Radionuclide measurement of left ventricular volume: comparison of geometric and counts-based methods.

Radionuclide measurements of left ventricular volume were determined in 20 patients by geometric and nongeometric, counts-based techniques using data from first-pass and equilibrium blood pool scintigraphy. Two geometric analytic approaches were used: directly measured long and short axes and the area-length method. Each approach was applied to the single-plane right anterior oblique images obtained by the first-pass technique and to biplane data, using the right anterior oblique first-pass and left anterior oblique blood pool data together. For the nongeometric determinations, background-corrected left ventricular counts were related to blood counts. This ratio was converted to volume by means of a linear regression relationship with angiographic volumes. All methods yielded high correlation coefficients (r greater than or equal to 0.93), but the standard errors of the estimates for the geometric techniques were high, and therefore the 95% confidence limits were wide. The use of biplane data improved the correlations, but area-length analysis of digitized data was no better than direct measurement of short axes from the analog images. The counts-based, nongeometric method provided the highest correlation and lowest standard error. These findings indicate that nongeometric left ventricular volume measurements using equilibrium blood pool scintigrams are the most accurate radionuclide technique. This approach also permits multiple determinations with a single dose of radiotracer.