Limitations of quantitative phase analysis of radionuclide angiograms for detecting coronary artery disease in patients with impaired left ventricular function

Abstract

Phase analysis of radionuclide ventriculograms is used for identifying ischemic wall motion abnormalities. Myocardial segments with an abnormal phase, that is, delayed onset of wall motion, can be localized on a phase distribution image, and the synchronicity of left ventricular (LV) wall motion can be assessed from a histogram of LV phase distribution. The standard deviation of the LV peak on this histogram (SDP-LV) describes the width of the peak and is used as an index of the synchronicity of wall motion. We examined in this study the sensitivity of SDP-LV for identifying coronary artery disease (CAD) and its specificity in patients with normal and various degrees of LV impairment. A total of 84 patients were studied. Forty-five patients had CAD and 39 had congestive cardiomyopathy or valvular heart disease. Patients were grouped according to their LV ejection fraction (EF). In group I (37 patients) resting LVEF was equal to or greater than 50%, in group II (24 patients) it ranged from 35% to 50%, and in group III (23 patients) it was less than 25%. SDP-LV was highly specific in groups I and II for CAD with sensitivities of 48% and 89% at rest that increased with exercise to 88% and 100%. In group III patients, SDP-LV remained highly sensitive but was no longer specific for CAD. Therefore, in severe LV impairment, phase analysis does not aid in distinguishing CAD from other causes of ventricular dysfunction. By contrast, phase analysis is highly sensitive and specific for CAD in patients with normal or moderately depressed LV function.