The spectrum of angiography‐derived IMR according to morphological and physiological coronary stenosis in patients with suspected myocardial ischemia

Abstract Background Coronary microvascular dysfunction is crucial in determining myocardial ischemia; however, the relationship between epicardial coronary diameter stenosis (DS) and the index of microcirculatory resistance (IMR) remains unclear. We sought to explore the distribution of coronary angiography‐derived IMR (angio‐IMR) in patients with suspected myocardial ischemia. Methods The study included 480 patients with suspected myocardial ischemia, all of whom underwent coronary angiography. According to the severity of coronary DS, patients were divided into three groups: mild (DS < 50%), intermediate (DS 50%–70%), and severe (DS > 70%). Angio‐IMR and fractional flow reserve (FFR) were calculated based on coronary angiography images through the principle of computational flow and pressure simulation. Results Of the 480 patients, the mean age was 67.23 ± 9.44 years, with 55.4% male. There were 193 (40.2%) patients in the mild group, 189 (39.4%) patients in the intermediate group, and 98 (20.4%) patients in the severe group. The average angio‐IMR of the mild group was 30.8 ± 14.9, which was significantly higher than those of the intermediate group (26.7 ± 13.0) and the severe group (17.9 ± 8.4) (p < .001). In the correlation analysis, angio‐IMR was negatively correlated with DS (rho = −0.331, p = .001) and positively correlated with angio‐FFR (rho = 0.483, p < .001). By multivariate logistic regression analysis, angio‐FFR ≤ 0.8 (odds ratio, 0.184; 95% confidence interval, 0.106–0.321) was the only independent predictor of coronary microvascular dysfunction. Conclusion In patients with suspected myocardial ischemia, coronary microcirculation is significantly associated with morphological and physiological coronary stenosis. (ClinicalTrials.gov: NCT05435898)


| INTRODUCTION
Both functional and structural abnormalities of the coronary epicardial vessels can lead to myocardial ischemia, and the presence of the broader coronary microvasculature also plays an important role. Increased coronary microvascular resistance can impair coronary blood flow, cause angina pectoris, and lead to diastolic and systolic dysfunction. 1,2 Coronary microvascular dysfunction (CMD) is widespread in obstructive and non-obstructive coronary artery disease.
Previous studies have shown that CMD exists in up to two-thirds of patients with non-obstructive coronary artery disease (CAD). [2][3][4] Therefore, the assessment of coronary microvascular function is critical.
At present, noninvasive examinations such as echocardiography, positron emission tomography, cardiac magnetic resonance, and invasive tests, mainly including the index of microcirculation resistance (IMR) and coronary flow reserve (CFR), were used for the evaluation of coronary microvascular function. 5,6 Among all the methods, wire-derived IMR is currently the most common one to evaluate microvascular function. 7 Still, it requires expensive pressure guide wires and injection of vasodilator drugs, which limits its wide application. In the past 2 years, coronary angiography-derived IMR (angio-IMR) has been developed with high diagnostic accuracy in CMD and could provide a better choice for evaluating microvascular function. [8][9][10][11][12][13][14] Previous studies showed that IMR was independent of epicardial stenosis, 15,16 indicating that microvascular function was stable and independent of coronary obstruction. However, many other studies showed a link between fractional flow reserve (FFR) and CMD. 17,18 Given that the relationship between coronary epicardial structural or functional stenosis and CMD is controversial, we sought to investigate the distribution of CMD using angiography-derived IMR in patients with suspected myocardial ischemia. (3) Severe myocardial bridge; (4) Chronic total occlusion of the coronary artery; (5) severe valvular heart disease; (6) A history of coronary artery bypass surgery; (7) Ostial lesions ≤3 mm from the aorta; (8) Coronary angiographic image quality of interrogated vessels did not meet the FLASH software requirements (poor contrast opacification, extreme vascular overlap, or distortion). 14,19 According to the study flow chart in Figure S1, patients were divided into three groups according to coronary diameter stenosis (DS). Our study was approved by the ethics committee, and all patients provided written informed consent.

| Data collection
A total of 480 patients were eventually included in this study. Basic demographic information, cardiovascular risk factors, laboratory tests, echocardiography, and CA results were documented in detail for all participants. Downloaded Digital Imaging and Communications in Medicine angiography images from the picture archiving and communication system and recorded real-time MAP during angiography. Morphological stenosis was assessed by visual assessment of coronary angiography. Vessels with the most severe DS were further analyzed for physiology, and vessels with the same degree of stenosis were randomly assigned. CAD is defined as at least one main coronary artery stenosis ≥50% in diameter, more than two vessels as multivessel disease, and DS < 50% as no-CAD.

| Angiography-derived IMR and FFR measurements
The angio-IMR and angio-FFR measurements were conducted using the FlashAngio system (including the FlashAngio console, FlashAngio software, and Flash Pressure transducer; Rainmed Ltd.). Details of measurement and procedures of angio-IMR and angio-FFR have been published previously. 14,19 The Digital Imaging and Communications in Medicine images and corresponding MAP were transferred to the FlashAngio workstation, and the interrogated vessels were selected for analysis. The positions from the inlet to the distal end (vessel length) along the target vessel path were marked, and the vessel contour and three-dimensional mesh were determined.
The image's start and end frames were selected to calculate the flow velocity (V diastolic ) by the thrombolysis in the myocardial infarction frame count method. 20

| Statistics analysis
Numerical data were presented as the mean ± SD or medians and interquartile ranges, and categorical variables were presented as percentages. For intergroup comparisons of numerical variables, independent sample Student's t-test or one-way analysis of variance was used as appropriate, and non-normally distributed continuous data were compared with the Wilcoxon rank-sum test. For comparisons of categorical variables, the Chi-square test and Fisher's exact tests were used, with a Bonferroni-adjusted significance level.
Spearman or Pearson correlation was applied for correlation analysis as appropriate. CMD was defined as angio-IMR > 25, 5 and binary logistic regression analysis was used to determine the risk factors of CMD. Covariates in the multivariable model were gender, age, conventional cardiovascular risk factors (hypertension, diabetes, smoking, dyslipidemia, obesity, stroke, prior percutaneous coronary intervention, left ventricular hypertrophy), DS < 50%, and angio-FFR < 0.8. All values were two-tailed and considered statistically significant at p-value < .05. All statistical analyses were performed using SPSS v.22 (IBM Inc). Data were visualized using GraphPad Prism v.9.0 (GraphPad Software Inc).

| Clinical characteristics according to DS
Patients' baseline characteristics stratified by coronary stenosis categories are shown in Table 1. The age of patients with <50% DS (mild group) was significantly lower than those with 50%-70%

DS (intermediate group) or those with >70% DS (severe group)
F I G U R E 1 Representative examples of angio-IMR and angio-FFR analysis. A 70 years old female patient with a 10 years history of hypertension was hospitalized with recurrent chest tightness and chest pain for 1 month. Coronary angiography suggested no significant stenosis in the three main coronary arteries, and angio-IMR analysis indicated the presence of microvascular dysfunction in the LAD artery (angio-FFR = 0.91, angio-IMR = 35.9). Angio-FFR, angiography-derived fractional flow reserve; Angio-IMR, angiography-derived index of microvascular resistance; LAD, left anterior descending.
(65.8 ± 9.7 vs. 68.3 ± 8.9 years or 65.8 ± 9.7 vs. 68.0 ± 9.6 years, respectively) (p = .022). In the mild group, the proportion of male patients was significantly less than that of the severe group In terms of drug use, the rates of beta-blockers (p = .044), antiplatelet aggregation drugs (p < .001), and angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers (p < .001) in the intermediate group and the severe group were significantly higher than that in the mild group, while there was no difference in statin use (p = .292). In the laboratory test, the glycated hemoglobin (6.80 ± 1.36 mmol/L) in the severe group was significantly higher than that in the mild group and the intermediate group

| Correlation of angio-IMR with DS and angio-FFR
In the correlation analysis between angio-IMR and continuous variables ( Table 2)

| Distribution spectrum of angio-IMR
The distribution of angio-IMR is shown in Figure 2. According to the number of obstructed blood coronary, the angio-IMR of patients in the no-CAD group was significantly greater than that in the 1-vessel group or the multivessel group (both p < .05) (Figure 2A). However, the angio-IMR was not significantly different among patients with three major coronary arteries (both p > .05) ( Figure 2B). Based on the different stenosis degrees, the angio-IMR of patients in the mild group was significantly higher than that in the intermediate and severe groups, as shown in Figure 2C (both p < .05). The angio-IMR of patients with the angio-FFR > 0.8 was higher than that of the patients with the angio-FFR ≤ 0.8 (p < .001) ( Figure 2D).  showed that angio-FFR < 0.8 (OR, 0.184; 95% CI, 0.106-0.321, p < .001) was an independent risk factor of CMD (Table S2).

| DISCUSSION
In this study, we pinpointed the following findings: (1) in patients with suspected myocardial ischemia, the distribution of angio-IMR was different in the stratification of the degree of epicardial stenosis by visual assessment; (2) angio-IMR and angio-FFR was moderately positively correlated, and angio-IMR was negatively correlated with DS; and (3) Angio-FFR > 0.8 was an independent predictor of CMD.
CMD is an essential mechanism of myocardial ischemia. 2  to wire-derived IMR, has greatly simplified the IMR assessment process. Several studies have shown that angio-IMR is highly consistent with wire-derived IMR and has high accuracy in diagnosing CMD. 8,11,14,21 Therefore, in this study, it is feasible to study the distribution of microcirculatory resistance in patients with suspected myocardial ischemia using angio-IMR.
The present study suggests that DS was negatively correlated with angio-IMR; angio-FFR was positively correlated with angio-IMR.
Patients with stenosis <50% had higher angio-IMR levels. In contrast, patients with severe stenosis had lower angio-IMR levels, suggesting more microcirculatory disturbances in patients with non-obstructive CAD, whereas patients with morphologically severe epicardial vessel stenosis versus functional stenosis have preserved microvascular function. Similarly, previous studies have also shown an association between IMR and functional or structural stenosis of epicardial vessels. In a cohort with moderate coronary stenosis, FFR and IMR were moderately positively correlated (r = .451, p < .001). 18 While another study also demonstrated that there was a weak positive correlation (r = .16, p < .01) between FFR and IMR. 17 In addition, a study of microcirculatory function in patients with acute chest pain assessed by positron emission tomography showed a higher incidence of CMD (42%) but lower mean CFR in low to moderaterisk patients than in high-risk patients. 22 However, inconsistent with the above findings, some argue that the wired-IMR for measuring microvascular resistance was flawed because it ignored the contribution of collateral vessels to the microvascular. Therefore, when there is severe occlusion of epicardial vessels, the influence of collateral vessels must be excluded, and the IMR measured by the pressure guidewire needs to be corrected. Yong et al. showed that IMR was independent of epicardial stenosis in acute myocardial infarction. 16 Subsequent studies have also confirmed that IMR has no relationship with epicardial vessel stenosis or FFR. 15 In the present study, we also analyzed the relationship between laboratory indicators and angio-IMR. The results showed a weak negative correlation between angio-IMR and fasting blood glucose, hemoglobin A1c, creatinine and B-type natriuretic peptide, and a weak positive correlation with glomerular filtration rate, which F I G U R E 3 Accordance and discordance of angio-FFR and angio-IMR across DS. According to different coronary diameter stenosis, in the angio-FFR and angio-IMR stratified models, the incidence of microcirculatory dysfunction decreased with the increase of stenosis, while the incidence of patients reflecting coronary epicardial vascular dysfunction increased. Angio-FFR, angiography-derived fractional flow reserve; Angio-IMR, angiography-derived index of microvascular resistance; DS, diameter stenosis.
indicated that the higher the angio-IMR, the lower the patient's blood glucose level, B-type natriuretic peptide level and the higher the glomerular filtration function, which suggested that there might be certain links between angio-IMR and the patient's blood glucose level, renal and cardiac functional status. However, the correlation coefficient was too low to make this relationship convincing. On the other hand, multifactorial regression analysis showed that traditional cardiovascular risk factors such as diabetes, dyslipidemia, and hypertension were not independent predictors of CMD, which was consistent with the results of a prospective observational study. 17 However, it will be of interest to investigate whether the stratifica- Unlike radionuclide myocardial imaging and coronary angiography, linear IMR and CFR require intraoperative medication to dilate coronary arteries, which is not a stable physiological state of the coronary arteries.
In this study, we found discordance between angio-FFR and epicardial vessel DS, especially in patients with 50%-70% stenosis, which is consistent with the FAME study. 24 Still, the proportion of discordant vessels was lower, which may be related to the study's sample size and the different composition of patients enrolled. The gap in CMD prevalence might be explained according to the FAME study. In addition, we found that the proportion of patients with pure microcirculatory disturbance decreased with increasing stenosis, and fewer patients had both functional ischemia and microcirculatory disorder. In patients with stable obstructive CAD, coronary vasodilator capacity and preservation of collateral flow help prevent stress-induced myocardial ischemia. [25][26][27] In the presence of CMD, FFR may underestimate the severity of epicardial vessel stenosis, which could also explain the inconsistency between lesion severity and the extent and severity of myocardial ischemia. In patients without epicardial stenosis, patients with angio-IMR were higher, and the proportion of CMD patients was more elevated.
In a regression analysis of risk factors predicting microvascular dysfunction, we found that diabetes, DS < 50%, and angio-FFR < 0.8 were significantly associated with CMD, independent of cardiovascular factors such as sex and age, and in a multifactorial analysis, angio-FFR < 0.8 was the only independent risk factor for CMD.
Previous studies have shown that female was an independent predictor of CMD, which may be associated with microvascular angina and have a poorer prognosis. 28,29 A recent study by the European Microvascular Group showed that microcirculatory dysfunction was significantly more frequent in women with angina, but no differences in prognosis were found in follow-up studies, and they emphasized that more attention should be paid to improving symptoms in this group of patients. 30 In addition, a study of coronary three-vessel microcirculation showed no conventional clinical risk factors predicting CMD. 17,22 The current study also found that patients with non-obstructive CAD with suspected myocardial ischemia were more likely to have microvascular dysfunction, which is consistent with previous studies. 4,31,32 However, our study suggests that changes in coronary flow physiology are more likely to affect coronary microcirculation than structural stenosis of epicardial vessels. It is now commonly accepted that therapeutic evaluation should be performed according to the different CMD categories, 6 and it remains an issue that needs to be addressed in the future treatment of patients with microcirculatory disturbances.
Patients with CMD have adverse clinical outcomes. Therefore, the assessment of microcirculatory function is particularly important.
Angiography-derived IMR measurements provide more options for coronary microcirculation assessment and have considerable promise in future applications. The current study investigated the distribution of angio-IMR according to different stenosis degrees of epicardial coronary arteries; it is the first time to combine angio-FFR and angio-IMR to perform functional analysis in patients with suspected myocardial ischemia, which has guiding significance for clinical decision-making.
There are some limitations to this study. First, based on the inclusion of patients with suspected myocardial ischemia, this study excluded patients who did not meet the requirements for angio-IMR measurement and also excluded patients who did not record realtime intraoperative aortic pressure. Second, this study was stratified according to different degrees of DS, and we visually selected the vessels with the most severe coronary stenosis rather than all three vessels, which could not reflect the actual myocardial ischemia state.
Third, this study is a retrospective study, which still needs further validation by prospective studies. Finally, 37 (7.2%) patients were excluded from angio-IMR analysis due to the noncompliant quality of angiographic images and lack of aortic pressure.

| CONCLUSION
Angio-IMR correlates with morphological and physiological stenosis of coronary arteries in patients with suspected myocardial ischemia.
The distribution of angio-IMR was distinguished in patients with different degrees of stenosis. In patients with suspected myocardial ischemia, angio-FFR was an independent predictor of CMD.

ACKNOWLEDGMENTS
This study was financially supported by the National Nature Science Foundation of China (82170388), Clinical Research Plan of SHDC