Application of list-mode based retrospective gating in patients with and without arrhythmia for myocardial perfusion SPECT

Introduction: Although prospective electrocardiographic (ECG) gating with a predetermined beat length window has been commonly used for myocardial perfusion imaging, unpredicted irregular beat length due to multiple premature beats or atrial fibrillation, may deteriorate tomographic consistency due to suboptimal curation of beats. We recently proposed a method based on retrospective analysis of list mode data allowing to mitigate such deteriorations. In this preliminary study we compare SPECT images from the retrospective gating method to the conventional prospective gating method.

Methods: We studied a total of 44 patients with coronary heart disease who underwent myocardial perfusion scan with a Symbia Intevo 16 SPECT/CT using 99mTc-perfusion tracers at resting condition. The prototype xSPECT cardiac method, consisting of list mode acquisition in continuous rotation, resulting in 120 views over 360 degrees with (2.4 mm)²-pixels and 16 ECG gates and subsequent iterative reconstruction with a prototype xSPECT ordered subset expectation maximization (OSEM) method with CT-based attenuation and emission data driven scatter, motion and respiratory corrections. The conventional prospective ECG gated SPECT acquisition was performed without CT in 6-degree step-and-shoot resulting in 60 views over 360 degrees, (6.6 mm)²-pixels, 16 ECG gates, with a fixed width beat window enabled auto-tracking of RR intervals, and our clinical standard Flash3D iterative reconstruction. All beat lengths during SPECT data acquisition were analyzed with RR interval histograms, and fractions of beats in the most commonly clinical used 20%-window beat length were measured. Left ventricular (LV) end-diastolic volume (EDV), end-systolic volume (ESV), and ejection fraction (EF) were compared between prospective and retrospective gating methods. Echography measurement of volume (Simpson method) and EF was used as reference. List-mode based simulated prospective gating with 20% window was also examined in patients with frequent arrhythmia and outlier beat length.

Results: The conventional gating resulted in the fraction of accepted beats of 89% ± 20% (accepted: 23% - 100%), while ≥19% of the beats were rejected in 6 patients (accepted: 23% - 81%,). In contrast, with retrospective gating resulted in 98% ± 2% of usable beats for reconstruction. The EF and EDV derived from the prospective and retrospective gated data were highly correlated (R² = 0.92, EF_retro = 0.65+ 0.91*EF_pro, p <0.0001; R2 = 0.98, EDV_retro = 12 + 0.94*EDV_pro, p <0.0001). Compared with the echography EF, coefficients of determination (R²) were 0.74 and 0.80 for prospective and retrospective gating, respectively. In 6 patients with arrhythmia, conventional clinical study with prospective gating had to use wider acceptance window yielding 92% ± 9% used beats. Mean EF and EDV were 58% ± 17%, 53% ± 16% (mean difference 5%, p <0.0001 by paired T-test) and 124 ± 65 mL and 129 ± 62 mL (mean difference 5 mL, p = 0.001 by paired T-test), respectively for prospective and retrospective gating.

Conclusions: Retrospective list-mode based gating could successfully use 98% of the beats, while losing only 2% of outlier beats. Functional parameters of LVEF and volumes calculated by retrospective gating methods demonstrated high correlations with the values with conventional prospective gating, allowing for clinical application.

• Download figure
• Open in new tab
• Download powerpoint