Determination of right ventricular mass in humans and dogs with ultrafast cardiac computed tomography.

Abstract

There is currently no clinical approach to precisely measure right ventricular (RV) mass. We postulated that the radiological mode of ultrafast computed tomography (CT) of 3-mm-thick slices with 0.7-mm resolution would allow sufficient resolution to accurately estimate RV mass. Using this radiological mode, we serially imaged the entire right ventricle from apex to base, gated to end diastole, and applied Simpson's rule to calculate mass of the RV free wall. Thirteen mongrel dogs (weight, 6-30 kg) were studied. The free wall mass of the right ventricle was in the range of 12.0-47.5 g and averaged 35.4 +/- 3.7 g (mean +/- SEE). The correlation between RV mass estimated by ultrafast CT and actual RV mass was r equaling 0.85, SEE equaling 5.5 g, slope equaling 0.99, and gamma intercept equaling -1.8. Intraobserver and interobserver variability (r = 0.99 and r = 0.99, respectively) was excellent with a standard deviation (SD) equal to 1.5 and 1.8, respectively. The effect of variable RV preload (right atrial pressure, -5 to +20 mm Hg) on accuracy of RV mass measurements produced minimal error (SD = 3.6 g) in RV mass measurements. Seven normal young healthy men were also studied. The free wall mass of the right ventricle was in the range of 48.3-67.4 g and averaged 54.6 +/- 2.8 g (mean +/- SEE). The left ventricular to right ventricular (LV:RV) ratio averaged 3.2 +/- 0.2:1. These results are in agreement with human autopsy data in healthy males reporting mean RV mass equal to 46 g and an LV:RV ratio equal to 3.4:1. Because imaging every 3-mm slice from apex to base requires two contrast injections, we determined the accuracy of RV mass measurements if only every fourth 3-mm slice with interpolation was used. RV mass measurements using every slice or every fourth slice with interpolation were excellent (dogs, r = 0.99; humans, r = 0.97). It is concluded that high resolution CT imaging (3-mm tomograms) allows accurate measurements of RV mass. It is possible to add this stop-action mode of ultrafast CT, to previous CT studies, using every fourth tomographic slice for mass determinations and only one additional contrast injection of 40-60 ml. This should permit the study of progression and regression of RV mass in patients with various diseases.