Simple area-based measurement for multidetector computed tomography to predict left ventricular size

Abstract

Measures of left ventricular (LV) mass and dimensions are independent predictors of morbidity and mortality. We determined whether an axial area-based method by computed tomography (CT) provides an accurate estimate of LV mass and volume. A total of 45 subjects (49% female, 56.0 +/- 12 years) with a wide range of LV geometry underwent contrast-enhanced 64-slice CT. LV mass and volume were derived from 3D data. 2D images were analysed to determine LV area, the direct transverse cardiac diameter (dTCD) and the cardiothoracic ratio (CTR). Furthermore, feasibility was confirmed in 100 Framingham Offspring Cohort subjects. 2D measures of LV area, dTCD and CTR were 47.3 +/- 8 cm(2), 14.7 +/- 1.5 cm and 0.54 +/- 0.05, respectively. 3D-derived LV volume (end-diastolic) and mass were 148.9 +/- 45 cm(3) and 124.2 +/- 34 g, respectively. Excellent inter- and intra-observer agreement were shown for 2D LV area measurements (both intraclass correlation coefficients (ICC) = 0.99, p < 0.0001) and could be confirmed on non-contrast CT. The measured 2D LV area was highly correlated to LV volume, mass and size (r = 0.68; r = 0.73; r = 0.82; all p < 0.0001, respectively). On the other hand, CTR was not correlated to LV volume, mass, size or 2D LV area (all p > 0.27). Compared with traditionally used CTR, LV size can be accurately predicted based on a simple and highly reproducible axial LV area-based measurement.