Cold infarction areas of varying size in the presence of left ventricular dysfunction: the impact on left ventricular ejection fraction determination by gated SPECT compared to radionuclide ventriculography.

The agreement in left ventricular (LV) ejection fraction (EF) determination between the gated perfusion SPECT (GPS) and radionuclide ventriculography (RVG) methods has been reported to show a systematic reduction with increasing EF values. In some reports, this EF difference exceeds 5%, indicating a critical underestimation rate by GPS in terms of clinical reliability. Although the accuracy of endocardial edge detection by means of GPS proves to be reliable, even in patients with large perfusion defects, the accuracy of EF calculation seems to depend on the level of individual EF values. Our objective was to investigate the RVG-GPS agreement in EF determination specifically in patients with myocardial infarction (MI) having perfusion defects including extremely depleted areas and LV dysfunction. Our patients had a EF<50%, a population in which the best agreement values between the above two methods should be found according to the literature. The idea was to test the accuracy of QGS algorithm in the presence of cold perfusion defects and a low probability of EF level influence. Thirty-six patients (26 males; 61.8+/-9.1 y and 10 female; 64.1+/-10.7 y) with MI older than two weeks, having perfusion defects, including all regions of varying size with no visible tracer uptake in rest GPS and RVG-EF<50% were included in the study. Rest-GPS was performed by injecting iv 925 MBq 99mTc-sestamibi (8-time bins) using a dual-headed gamma camera and rest-RVG was performed within the following three days by injecting iv 740 MBq 99mTc-pertechnetate-pyrophosphate (24 time bins). Myocardial perfusion was visually analyzed on a 17 segment-model and summed rest perfusion scores (SRS) were determined. The cold defect number (CDN) was calculated by selecting the myocardial segments with 0%-9% of maximal tracer uptake (grade 4) to identify the extent of the depleted tissue in each patient. The patients with (3)4 adjacent myocardial segments with grade 4 perfusion were considered as having large cold defects (Group 1: GR1). Patients with two or three CDN constituted the Group 2 (GR2; medium cold defects) and those with only one CDN constituted Group 3 (GR3; small cold defects). Hereafter, the relative weight of cold perfusion defects (CD%) and infarcted segments (IS%) were calculated for all patients and subgroups. Of 36 patients studied, 14 patients (39%, GR1) had large cold defects. Eleven patients (30.5%, GR2) had moderate and 11 (30.5%, GR3) had small cold defects. The overall mean value of CDN was 3.03+/-1.96 (1-9). Mean EF values in RVG and GPS among Groups 1, 2 and 3 were 28.78%+/-6.32%, 38.46%+/-6.43%, 38.73%+/-8.55% and 27.0%+/-6.93%, 37.82%+/-8.80%, 33.27%+/-11.65%, respectively. The percentage of patients showing an EF difference pound 5% between RVG and GPS in Groups 1, 2 and 3 were 93%, 73% and 27%, respectively. The CD% and IS% were 19+/-12, 30+/-7, 14+/-3, 6 and 52+/-20, 61+/-18, 50+/-17, 42+/-23 in overall, GR1, GR2 and GR3, respectively. It is concluded that: the negligible underestimation of EF in GR1 by GPS compared to RVG confirms the stability of the geometric modeling algorithm of QGS in this particular patient sub-group. Although the agreement results in patients with smaller defects were probably influenced through the realization of RVG and GPS studies performed on different days, it seems to be rational to set the threshold of RVG-GPS agreement dependency on EF levels in patients with myocardial infarction lower than 50%.