Quantitative measurements of myocardial perfusion and function from SPECT (and PET) studies depend on the method used to perform those measurements

Abstract

See related article, pp. 911-924One would perhaps be justified to remark that the title of this editorial only states the obvious and that its premise ought to be accepted by any reasonable reader, even if not particularly familiar with nuclear cardiology. Nevertheless, the paper to which the editorial refers, 1 describing good concordance but also substantial differences amongst three commercially available software packages to quantify LV perfusion and function from gated myocardial SPECT, offers us the opportunity to reflect on some of the specific reasons why those differences exist. As the authors describe in their comprehensive review of the published literature, a wealth of studies exists assessing the cross-correlation of the outputs of different quantitative gated SPECT algorithms, and the current study confirms previously reported findings, albeit in a larger population of 634 consecutive patients. Let us try to interpret and clarify those findings, while framing them in a more general context. DIFFERENT ALGORITHMS ARE BASED ON DIFFERENT COMPUTATIONAL TOOLS (AND WORKING ASSUMPTIONS)The specific software packages used by Alexiou et al are Emory University's Cardiac Toolbox (ECTb), 2Cedars-Sinai's Cardiac Suite (QGS/QPS), 3 and General Electric's Myovation. This writer is not aware of any published papers detailing the technical principles on which the Myovation software is based, but the other two algorithms and their differing characteristics have been well described in the literature. As far as measurement of global function is concerned, the ECTb performs a Fourier decomposition of the time-volume (T-V) curve that describes the LV cavity volume as a function of the cardiac cycle, then takes the first Fourier harmonic to estimate end-systolic (ES) and end-diastolic (ED) volumes-as a result, T-V curve estimates are less dependent on temporal sampling, and ejection fractions (EFs), EDVs and ESVs from 8-frame and 16-frame gated acquisitions will be relatively close in value. 2 In contrast, QGS/QPS measures the actual dynamic range of the T-V curve, which is directly proportional to the frequency of temporal sampling-in other words, 8-frame ESVs will be larger, EDVs smaller, and EFs lower compared to 16-frame gated acquisitions, with a substantially uniform difference of 3-4 EF points over the entire EF range. 4 With regard to myocardial thickening, the ECTb algorithm assumes a uniform myocardial thickness of 1 cm at ED, and increases it with a 1:1 relationship to the apparent myocardial ''brightening'' from ED to ES, since partial volume effect is fairly linear for the reconstructed spatial resolutions and myocardial thicknesses of interest. 2 QGS/QPS is also based, in part, on myocardial brightening, but uses phantom calibration to estimate ED thickness, and asymmetrical Gaussian fitting with a constant myocardial mass constraint to follow the endocardium and epicardium's location throughout the cardiac cycle. 3 As for perfusion quantification, aside from the different myocardial sampling techniques (hy...

New hardware, new software, different results: The estimation of left ventricular volumes and ejection fractions with gated myocardial perfusion SPECT

“…6 It is not surprising that there would be some changes in ventricular volumes and LVEF when acquisition and image reconstruction methods are varied. In a recent instructive editorial, 7 Germano reinforces the notion that repeatability is probably more important than absolute accuracy in a clinical setting.…”

Automated Analysis vs. Expert Reading in Nuclear Cardiology: Correlations with the Angiographic Score

“…MYO does not provide standardised segmental perfusion scores. Therefore, we converted the average segmental count values (relative to maximum pixel values in the relevant polar plot) to categorical scores according to >70%, 50–69%, 30–49%, 10–29% and <10% thresholds, as previously described [ 14 ]. Finally, automated myocardial perfusion measurements were based on the results of the commercially available packages, without any institutional adjustments.…”

Long-Term Prognostic Value of Automated Measurements in Nuclear Cardiology: Comparisons with Expert Scoring

“…The automated measurements of SSS, SRS, and SDS were recorded for the participants, as derived by ECTb and QPS software packages (using the institutional normal database for each programme). Notably, MYO software package does not provide standardised segmental perfusion scores, and we converted the average segmental count values (relative to maximum pixel values in the relevant polar plot) to categorical scores according to >70%, 50-69%, 30-49%, 10-29%, and <10% thresholds, as previously described [14].…”