Ability of the coronary angiography-derived index of microcirculatory resistance to predict microvascular obstruction in patients with ST-segment elevation

Background The coronary angiography-derived index of microvascular resistance (caIMR) correlates well with the index of microcirculatory resistance (IMR), which predicts microvascular obstruction (MVO). However, the relationship between caIMR and MVO remains unclear. Aim To evaluate the predictive ability of caIMR of MVO after ST-segment elevation myocardial infarction (STEMI). Methods CaIMR was calculated using computational flow and pressure simulation in patients with STEMI in whom MVO status had been assessed by cardiac magnetic resonance (CMR) after successful primary percutaneous intervention at Peking University First Hospital between December 2016 and August 2019. The clinical, biochemical, echocardiographic, and CMR characteristics were assessed according to MVO status. The predictive value of the clinical parameters and caIMR was evaluated. Results Fifty-three eligible patients were divided into an MVO group (n = 32) and a no-MVO group (n = 21). The caIMR tended to be higher in the MVO group (41.6 U vs. 30.1 U; p = 0.136). CaIMR and peak cardiac troponin-I (cTNI) were independent predictors of MVO (per 1-U increment in caIMR: odds ratio [OR] 1.044, 95% confidence interval [CI] 1.004–1.086, p = 0.030; per 1 ng/L increase in peak cTNI: OR 1.018, 95% CI 1.003–1.033, p = 0.022). In receiver-operating characteristic curve analysis, when a cut-off value of 45.17 U was used, caIMR had some ability to predict MVO (area under the curve 0.622, 95% CI 0.478–0.752, p = 0.127). Conclusions CaIMR and peak cTNI were independent predictors of short-term MVO in patients with STEMI who had undergone successful primary percutaneous coronary intervention and may help to identify those at high risk of MVO.


Introduction
The coronary microcirculation consists of pre-arterioles and arterioles that are responsible for maintenance of appropriate myocardial perfusion.Functional and structural abnormalities of the microcirculation contribute to an imbalance in the supply and demand for oxygen, resulting in myocardial ischemia, a condition now referred to as coronary microvascular dysfunction (CMD) (1).CMD has been observed in almost half of patients with ST-segment elevation myocardial infarction (STEMI) despite successful restoration of epicardial coronary artery blood flow (2).However, the coronary microcirculation cannot be visualized by current angiographic systems.Further development of techniques would allow assessment of the functional status of the coronary microcirculation.Measurement of the index of microvascular resistance (IMR) is a gold standard method for functional assessment of the coronary microvasculature (3) and has been recommended for patients with angina who are suspected to have CMD.The IMR is a quantitative and reproducible hyperemic index that is derived from measurements obtained by a pressure-temperature sensor-tipped coronary guidewire.This index is relatively specific for the microcirculation and is not affected by hemodynamic perturbations (4).However, it has had limited application in clinical practice because of the requirement for dedicated intracoronary instruments, need for administration of hyperemic agents, and a relatively long procedural time.
MVO indicates a severe perfusion defect and profound damage to the microcirculation and has a significant association with adverse outcomes, including an increased risk of all-cause mortality and re-admission for heart failure (5).Cardiac magnetic resonance (CMR) is the gold standard for evaluation of MVO but has limited applications in clinical practice because of contraindications to CMR and inaccessibility at the time of primary percutaneous coronary intervention (PPCI).
With advances in technique, caIMR allows the function of the coronary microcirculation to be assessed at the time of the index procedure without the need for a pressure sensor wire and hyperemic agents.Previous studies have confirmed a good correlation and high agreement between caIMR and pressurewire-based IMR (6-8).Another recent study demonstrated that caIMR has prognostic value in patients with successfully revascularized STEMI (5).However, there is limited information on the relationship between caIMR and MVO.In this study, we evaluated the ability of caIMR to predict the presence of MVO.

Study population
Eighty-six patients who were admitted with STEMI to Peking University First Hospital and underwent PPCI with measurement of CMR between December 2016 and August 2019 were enrolled in a prospective cohort study.The cohort study protocol was approved by the Ethics Committee of Peking University First Hospital and all the patients provided written form consent.This is the post-hoc analysis of this cohort study using anonymous clinical data.STEMI was diagnosed based on chest pain lasting for at least 30 min and ST-segment elevation >2 mm in at least two contiguous leads.Data for patients with angiographic images that were inadequate for functional analysis (n = 3), those in whom CMR was performed more than 30 days after the intervention (n = 19), and those in whom CMR was performed after selective percutaneous intervention (n = 11) were excluded.Finally, data for 53 patients with successfully revascularized STEMI were eligible for inclusion in the analysis.

CMR analysis
CMR imaging was performed before discharge (generally 1 week after the index event).All patients were examined using a 1.5-Tesla magnetic resonance imaging scanner (GE Healthcare, Chicago, IL, USA).CMR images were reviewed using commercially available software (Circle Cardiovascular Imaging, Calgary, AB, Canada) by two observers working independently and blinded to the clinical data and the results of physiological assessment.MVO was defined as a hypointense region within a hyperenhanced area on the delayed enhancement images.Hyperenhancement was defined as a region with a signal intensity threshold that was 5 standard deviations above the mean signal intensity of the remote reference myocardium.

Measurement of CaIMR
Coronary angiograms were centrally analyzed in a blinded fashion by a core laboratory using the FlashAngio system (RainMed Medical Technology Co. Ltd., Suzhou, China).The caIMR was estimated in three steps.First, we chose angiograms from at least two different projections to construct threedimensional mesh models of the coronary arteries.Second, we estimated caFFR using a computational fluid dynamics method.Third, we estimated caIMR using the following formula: Ai et al (9) defined caIMR to assess microcirculation as: L represents the length of the target vessel from the inlet to the distal position (L = 75 mm).Meuwissen et al. (10) proposed the HMR(hyperemic microvascular resistance) to evaluate microvascular dysfunction as: (2) (Pd) hyp is the mean pressure at the distal position at maximal hyperemia.And V hyp is the mean flow velocity at the distal position when hyperemia is maximal V diastole is the mean flow velocity at the distal position at diastole V diastole was derived using the TIMI (Thrombolysis in Myocardial Infarction) frame count method.The entire diastolic period provides higher flow velocity and lower microvascular resistance than the whole car diac cycle.And an intracoronary injection of contrast medium can induce some degree of hyperemia.So V hyp is assumed to be proportional to V diastole .K is a constant (K = 1.1), obtained from a previous study (6).caFFR is the coronary angiography-derived fractional flow reserve. (Pd The mean pressure at the aorta at maximal hyperemia ((Pa) hyp ) was estimated on the basis of mean arterial pressure (MAP) during the index procedure, which equals to MAP-MAP*0.2when MAP ≥ 95 mmHg and MAP-MAP*0.15when MAP < 95 mmHg, according to a previous study.The pressure drop (ΔP s ) across a stenosis was computed from the CFD simulation,which will not be published in details because it is the property of Rainmed Ltd.ΔP is the pressure drop along the meshed coronary arteries in the vessel path from the inlet to the most distal position.
According to formulations (Equations 1-4), we can deduce: Evaluation of caIMR and derivation of the formula have been described in detail by Ai et al. (9).It takes less than 1 min to measure CaIMR.We chose the major epicardial artery with the highest caIMR.

Collection of data
Demographic and clinical data, information on cardiac risk factors, echocardiographic and CMR data, and the results of laboratory investigations during hospitalization were retrieved manually from the hospital electronic medical records system.

Statistical analysis
Categorical variables are expressed as the number and relative frequency (percentage).Continuous variables are shown as the mean ± standard deviation if they were distributed normally and as the median [interquartile range (IQR)] if not.Continuous data with a normal distribution were compared using the Kolmogorov-Smirnov test and by the independent t-test or Mann-Whitney U test if not normally distributed.Categorical variables were examined using the chi-squared test.The ability of peak cardiac troponin-I (cTNI) and caIMR to predict the presence of MVO was examined by receiver-operating characteristic curve analysis.The optimal cut-off values were those that yielded the greatest product of sensitivity and specificity in predicting MVO.The areas under the curve (AUCs) for peak cTNI, caIMR, and a combined model were compared using Delong's test with MedCalc statistical software (MedCalc Software Ltd., Ostend, Belgium).Predictors of MVO were sought in a univariate logistic regression model.All variables with a p-value of <0.20 were entered into a multivariable logistic regression model.The results of the logistic regression analysis are presented as odds ratios with 95% confidence intervals (CIs).All other statistical analyses were performed using SPSS statistics version 26.0.0.0 (IBM Corp., Armonk, NY, USA).All tests were two-tailed, and a p-value < 0.05 was considered statistically significant.

Results
Fifty-three of the 86 patients with STEMI during our study period met the study inclusion criteria (Figure 1) and were classified based on the presence of MVO into an MVO group (n = 32, 60.4%) and a no-MVO group (n = 21, 39.6%).

Baseline characteristics
The baseline demographic, clinical, biochemical, and procedural characteristics are presented for the whole cohort and when stratified according to the presence of MVO in Table 1.There was no significant difference in age or sex distribution, cardiac risk factors, or procedural characteristics according to MVO status.Mean door-to-balloon time was 72.5 ± 32.9 min.The most frequent culprit vessel was the left anterior descending artery (60.6%) in the MVO group and the right coronary artery (47.6%) in the no-MVO group.TIMI flow grade 3 was achieved after PCI in 96.2% of cases (MVO group, 93.8%; no-MVO group, 100%).There was no significant between-group difference in low-density lipoprotein cholesterol, high-sensitivity C-reactive protein, estimated glomerular filtration rate, brain natriuretic peptide, or glycated hemoglobin.Patients with MVO had significantly higher values for peak cTnI (110.3 ng/L vs. 30.5 ng/L, p = 0.0001) and creatine kinase-myocardial band (273.1 U/L vs. 115.7 U/L, p = 0.004) and tended to have a higher caIMR (41.6 U vs. 30.1 U, p = 0.136, Figure 2).Figure 3 shows a case examples of patients with microcirculatory dysfunction in the culprit vessel after successful primary PCI.

Discussion
This study evaluated the ability of caIMR to predict MVO in patients with STEMI after successful PPCI and had two main findings: (1) peak cTNI and caIMR values were independently associated with MVO detected on CMR and (2) a high caIMR tended to be more common in patients with MVO.
The finding of MVO on CMR indicates severe damage at the microvascular level and predicts adverse ventricular remodeling and an increased risk of mortality and morbidity (11)(12)(13).However, CMR is not available at the time of percutaneous coronary intervention or widely accessible in routine practice.Therefore, it cannot be used to guide prompt optimized medical therapy and improve the microcirculation.Previous studies of the relationship between IMR and MVO have shown that the IMR is higher in patients with STEMI and MVO (14-16).However, IMR is typically measured using a thermodilution wire with a sensor near its tip, which has less torquability in comparison with workhorse wires, leading to technical difficulty and increased cost (17).Furthermore, IMR requires pharmacological induction of hyperemia, which may be uncomfortable for patients and has a risk of morbidity from arrhythmia.Previous research suggests that caIMR may be a promising alternative indicator of IMR (8,9).The caIMR is a readily available index and has been confirmed to have an incremental prognostic value over clinical and angiographic factors (6).In this study, although the between-group difference in caIMR did not reach statistical significance, patients with STEMI and MVO tended to have a higher caIMR than those without MVO.This finding is concordant with data from Cuculi et al., who   reported a mean invasive IMR value of 42.9 U in post-STEMI patients with MVO and 31.1 U for those without MVO (14).The failure of our result to reach statistical significance may be attributable to a relatively small infarct size.In our study, the mean left ventricular ejection fraction in patients with MVO was preserved at 49.8%, which indicates a small infarct size with a limited effect on IMR, as demonstrated by Marques et al. ( 18).
According to a previous study, obesity was associated with a high caIMR (19).A relatively higher body mass index of patients in the group without MVO may has played a role in our findings.
The limited number of study participants may also have constrained the statistical power of our analysis.
Several previous studies have suggested that cTNI and highsensitivity cTNI have a predictive value (20-22).Hallen et al. found that the cTNI level at 24 or 48 h was independently associated with MVO (23).Other studies have shown that the peak cTNI concentration correlates well with the extent of MVO and infarct mass and can predict clinical outcomes (24-27).Thus, patients with MVO have a larger myocardial infarct size than those without MVO.In our study, the peak cTNI level was an independent predictor of the presence of MVO, which is consistent with previous reports (23,28).
CaIMR is a readily available, quantitative, and wire-free method that can be used to assess acute microvascular dysfunction, which has been shown to correlate well with IMR and be an independent predictor of long-term outcomes.In line with another recent study (28), we found caIMR to be an independent predictor of MVO that may be used to identify patients at high risk of developing MVO and guide personalized precision medicine during and following PPCI.Although the ability of caIMR to discriminate MVO was inferior to that of peak cTNI or a combination of peak cTNI and caIMR, caIMR is an index that is readily available in a catheterization laboratory and has advantages over other methods at the time of STEMI.The clinical implication of this finding is that caIMR is useful for risk stratification of persons who may benefit most from early adjunctive treatment and individualized treatment strategies aimed at microvascular recovery following PPCI.
Efforts are ongoing to identify effective adjunctive therapies to improve microvascular function and clinical outcomes after STEMI, including intracoronary administration of glycoprotein IIb/IIIa inhibitors, use of coronary vasodilators, and periprocedural antiinflammatory interventions (29).CaIMR may help us to detect patients at high risk of developing MVO and adverse clinical outcomes.Perhaps most important, better assessment of microvascular dysfunction may pave the way for development of therapies aimed specifically at microvascular dysfunction.

Limitations
This study has several limitations, First, it had a single-center observational design and included a small sample size.Therefore, it may not have been fully representative of patients encountered in routine clinical practice.Second, we did not evaluate the extent of the microvasculature or quantify the area at risk or infarct size using CMR, which limits any further exploration of  Discriminatory ability for prediction of the presence of MVO.The graphs compare the receiver-operating characteristic curves for the ability of a combination of post-PCI caIMR and peak cTNI (solid black line), peak cTNI (dashed-dotted black line), and post-PCI caIMR (dashed-doubledotted black line) to predict CMR-defined MVO.(A) The peak cTNI tended to be a better predictor than caIMR (AUC 0.829 (95% CI 0.700-0.918,p < 0.0001 vs. AUC 0.622, 95% CI 0.478-0.752,p = 0.127; difference in AUC 0.207, p = 0.057).(B) There was no significant difference between peak cTNI and the combination of peak cTNI and caIMR (AUC 0.854, 95% CI 0.73-0.936,p < 0.001 vs. AUC 0.829, 95% CI 0.700-0.918],p < 0.0001; difference in AUC 0.025, p = 0.463); (C) The AUC was higher for the combination of cTNI and peak cTNI than for caIMR (AUC 0.854, 95% CI 0.73-0.936,p < 0.0001 vs. AUC 0.622, 95% CI 0.478-0.752],p = 0.127; difference in AUC 0.232, p = 0.008).AUC, area under the curve; caIMR, coronary angiography-derived index of microcirculatory resistance; cTNI, cardiac troponin I; CI, confidence interval; MVO, microvascular obstruction.the association between caIMR and MVO.Third, we did not measure the IMR derived using a conventional pressure wire.Therefore, we could not validate the diagnostic accuracy of caIMR.However, a recent study that used the same software and methodology as in our study to calculate caIMR confirmed a high correlation between caIMR and IMR (9).Fourth, clinical outcomes are not included in this study.Fifth, caIMR was measured based on the angiography-derived mean pressure and diastolic flow velocity in the contrast-induced sub-hyperemia which may result the discordance between caIMR and IMR.However, previous study has confirmed the good diagnostic of caIMR in patients with stable/unstable angina pectoris (9).Sixth, caIMR is not repeatable.CFR (coronary flow reserve) by transthoracic doppler methods is a noinvasive method with high feasibility, which has been proved to have excellent prediction of clinical outcomes, precise discrimination between microcircle and epicardial conduit status and excellent assessment of the severity of coronary artery stenosis (30)(31)(32).CFR may be integrated with caIMR during the post angioplasty follow-up in our future study.

Conclusion
Peak cTNI and caIMR are predictors of MVO in patients with STEMI that has been successfully revascularized.caIMR holds promise as a method for evaluation of CMD and prediction of the presence of MVO at the time of STEMI, which would facilitate prompt adjunctive therapy aimed at microvascular recovery.A further study is warranted to clarify in more detail the diagnostic performance and prognostic value of caIMR in patients with STEMI.

FIGURE 1 Flow
FIGURE 1Flow chart showing the procedure used for patient selection.caIMR, coronary angiography-derived index of microvascular resistance; CMR, cardiac magnetic resonance; MVO, microvascular obstruction; PCI, percutaneous coronary intervention; STEMI, ST-segment elevation myocardial infarction.

FIGURE 2
FIGURE 2Boxplot showing the distribution of caIMR of the two groups.caIMR, coronary angiography-derived index of microvascular resistance; MVO, microvascular obstruction.

TABLE 1
Clinical and procedural characteristics.

TABLE 2
Echocardiographic and cardiac magnetic resonance parameters.

TABLE 3
Independent predictors of microvascular obstruction assessed by cardiac magnetic resonance.

TABLE 4
Performance of independent predictors of microvascular obstruction at optimal cut-off values.