Assessment of myocardial perfusion defects with first-pass arterial-phase spectral photon counting CT

Background The introduction of spectral detector CT scanners presents an opportunity to assess coronary arteries and myocardial perfusion in single acquisition. By measuring the energy of individual photons, these systems facilitate comprehensive characterization of myocardial tissue(1-3). Purpose The aim of this study was to assess the feasibility of identifying myocardial perfusion defects using PCCT-derived myocardial tissue information from single arterial-phase coronary CT angiography (CCTA). Methods Patients referred for PCCT CCTA (NAEOTOM Alpha) who had at least one reference examination within three months were enrolled in this IRB-approved study. Reference exams included stress perfusion single-photon emission computed tomography (SPECT), stress perfusion ammonia positron emission tomography (PET), cardiac magnetic resonance imaging (CMR) with stress perfusion and/or late gadolinium enhancement (LGE), rest or stress transthoracic echocardiography, or invasive coronary angiography. Virtual monoenergetic images were created at eight energy levels (40-140 keV in 10 kV steps) and analysed on a dedicated workstation (Syngo.via). Each short-axis slice was spatially matched to the corresponding myocardial muscle segment on the reference examination. For invasive coronary angiography vessel regions were matched. Regions of interest (ROIs) were manually placed within the left ventricular myocardium on short-axis PCCT CCTA reconstructions. Based on the reference standard, mean Hounsfield Units in ROIs (0.5–1.0 cm²) were measured in the following AHA segments with: decreased perfusion, the adjacent segment, and remote healthy myocardium. The septum was analysed separately due to increased blooming artefact (Figure 1). Each ROI was evaluated across all eight virtual monoenergetic reconstructions. Results Overall, we included 58 patients (1,267 myocardial ROIs). On a per-patient basis at 40 keV, spectral analysis of first-pass arterial-phase PCCT demonstrated a sensitivity of 68.0%, specificity of 86.7%, PPV of 80.9%, NPV of 76.4%, and an overall diagnostic accuracy of 79.6% for the detection of significant myocardial perfusion defects. In patients without visually apparent perfusion defects on CT, attenuation differences between myocardial regions with decreased perfusion and all other regions were statistically significant at all monoenergetic levels (p < 0.001), except at 120 keV (p = 0.68) and 140 keV (p = 0.292) (Figure 2). In contrast, in patients with visually apparent myocardial scar on CT, attenuation values differed significantly among all four tissue types across all monoenergetic energy levels (p < 0.001). Conclusion Spectral photon counting CT may enable identification of reduced myocardial perfusion from a single first-pass arterial-phase CCTA acquisition when using multiple virtual monoenergetic reconstructions as compared to reference.