Detection and Classification of Myocardial Infarction Transmurality Using Cardiac MR Image Analysis and Machine Learning Algorithms.

The presence of abnormalities when the left ventricle is deformed is related to the patients' prognosis after a first myocardial infarction. These deformations can be detected by performing a cardiac magnetic resonance (CMR) study. Currently, late gadolinium enhancement (LGE) is considered to be the gold standard when performing CMR imaging. However, CMR with LGE overestimates infarct size and underestimates recovery of dysfunctional segments after myocardial infarction. Based on this statement, the objective is to detect, characterize, and quantify the extent of myocardial infarction in patients with cardiac pathologies, using parameters derived from CMR, in order to obtain greater precision in patients' recovery predictions than when only studying LGE images. For this purpose, we studied the infarct presence and extension from a total of 105 images from 35 patients, and calculated myocardium strain and torsion to characterize and quantify the affected tissue. A total of twenty-one parameters were selected to create predictive models. Moreover, we compared two feature extraction methods, and the performance of five machine learning algorithms. Results show that both temporal and strain parameters are the most relevant to detect and characterize the extent of myocardial infarction. The use of imaging techniques and machine learning algorithms have great potential and show promising results when it comes to detecting the presence and extent of myocardial infarction. The current study proposes a novel approach to detect, quantify, and characterize cardiac infarction by using strain and torsion parameters from different CMR images and different Machine Learning algorithms. This would potentially overcome LGE, the current state of the art technique, in estimating the extension of damaged tissue and enable an objective diagnosis and clinical decision.

Synthetic Generation of Cardiac MR Images Combining Convolutional Variational Autoencoders and Style Transfer.

The number of studies in the medical field that uses machine learning and deep learning techniques has been increasing in the last years. However, these techniques require a huge amount of data that can be difficult and expensive to obtain. This specially happens with cardiac magnetic resonance (MR) images. One solution to the problem is raise the dataset size by generating synthetic data. Convolutional Variational Autoencoder (CVAe) is a deep learning technique which allows to generate synthetic images, but sometimes the synthetic images can be slightly blurred. We propose the combination of the CVAe technique combined with Style Transfer technique to generate synthetic realistic cardiac MR images. Clinical Relevance-The current work presents a tool to increase in a simple easy and fast way the cardiac magnetic resonance images dataset with which perform machine learning and deep learning studies.

Exercise ECG Testing and Stress Cardiac Magnetic Resonance for Risk Prediction in Patients With Chronic Coronary Syndrome.

PURPOSE: Vasodilator stress cardiac magnetic resonance (VS-CMR) has become crucial in the workup of patients with known or suspected chronic coronary syndrome (CCS). Whether traditional exercise ECG testing (ExECG) contributes prognostic information beyond VS-CMR is unclear. METHODS: We retrospectively included 288 patients with known or suspected CCS who had undergone ExECG and subsequent VS-CMR in our institution. Clinical, ExECG, and VS-CMR variables were recorded. We defined the serious adverse events (SAE) as a combined endpoint of acute coronary syndrome, admission for heart failure, or all-cause death. RESULTS: During a mean follow-up of 4.2 ± 2.15 yr, we registered 27 SAE (15 admissions for acute coronary syndrome, eight admissions for heart failure, and four all-cause deaths). Once adjusted for clinical, ExECG, and VS-CMR parameters associated with SAE, the only independent predictors were HRmax in ExECG (HR = 0.98: 95% CI, 0.96-0.99; P = .01) and more extensive stress-induced perfusion defects (PDs, number of segments) in VS-CMR (HR = 1.19: 95% CI, 1.07-1.34; P < .01). Adding HRmax significantly improved the predictive power of the multivariable model for SAE, including PDs (continuous reclassification improvement index: 0.47: 95% CI, 0.10-0.81; P < .05). The annualized SAE rate was 1% (if PD < 2 segments and HRmax > 130 bpm), 2% (if PD < 2 segments and HRmax ≤ 130 bpm), 3.2% (if PD ≥ 2 segments and HRmax > 130 bpm), and 6.3% (if PD ≥ 2 segments and HRmax ≤ 130 bpm), P < .01, for the trend. In patients on ß-blocker therapy, however, only PDs in VS-CMR, but not HRmax, predicted SAE. CONCLUSIONS: We conclude that ExECG contributes significantly to prognostic information beyond VS-CMR in patients with known or suspected CCS.

Role of PCSK9 in the course of ejection fraction change after ST-segment elevation myocardial infarction: a pilot study.

AIMS: Proprotein convertase subtilisin/kexin type 9 (PCSK9) has emerged as a therapeutic target for reducing plasma low-density lipoprotein cholesterol. Beyond lipid control, recent findings suggest a deleterious effect of this protein in the pathogenesis of postmyocardial infarction left ventricle remodelling and heart failure-related complications. The aim of this study was to assess the relationship between circulating PCSK9 and 6 month cardiac magnetic resonance imaging-derived left ventricular ejection fraction (LVEF) after a first ST-segment elevation myocardial infarction (STEMI). METHODS AND RESULTS: We prospectively evaluated 40 patients with a first STEMI, LVEF < 50% and treated with primary percutaneous coronary intervention in which PCSK9 was measured 24 h postreperfusion. All patients underwent cardiac magnetic resonance imaging 1 week and 6 months after STEMI. Baseline characteristics were compared across median values of PCSK9. The association between PCSK9 levels and LVEF at 6 months was evaluated by analysis of covariance. The mean age of the sample was 60 ± 12 years and 33 (82.5%) were male patients. The infarct location was anterior in 27 patients (67.5%), and 9 patients (22.5%) were Killip class ≥ II. The mean 1 week and 6 month LVEF were 41 ± 7% and 48 ± 10%, respectively. The mean PCSK9 was 1.93 ± 0.38 U/mL. Testing the association between serum PCSK9 and 6 month LVEF with analysis of covariance revealed an inverse relationship (r = -0.35, P = 0.028). After multivariate adjustment, circulating PCSK9 remained significant and inversely associated with 6 month LVEF (P = 0.002). CONCLUSIONS: In patients with a first STEMI with reduced ejection fraction at index admission and treated with primary percutaneous coronary intervention, circulating PCSK9 was associated with lower LVEF at 6 months.

ST2 and left ventricular remodeling after ST-segment elevation myocardial infarction: A cardiac magnetic resonance study.

BACKGROUND: The association of soluble interleukin-1 receptor-like 1 (ST2) with left ventricular (LV) remodeling is unclear in patients with a first ST-segment elevation myocardial infarction (STEMI). The objective of this work was to assess the relationship between ST2, a marker of inflammation, and cardiac magnetic resonance (CMR) imaging-derived LV remodeling after a first STEMI. METHODS: We prospectively evaluated 109 patients with a first STEMI treated with primary percutaneous coronary intervention who had ST2 assessed 24 h post-reperfusion. All patients underwent CMR imaging 1 week and 6 months after STEMI. The independent associations between ST2, LV diastolic and systolic volume indices, and LV ejection fraction (LVEF) were evaluated by linear mixed models. RESULTS: The mean age of the sample was 59 ±â€¯12 years, 85 patients (78%) were male, and 13 (11.9%) had a LVEF ≤40%. The median (IQR) of ST2 was 55.3 (38.7-94.1) pg/mL. At 1-week CMR higher ST2 was related to more infarct size and less myocardial salvage index (p < 0.01). Overall, after comprehensive multivariable adjustment, higher baseline ST2 was associated with progressive LV volume indices dilation and LVEF deterioration (p < 0.05). This effect was stronger in patients with severe 1-week structural damage, namely those with large infarct size, extensive microvascular obstruction or LVEF ≤40%. CONCLUSIONS: In patients with a first STEMI treated with primary percutaneous coronary intervention, soluble ST2 predicts dynamic changes in CMR-derived LV volumes and LVEF. Future studies must assess whether targeting interleukin-1 leads to lower ST2 levels and less LV remodeling.

Left ventricular ejection fraction recovery in patients with heart failure treated with intravenous iron: a pilot study.

AIMS: In patients with heart failure with reduced ejection fraction (HFrEF) and iron deficiency, treatment with intravenous iron has shown a clinical improvement regardless of anaemic status. Cardiac magnetic resonance (CMR) T2* sequence has shown a potential utility for evaluating myocardial iron deficiency. We aimed to evaluate whether T2* sequence significantly changes after ferric carboximaltose (FCM) administration, and if such changes correlate with changes in left ventricle ejection fraction (LVEF). METHODS AND RESULTS: In this pilot study, we included eight patients with chronic symptomatic (New York Heart Association II-III) HFrEF and iron deficiency. A CMR, including T2* analysis, was performed before and at a median of 43 days (interquartile range = 35-48) after intravenous FCM administration. Pearson or Spearman correlation coefficient (r) was used for bivariate contrast as appropriate. A partial correlation analysis was performed between ΔLVEF and ΔT2* while controlling for anaemia status at baseline. Anaemia was present in half of patients. After FCM administration, T2* decreased from a median of 39.5 (35.9-48) to 32 ms (32-34.5), P = 0.012. Simultaneously, a borderline increase in median of LVEF [40% (36-44.5) to 48.5% (38.5-53), P = 0.091] was registered. In a bivariate correlational analysis, ΔT2* was highly correlated with ΔLVEF (r = -0.747, P = 0.033). After controlling for anaemia at baseline, the association between ΔT2* and ΔLVEF persisted [r(partial): -0.865, R2(partial): 0.748, P = 0.012]. A median regression analysis backed-up these findings. CONCLUSIONS: In a small sample of patients with HFrEF and iron deficiency, myocardial iron repletion assessed by CMR was associated to left ventricular remodelling. Further studies are warranted.

White blood cell subtypes after STEMI: temporal evolution, association with cardiovascular magnetic resonance--derived infarct size and impact on outcome.

The evolution of white blood cells after ST elevation myocardial infarction (STEMI) and their association with infarct size and major adverse cardiac events (MACE) remains unclear. Two hundred eleven patients underwent CMR after STEMI. Infarct mass (grams) was determined. Neutrophil, lymphocyte, and monocyte counts (×1,000 cells/ml) were measured upon arrival and at 12, 24, 48, 72, and 96 h. Patients with large infarctions (3rd tertile ≥ 28.5 g vs. 1st and 2nd tertiles < 28.5 g) showed a larger neutrophil count at 12 h (14.8 ± 4.8 vs. 11.4 ± 3.3, p < 0.0001) and an increased monocyte count (maximum at 24 h (0.65[0.50-0.91] vs. 0.55[0.42-0.71], p = 0.004)) but no difference in lymphocyte count. Neutrophil count at 12 h independently predicted large infarctions (OR 1.14, 95%CI [1.04-1.26], p = 0.008). During follow-up (median 504 days), 25 MACE occurred. Neutrophil count at 96 h independently predicted MACE (HR 1.2, 95%CI [1.1-1.4], p = 0.003). Large infarctions show a marked neutrophil peak and an increasing monocyte count. Neutrophil count independently predicts large infarctions and MACE.

Atrial stunning as predictor of early relapse into atrial fibrillation after cardioversion.

Although the high rate of success after cardioversion, less than 50% of patients maintain sinus rhythm for the first year. In view for the high percentage of relapse into atrial fibrillation, it is interesting to analyze the relationship between atrial stunning after cardioversion and relapse into atrial fibrillation. Thus, we evaluated 101 patients with atrial fibrillation and successful cardioversion. Atrial mechanical function was assessed by measures of transmitral peak A wave velocity, determined before and weekly after cardioversion during 1 month. Fifty-five percent of patient relapse into atrial fibrillation during follow-up. No significant differences were found in clinical and echocardiographic variables between the group with and without relapse. However, the group of patients who relapsed into atrial fibrillation showed a lower peak A wave velocity immediately after cardioversion than patients who maintain in sinus rhythm at month (0.44+/-0.27 vs. 0.60+/-0.38 m/s p<0.01). Impaired atrial function improves during the first 14 days after cardioversion.