Myocardial imaging by digital subtraction angiography for left ventricular mass measurement.

Digital ECG- and respiration gated background subtraction and image combination were used to visualize the myocardial wall in routine left ventricular angiocardiography. Muscle mass was computed from manually traced biplane endo- and epicardial contours according to the multiple slices algorithm. Experimental validation and first clinical applications are reported. In animal experiments (13 pigs, 16 to 25 kg) a close correlation to postmortem measurement was found: r = 0.894 and SEE = 7.4 g for end-diastole, r = 0.938 and SEE = 6.2 g for end-systole. The postmortem measurement was always within the 95%-confidence range of +/- 10%. In 16 patients (five days to 60 years, 3.2 to 81 kg) assessment of myocardial mass by myocardial imaging and conventional angiocardiography was compared. Conventional calculation based on manually outlined cavity contour and mean thickness of a lateral wall segment in the unprocessed angiogram. At end-diastole myocardial imaging and the conventional method correlated closely with r = 0.986 and SEE = 18.7 g. Conventional mass determination was inaccurate in end-systole (r = 0.928, SEE = 56.9 g) with values exceeding the diastolic ones by 26% +/- 19% (mean, SD), sometimes by up to 100%. Enclosure of the papillary muscle in the averaged wall segment probably leads to a mostly overrated mean wall thickness. There was no significant difference between diastolic and systolic results of myocardial imaging. In further investigations myocardial imaging showed good agreement to two-dimensional echocardiographic mass determination and, furthermore, has successfully been applied to study contraction pattern and to evaluate percentage of non-perfused muscle volume.