Perfusion dyssynchrony analysis

Abstract

Aims: We sought to describe perfusion dyssynchrony analysis specifically to exploit the high temporal resolution of stress perfusion CMR. This novel approach detects differences in the temporal distribution of the wash-in of contrast agent across the left ventricular wall.

Methods and results: Ninety-eight patients with suspected coronary artery disease (CAD) were retrospectively identified. All patients had undergone perfusion CMR at 3T and invasive angiography with fractional flow reserve (FFR) of lesions visually judged >50% stenosis. Stress images were analysed using four different perfusion dyssynchrony indices: the variance and coefficient of variation of the time to maximum signal upslope (V-TTMU and C-TTMU) and the variance and coefficient of variation of the time to peak myocardial signal enhancement (V-TTP and C-TTP). Patients were classified according to the number of vessels with haemodynamically significant CAD indicated by FFR <0.8. All indices of perfusion dyssynchrony were capable of identifying the presence of significant CAD. C-TTP >10% identified CAD with sensitivity 0.889, specificity 0.857 (P < 0.0001). All indices correlated with the number of diseased vessels. C-TTP >12% identified multi-vessel disease with sensitivity 0.806, specificity 0.657 (P < 0.0001). C-TTP was also the dyssynchrony index with the best inter- and intra-observer reproducibility. Perfusion dyssynchrony indices showed weak correlation with other invasive and non-invasive measurements of the severity of ischaemia, including FFR, visual ischaemic burden, and MPR.

Conclusion: These findings suggest that perfusion dyssynchrony analysis is a robust novel approach to the analysis of first-pass perfusion and has the potential to add complementary information to aid assessment of CAD.

Keywords: adenosine; dyssynchrony analysis; gadolinium; perfusion.

Figure 1

Areas of abnormal myocardial perfusion are characterized by reduced and delayed wash-in of contrast agent. These features are the basis for visual analysis. Quantitative analysis detects and measures absolute differences of perfusion (vertical arrow). To achieve this, myocardial signal intensity curves require temporal realignment before deconvolution with the arterial input function. This is particularly important when high-resolution voxel-wise perfusion quantification is performed. Dyssynchrony analysis instead does not take into account changes in the amplitude of signal intensity but rather isolates and measures the temporal dyssynchrony of the wash-in curves (horizontal arrow). SI, signal intensity.

Figure 2

Schematic representation of perfusion dyssynchrony analysis. First-pass myocardial signal intensity curves are shown for two myocardial segments (Segment A and Segment B) and for the AIF measured from the LV cavity. A₁ and B₁ indicate the point of maximum signal intensity upslope in each segment. A₂ and B₂ indicate the peak of signal intensity in each segment. In this schematic example, the calculated perfusion dyssynchrony indices are coefficient of variation of time to maximum upslope (C-TTMU) 22%; variance of time to maximum upslope (V-TTMU) 2.3 s²; coefficient of variation of time to peak signal (C-TTP) 8.8%; variance of time to peak signal (V-TTP) 1 s².

Figure 3

Intra- (A) and inter-observer (B) variabilities of perfusion dyssynchrony analysis. Bland–Altman graphs for each parameter.

Figure 4

Intra- (A) and inter-observer (B) variabilities of perfusion dyssynchrony analysis. Pearson's r analysis for each parameter.

Figure 5

Results of image quality analysis. No significant differences were observed among groups for image quality (A), respiratory motion (B), and dark rim artefacts (C).

Figure 6

Proposed pathophysiology of myocardial perfusion dyssynchrony. In normal hearts (left), myocardial perfusion shows a temporally homogeneous perfusion. The presence of flow-limiting CAD influences the propagation of the contrast agent through the coronary circulation. Temporal dyssynchrony is caused by the interaction between flow-limiting stenoses and down-stream coronary capacitance. This effect is proportional to the number of flow-limiting coronary lesions. LAD, left anterior descending coronary artery; LCX, left circumflex coronary artery; LM, left main coronary artery; RCA, right coronary artery.