and imaging To determine the ability of nuclear magnetic resonance (NW) imaging to assesa left ventricular (LV) volumes and ejection fraction (EF), we studted 24 patknb wfthk 48 hours of atngie-piane LV angkgraphy. in ail patients, a tranaverae, ak$ie-ptane NWt acqwiattton techntque we8 employed with LV end-diaatoik (ED) and end-syatoiic (ES) volumes (V) cekukted by a modtftad area-iength algorithm. in nine patients, a muitiaiice ecqwkttkn technique waa empieyed with LVEDV and LVESV calculated by a Simpson’s rule algorithm. NWt-detarmked LVV and EF correlated reaaonabiy well with angiographk values (LVEDV: r = 0.75; LVESVz r = 0.90, and LVEF: I = 0.76). The single-plane NMR technique aignificantiy underestimated LVEDV @ < 0.01) whereas no significant difference was demonstrated for LVESV. AS a result, angkgraphic LVDF was aigniflcantiy underestimated (p < 0.05). This wndereattmetkn is itttaty related to off-aria imaging and to the geometrk constraints of a single-plane aigorithrn. in comparing muitiafke NW? to anglographic data, no signifkant dtfference wsa demonstrated for LyfDV, LVESV, or LVEF. Thus, qusntttation of LVV end EF with NMR is feasible, and comparison to angkgraphic votwmes is aimiiar to results reported from other noninvaaive imagtng modattties. improvement in current acquisition techniques and software shot&i reawit tn further qwantttattve potenttai. (AM HURT J 1987;113:24.) Lee R. Dilworth, M.D., Alex M. Aisen, M.D., G. B. John Mancini, M.D., F.R.C.P.(C), Ian Lande, M.D., and Andrew J. Buda, M.D. Ann Arbor, Mich. The assessment of left ventricular (LV) function is fundamental to the diagnostic evaluation, prognosti- cation, and therapy of patients with cardiac disease. Traditionally, LV volume and ejection fraction have been important parameters in this assessment. Ever since Sandler, Dodge, et al.‘, 2 validated angiographic methods of measuring LV volumes and ejection fraction, a number of noninvasive techniques have been used to estimate these parameters. Both echo- cardiographic and gated blood pool estimates of LV volume have yielded variable results when compared to angiographic measurements. Nuclear magnetic resonance (NMR) imaging is a new technique that has been used primarily for tissue characterization and qualitative structural analyses3 However, gated From the Cardiology Division, Department of Internal Medicine and Department of Radiology, University of Michigan Medical School. This work was supported in part by a Grant-in-Aid from the American Heart Association of Michigan; by Grant HL29716 from the National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Md.; and by The Veterans Administration, Washington, DC. Received for publication Jan. 30,1986; accepted May 26, 1986. Reprint requests: Andrew J. Buda, M.D., University of Michigan Medical Center, Cardiology Division, 1500 E. Medical Center Dr., TC-3910, Ann Arbor, MI 48019-0366. cardiac NMR images provide excellent endocardial definition and thus should be potentially useful for quantitating LV volumes and ejection fraction. To determine the accuracy of NMR determination of LV volumes and ejection fraction, we used two methods for determining these parameters by NMR and compared them to results from standard meth- ods that used LV angiography. METHODS Patient population. Twenty-seven patients (17 men and 10 women, mean age 59 years, range 36 to 70 years) in normal sinus rhythm were studied prospectively. After informed consent had been obtained, all patients under- went NMR and routine cardiac catheterization, including conventional or digital angiography, within 48 hours. Three patients, two with technically unsatisfactory NMR studies (due to an irregular cardiac rhythm resulting in poor imagequality) and onewith a technically inadequate ventriculogram (due to post injection ventricular tachy- cardia), were excluded from further analysis. Among the remaining 24 patients, six had normal cardiac evaluations, 14 had coronary artery disease(five with a history of myocardial infarction), two had variant angina, one had dilated cardiomyopathy, and onehad idiopathic hypertro- phic subaortic stenosis. NMR acqutattion techniques and quantttatlve analysis. 24