Noninvasive quantification of left ventricular myocardial mass by gated proton nuclear magnetic resonance imaging.

The validity of cardiac nuclear magnetic resonance imaging for determination of left ventricular myocardial mass was evaluated in nine dogs. A gated spin echo-pulsing sequence was used for in vivo imaging, obtaining 0.7 cm thick slices of the heart spaced by 1 cm. On each imaged slice, the left ventricular surface area was reproducibly determined by planimetry and was multiplied by slice spacing and specific gravity of the myocardium (1.05) to obtain slice mass. Total left ventricular mass was calculated by adding slice masses in short-axis (method I), transaxial (method II) and vertical long-axis (method III) orientations using Simpson's rule. With each method, masses of the portions of the left ventricle subject to partial volume effect either were not accounted for or alternatively were estimated from the same or an orthogonal imaging plane. Calculated left ventricular mass was compared with the actual excised left ventricular weight. With NMR imaging of in situ nonbeating hearts, best results were obtained when either method I or method II was used and partial volume effect was estimated either from the same or an orthogonal plane. With in vivo NMR imaging, best results were noted when method I was used and mass of the partial volume apex was calculated from transaxial slices: Y (in vivo NMR image) = 8.3 + 0.99X, r = 0.996, SEE = 3.14. For this method, the interobserver reliability coefficient and standard error of the measurement were 0.97 and 5.4, respectively. Compared with method I, in vivo methods II and III were associated with larger errors (SEE ranging from 13.03 to 19.03) regardless of the approach used to estimate partial volume effect. It is concluded that NMR imaging is a highly accurate noninvasive method for in vivo measurement of left ventricular mass in dogs and offers promise for accurate measurement of left ventricular mass in patients.