Blood Flow Energy Identifies Coronary Lesions Culprit of Future Myocardial Infarction

Abstract

The present study establishes a link between blood flow energy transformations in coronary atherosclerotic lesions and clinical outcomes. The predictive capacity for future myocardial infarction (MI) was compared with that of established quantitative coronary angiography (QCA)-derived predictors. Angiography-based computational fluid dynamics (CFD) simulations were performed on 80 human coronary lesions culprit of MI within 5 years and 108 non-culprit lesions for future MI. Blood flow energy transformations were assessed in the converging flow segment of the lesion as ratios of kinetic and rotational energy values (KER and RER, respectively) at the QCA-identified minimum lumen area and proximal lesion sections. The anatomical and functional lesion severity were evaluated with QCA to derive percentage area stenosis (%AS), vessel fractional flow reserve (vFFR), and translesional vFFR (ΔvFFR). Wall shear stress profiles were investigated in terms of topological shear variation index (TSVI). KER and RER predicted MI at 5 years (AUC = 0.73, 95% CI 0.65-0.80, and AUC = 0.76, 95% CI 0.70-0.83, respectively; p < 0.0001 for both). The predictive capacity for future MI of KER and RER was significantly stronger than vFFR (p = 0.0391 and p = 0.0045, respectively). RER predictive capacity was significantly stronger than %AS and ΔvFFR (p = 0.0041 and p = 0.0059, respectively). The predictive capacity for future MI of KER and RER did not differ significantly from TSVI. Blood flow kinetic and rotational energy transformations were significant predictors for MI at 5 years (p < 0.0001). The findings of this study support the hypothesis of a biomechanical contribution to the process of plaque destabilization/rupture leading to MI.

Keywords: Computational fluid dynamics; Fractional flow reserve; Kinetic energy; Myocardial infarction; Quantitative coronary angiography; Rotational energy; Wall shear stress.

Fig. 1

Workflow of the study. From a population of 80 patients, 80 coronary lesions site of myocardial infarction (MI) within 5 years (future culprit) were compared to 108 non-culprit lesions for future MI. The three-dimensional coronary vessel geometries were reconstructed from quantitative coronary angiography (QCA) images and used to obtain anatomical (%AS) and functional (vFFR, ΔvFFR) clinical indicators of coronary disease, and to perform computational fluid dynamics (CFD) simulations. From CFD simulations: (i) wall shear stress was quantified in terms of topological shear variation index (TSVI); (ii) blood flow energy transformations in the converging segment of the lesion were quantified in terms of ratio between kinetic (KER) or rotational energies (RER) at MLA and proximal luminal sections of the lesion. The predictive power for future MI of %AS, vFFR, ΔvFFR, and TSVI was evaluated by means of receiver operating characteristic (ROC) curves and compared with the predictive power of KER and RER

Fig. 2

Left panel: QCA-based definition of the atherosclerotic lesion, with proximal and distal lesion cross-sections identified by the intersection between measured lumen area curve and linear regression identifying the reference interpolated lumen area, while minimum lumen area (MLA) section was identified as the section presenting the lowest surface area value inside the lesion segment (LS), defined as the segment between proximal and distal lesion luminal sections. The proximal lesion segment (PLS) was defined as the segment identified by proximal and MLA luminal sections. Right panel: schematic representation of velocity (u) and vorticity (ω) vectors for a fluid element, and mathematical definition of specific kinetic energy (KE_u), kinetic energy ratio (KER), ω, enstrophy (ε), and rotational energy ratio (RER)

Fig. 3

A Volume visualizations of cycle-average specific kinetic energy (KE_u) and enstrophy (ε) in two explanatory lesion types, one future culprit and one non-future culprit. The distributions of cycle-average KE_u and ε on the proximal and minimum lumen area sections are also presented. B Box plots of KER, RER, %AS, vFFR, and ΔvFFR comparing future culprit (FC) and non-future culprit (NFC) lesions (statistically significant differences between FC and NFC lesions are reported in terms of p-values). The p-values are reported for each quantity

Fig. 4

Receiver operating characteristic (ROC) curves for myocardial infarction (MI), ST-elevation MI (STEMI), non-ST-elevation MI (NSTEMI) for %AS, vFFR, ΔvFFR, TSVI, KER, and RER. The values of area under the curve (AUC) with 95% confidence interval (CI) and p values are reported for each quantity

Fig. 5

Time-to-event curves. Significantly divergent Kaplan–Meier curves for future myocardial infarction (MI) are represented at 4-year follow-up for KER and RER. Red and green curves refer to values above or below the threshold values obtained from the ROC analysis, respectively. Hazard ratio (HR) refers to the whole follow-up time interval (i.e., 5 years)

Fig. 6

Scatter plots of rotational energy ratio (RER) and kinetic energy ratio (KER) vs. %AS (panel A), vFFR (panel B), and ΔvFFR (panel C). Spearman correlation coefficients (r) and corresponding p-values are reported

Fig. 7

Scatter plots of rotational energy ratio (RER) and kinetic energy ratio (KER) vs. topological shear variation index (TSVI). Spearman correlation coefficients (r) and corresponding p-values are reported

Fig. 8

Velocity streamlines visualization color-colored with cycle-averaged local enstrophy (ε) values. Enstrophy is higher in the converging flow segment of the lesion, where the streamlines are basically straight, while lower values of enstrophy are found in the anatomical expansion of the vessel located distal to the MLA section, where flow recirculation regions are located