Background: Two-dimensional (2D) echocardiographic automated border detection (ABD) can provide on-line measurement of left ventricular cavity area and fractional area change. However, this new quantitative method has not been extensively validated.
Methods: Values of manually traced areas on 2D-echo images were compared with those obtained from ABD in 34 consecutive normal subjects (age 16-65 years). Only subjects with more than 70% of endocardial border circumferences clearly seen in both selected imaging planes were included in the study. We evaluated left ventricular end-diastolic and end-systolic area, and fractional area change obtained from mid-left ventricular short axis and apical 4-chamber view. Left ventricular volumes (area/length method) and ejection fraction were manually calculated off-line from apical 4-chamber view.
Results: From the short axis view, left ventricular cavity area measurements with ABD were obtained in 85% of subjects. The values closely correlated with off-line measurements: end-diastolic area 15.6 +/- 3.1 vs 14.8 +/- 3.3 cm2, r = 0.88 SEE = 1.58; end-systolic area 7.2 +/- 1.7 vs 6.7 +/- 1.7 cm2, r = 0.88 SEE = 0.80. A good correlation was also found for the apical 4-chamber view; end-diastolic area 25.9 +/- 5.9 vs 25.3 +/- 5.5 cm2, r = 0.97 SEE = 1.36; end-systolic area 16.3 +/- 4.1 vs 15.0 +/- 3.8 cm2, r = 0.92 SEE = 1.51. In this view ABD measurements were obtained in 79% of subjects. A significant correlation was also found between the end-diastolic volume and short axis (r = 0.54, SEE = 2.63; p 0.003) and apical 4-chamber (r = 0.66, SEE = 4.51; p = 0.0002) ABD diastolic area. Similarly, the end-systolic volume was significantly correlated with short axis (r = 0.57, SEE = 1.42; p = 0.001) and apical 4-chamber (r = 0.55, SEE = 3.54; p = 0.003) ABD systolic area. However, the on-line fractional area change correlated with off-line ejection fraction better from short axis view: (r = 0.72 SEE = 3.52) than from apical 4-chamber view (r = 0.45 SEE = 6.84).
Conclusions: These data indicate that: 1) left ventricular areas measured by ABD correlate well with manually measured areas and volumes; 2) short axis ABD fractional area change may be a reliable substitute of off-line manually traced ejection fraction in normal subjects.