In this study, the clinical utility of phase analysis using gated blood pool scintigraphy to detect aortic regurgitation (AR) was assessed. The study population included 33 patients with AR documented by cineaortography, and 53 patients without AR diagnosed by clinical features and pulsed Doppler echocardiography or cineaortography. Gated blood pool images were stored in modified left anterior oblique views by the multiple gated method (28 frames per beat) after the in vivo labeling of erythrocytes using 25 mCi 99m-Tc. Amplitude was depressed at the subaortic region toward the apex in patients with AR, and occasionally the phase angle occurred earlier in the same region. This is because diastolic filling began earlier and stroke count was decreased more in the subaortic region than in other regions of the left ventricle in examining the regional time-activity curve. Based on this characteristic finding in the amplitude image, the specificity was 81% and the sensitivity was 79% in the diagnosis of AR. The degree of AR was estimated using the amplitude image as follows: grade 1, depression of amplitude localized to the subaortic region; grade 2, depression of amplitude detected from the subaortic region to the midportion of the left ventricle; grade 3, depression of amplitude detected beyond the midportion of the left ventricle. We compared results with those of the radionuclide (RI's classification) and angiographic methods (Sellers' classification) as to the severity of AR. AR could not be detected using the radionuclide method in all four patients with Sellers' grade 1, and in three of nine patients with Sellers' grade 2.(ABSTRACT TRUNCATED AT 250 WORDS)