Electrocardiogram gated single photon emission computed tomography of the intracardiac blood pools is a recent development that involves the acquisition of images in multiple projections after in vivo erythrocyte labelling with technetium-99m and reconstruction of these images into tomographic sections in any desired plane. The technique was used in 25 subjects to measure left ventricular volume, by summing the areas of the ventricle in each of the tomographic sections, and the results compared with those using a counts based (non-geometric) technique from planar radionuclide ventriculography. Endocardium was defined with the aid of a contour at 43% of maximum left ventricular counts, and this contour was validated for a left ventricular phantom. Correlation between tomographic and counts based left ventricular volume was close. Similarly, ejection fraction correlated well. The technique is therefore an accurate method for determining left ventricular volume and ejection fraction, avoiding the assumptions about shape made by other geometric methods.

No abstract

SUMMARY The phase image generated by radionuclide angiocardiography illustrates the regional tming of ventricular wall motion. In this study the phase image was used to investigate the patterns of ventricular contraction in 103 subjects with either normal hearts or a conduction abnormality. In 38 normal subjects the right ventricle contracted on average 7 ms after the left, and the last region to contract was the right ventricular outflow tract. In 15 subjects with left bundle branch block the left ventricle contracted 69 ms after the right, contraction spreading from the septum to the lateral wall. In 12 subjects with right bundle branch block right ventricular contraction occurred 54 ms after the left. In 11 subjects with right bundle branch block and left anterior fascicular block both right and left ventricular contraction were delayed, the right more so than the left. In three of five subjects with the Wolff-Parkinson-White syndrome and four with frequent ventricular extrasystoles areas of early contraction corresponded to areas of early depolarisation.It is concluded that ventricular contraction can be studied non-invasively and follows a pattern to be expected from the pattern of electrical depolarisation.Radionuclide angiocardiography is now firmly established in the non-invasive assessment of ventricular function. Traditionally, measurement of wall motion has been used as an index of regional contraction, but the recently introduced Fourier phase and amplitude images'-4 circumvent the problem of edge detection and provide a more complete description of wall motion. Fourier analysis quantifies the timing and magnitude of the cyclical changes in radioactive counts for each display element of the ventriculogram, and the phase and amplitude images so generated indicate the timing and magnitude of regional ventricular wall motion.The purpose of this study was to use the phase image to investigate the timing and synchronisation of left and right ventricular contraction in a variety of conduction disorders including left and right bundle branch blocks, left anterior fascicular block, the Wolff-Parkinson-White syndrome, endocardial pacing, and ventricular extrasystoles.Requests for reprints to Dr S R Underwood, Department of Cardiology, The Middlesex Hospital, Mortimer Street, London WIN 8AA.Accepted for publication 13 December 1983 Patients and methods One hundred and three subjects were studied at rest by electrocardiogram gated equilibrium radionuclide angiocardiography. Sixty-five had electrocardiographic evidence of a conduction abnormality (Table), and the phase image was used to compare the tiining of contraction in these subjects with that in the 38 whom electrical conduction was normal. All subjects had a normal ejection fraction and amplitude image for the ventricle with abnormal conduction, and the extent of ventricular contraction was therefore normal.

SUMMARY Attenuation of counts from the more distant inferior portion of the left ventricular blood pool in equilibrium radionuclide ventriculography may mean that inferior infarction is less likely to be detected than anterior infarction. Fourier amplitude and phase images can be used to map the extent and timing of regional ventricular wall motion and this study assesses their use for the detection of anterior and inferior infarction. Normal regional values of amplitude and phase were established in 38 individuals without evidence of cardiac abnormality. In 20 patients with anterior infarcts, though the sensitivity of the combined left anterior oblique amplitude and phase images was high (95%) it was lower (77%) in 39 with inferior infarcts, principally because the sensitivity of the phase image for the detection of inferior infarcts was only 39%. Right anterior oblique images generated from a first pass study detected all 13 patients with inferior infarcts. The mean left ventricular ejection fraction was significantly lower in the patients with anterior infarcts (37%) than in those with inferior infarcts (48%). Although the mean wall motion score on x ray contrast ventriculography was slightly lower in the patients with anterior infarction, the high sensitivity of the right anterior oblique amplitude and phase images in inferior infarction suggests that attenuation of counts is an important cause of reduced sensitivity of the left anterior oblique images. This may also partly explain the lower ejection fractions in inferior infarction.It follows that both a right anterior oblique first pass study and a left anterior oblique equilibrium study are necessary for an accurate description of regional wall motion. images show motion of all parts of the left ventricle and they differ from methods that display contours; these only show motion of walls that are perpendicular to the plane of projection. Together, the images detect abnormal wall motion because abnormalities of both the extent and timing of contraction occur after myocardial infarction.45In the usual left anterior oblique projection, each left ventricular pixel represents a narrow column of blood passing from the anterior to the inferior wall. Changes in counts are the result of motion of both of the walls, and abnormal motion of either will lead to abnormalities of amplitude and phase. This may make it unclear which part of the ventricle is giving rise to an abnormality, but in practice the ambiguity can be resolved by the recognition of common pattems of wall motion. For instance, an amplitude defect in the centre of the left ventricle could represent either anterior or inferior hypokinesis, but 411

Stress induced changes in left ventricular ejection fraction are widely used in the detection and assessment of coronary artery disease. This study demonstrates that right ventricular dysfunction may also occur, and assesses its significance in terms of coronary artery anatomy. This study involved 14 normal subjects and 26 with coronary artery disease investigated by equilibrium radionuclide ventriculography, at rest and during maximal dynamic exercise. Mean normal resting right ventricular ejection fraction (RVEF) was 0.40 (SD 0.118), and all normal subjects increased RVEF with stress (mean delta RVEF + 0.13 SD 0.099). Mean delta RVEF in the subjects with coronary artery disease was significantly lower at 0.00 (SD 0.080), but there was overlap between the two groups. The largest falls in RVEF were seen if the right coronary artery was occluded without retrograde filling. In this subgroup with the most severely compromised right ventricular perfusion (nine subjects), RVEF always fell with stress, and mean delta RVEF was -0.08 (SD 0.050). There was no significant correlation between delta LVEF and delta RVEF, implying that the right ventricular dysfunction was due to right ventricular ischaemia, rather than secondary to left ventricular dysfunction. Stress induced right ventricular ischaemia can therefore be detected readily by radionuclide ventriculography.