 (14.0) |
| EF AI, % | 50.9 (9.7) | 54.2 (7.3) | 53.6 (6.8) | 55.4 (6.4) |
| EDV reference, mL | 132.1 (41.1) | 115.8 (30.5) | 108.9 (32.7) | 90.4 (23.8) |
| ESV reference, mL | 65.2 (36.3) | 50.4 (18.9) | 44.7 (17.2) | 40.8 (13.5) |
| EF reference, % | 52.8 (10.7) | 57.1 (6.2) | 59.1 (6.6) | 55.1 (7.1) |
| EDV differenceb, mL | −11.6 (17.5) | −9.1 (8.4) | −8.3 (14.4) | 6.5 (14.0) |
| ESV differenceb, mL | −4.0 (14.6) | −1.2 (6.0) | 2.2 (8.8) | 2.7 (11.2) |
| EF differenceb, % | −1.9 (7.6) | −2.9 (4.0) | −5.5 (6.5) | 0.3 (8.6) |
Mean (SD) for absolute values of measurements grouped by data set.
AI, artificial intelligence; EDV, end-diastolic volume; EF, ejection fraction; ESV, end-systolic volume; HUNT, Trøndelag Health Study; rep, test–retest reproducibility.
aValues are averages of all available measurements by each method for all subjects.
bPaired differences calculated as AI minus reference.