Three-Dimensional Echocardiographic Reconstruction of the Left Ventricle by a Transesophageal Tomographic Technique: In Vitro and In Vivo Validation of its Volume Measurement.

Abstract

Accurate determination of left ventricular (LV) volume has important therapeutic and prognostic implications in patients with cardiac disease. Volume estimations by two-dimensional techniques are not very accurate due to geometric assumptions. OBJECTIVES: To validate LV volume determinations by a new transesophageal three-dimensional echocardiographic technique. We performed three-dimensional reconstruction of the LV using an echo-computed tomographic (CT) technique based on serial pullback parallel slice imaging technique in both in vitro and in vivo settings. Fourteen latex balloons with various sizes (30-235 mL) and shapes (conical, pear shaped, round, elliptical, and aneurysms in various locations) filled with known volumes of water were imaged in a water bath. From the static three-dimensional image, the LV long axis was defined and the LV was sectioned perpendicular to this axis into 2-mm slices. The volume of each slice was calculated with the observer blinded to the actual volume as the product of the slice thickness and the manually traced perimeter of the slice and the LV volume as the sum of the volumes of the slices (Simpson's method). The calculated LV volume closely correlated with the actual volume (r = 0.99, P < 0.0001, calculated volume = 1.06x - 11.3, Deltavolume = -5.7 +/- 10.0 cc). Using the same system, transesophageal echocardiographic (TEE) images of the LV were obtained in 15 patients gated to respiration and ECG. Satisfactory dynamic three-dimensional reconstruction of the LV was possible in ten patients. The three-dimensional LV volumes (systolic and diastolic) using Simpson's method correlated well with those obtained from biplane or multiplane TEE images using the area length method (r = 0.89, p < 0.0001, y = 12.7 + 0.84x, Deltavolume = 1.3 +/- 18.1 cc). The LV major-axis diameters by the two methods showed very close correlations as well (r = 0.86, P < 0.0001, y = 19 + 0.74x, Deltadiameter = 1.0 +/- 7.2 mm). We conclude that three-dimensional LV volume calculation by the echo-CT technique is intrinsically sound, is independent of LV geometry, and with some limitations, is applicable in vivo. (ECHOCARDIOGRAPHY, Volume 13, November 1996)