We evaluated the accuracy of the non-imaging nuclear probe for measuring various parameters of left ventricular function by comparing these with simultaneous measurements acquired via tantalum marker cine-fluoroscopy. Eight patients with surgically implanted mid-myocardial tantalum markers were studied during cardiac catheterization. High temporal resolution tantalum marker-derived volume/time curves were generated and calibrated to absolute endocardial volumes by comparison with contrast left ventricular angiography. Left ventricular function parameters were acquired at baseline and during atrial pacing, nitroprusside infusion and volume loading. Simultaneous measurements obtained with the nuclear probe and via tantalum marker cine-fluoroscopy were compared. The following correlation coefficients were obtained: ejection fraction, r = 0.49, P less than 0.001; peak ejection rate, r = 0.41, P less than 0.01; mean ejection rate, r = 0.62, P less than 0.001; time to peak ejection, r = 0.73, P less than 0.001; peak filling rate, r = 0.73, P less than 0.001; time to peak filling, r = 0.58, P less than 0.001. Relative changes in end-diastolic volume were accurately tracked by the nuclear probe in six of seven patients (r values 0.77-0.93). These results suggest that measurements of certain left ventricular volume parameters with the nuclear probe may not accurately reflect true volume changes as measured by tantalum marker cine-fluoroscopy. Although, the probe may be useful in tracking relative changes in end-diastolic volume, we would advise caution in using this instrument for studies requiring accurate measurements of ejection and filling parameters.