MyI-Net: Fully Automatic Detection and Quantification of Myocardial Infarction from Cardiovascular MRI Images.

Myocardial infarction (MI) occurs when an artery supplying blood to the heart is abruptly occluded. The "gold standard" method for imaging MI is cardiovascular magnetic resonance imaging (MRI) with intravenously administered gadolinium-based contrast (with damaged areas apparent as late gadolinium enhancement [LGE]). However, no "gold standard" fully automated method for the quantification of MI exists. In this work, we propose an end-to-end fully automatic system (MyI-Net) for the detection and quantification of MI in MRI images. It has the potential to reduce uncertainty due to technical variability across labs and the inherent problems of data and labels. Our system consists of four processing stages designed to maintain the flow of information across scales. First, features from raw MRI images are generated using feature extractors built on ResNet and MoblieNet architectures. This is followed by atrous spatial pyramid pooling (ASPP) to produce spatial information at different scales to preserve more image context. High-level features from ASPP and initial low-level features are concatenated at the third stage and then passed to the fourth stage where spatial information is recovered via up-sampling to produce final image segmentation output into: (i) background, (ii) heart muscle, (iii) blood and (iv) LGE areas. Our experiments show that the model named MI-ResNet50-AC provides the best global accuracy (97.38%), mean accuracy (86.01%), weighted intersection over union (IoU) of 96.47%, and bfscore of 64.46% for the global segmentation. However, in detecting only LGE tissue, a smaller model, MI-ResNet18-AC, exhibited higher accuracy (74.41%) than MI-ResNet50-AC (64.29%). New models were compared with state-of-the-art models and manual quantification. Our models demonstrated favorable performance in global segmentation and LGE detection relative to the state-of-the-art, including a four-fold better performance in matching LGE pixels to contours produced by clinicians.

Short-term adverse remodeling progression in asymptomatic aortic stenosis.

OBJECTIVES: Aortic stenosis (AS) is characterised by a long and variable asymptomatic course. Our objective was to use cardiovascular magnetic resonance imaging (MRI) to assess progression of adverse remodeling in asymptomatic AS. METHODS: Participants from the PRIMID-AS study, a prospective, multi-centre observational study of asymptomatic patients with moderate to severe AS, who remained asymptomatic at 12 months, were invited to undergo a repeat cardiac MRI. RESULTS: Forty-three participants with moderate-severe AS (mean age 64.4 ± 14.8 years, 83.4% male, aortic valve area index 0.54 ± 0.15 cm2/m2) were included. There was small but significant increase in indexed left ventricular (LV) (90.7 ± 22.0 to 94.5 ± 23.1 ml/m2, p = 0.007) and left atrial volumes (52.9 ± 11.3 to 58.6 ± 13.6 ml/m2, p < 0.001), with a decrease in systolic (LV ejection fraction 57.9 ± 4.6 to 55.6 ± 4.1%, p = 0.001) and diastolic (longitudinal diastolic strain rate 1.06 ± 0.2 to 0.99 ± 0.2 1/s, p = 0.026) function, but no overall change in LV mass or mass/volume. Late gadolinium enhancement increased (2.02 to 4.26 g, p < 0.001) but markers of diffuse interstitial fibrosis did not change significantly (extracellular volume index 12.9 [11.4, 17.0] ml/m2 to 13.3 [11.1, 15.1] ml/m2, p = 0.689). There was also a significant increase in the levels of NT-proBNP (43.6 [13.45, 137.08] pg/ml to 53.4 [19.14, 202.20] pg/ml, p = 0.001). CONCLUSIONS: There is progression in cardiac remodeling with increasing scar burden even in asymptomatic AS. Given the lack of reversibility of LGE post-AVR and its association with long-term mortality post-AVR, this suggests the potential need for earlier intervention, before the accumulation of LGE, to improve the long-term outcomes in AS. KEY POINTS: • Current guidelines recommend waiting until symptom onset before valve replacement in severe AS. • MRI showed clear progression in cardiac remodeling over 12 months in asymptomatic patients with AS, with near doubling in LGE. • This highlights the need for potentially earlier intervention or better risk stratification in AS.

Chronic infarct size after spontaneous coronary artery dissection: implications for pathophysiology and clinical management.

AIMS: To report the extent and distribution of myocardial injury and its impact on left ventricular systolic function with cardiac magnetic resonance imaging (CMR) following spontaneous coronary artery dissection (SCAD) and to investigate predictors of myocardial injury. METHODS AND RESULTS: One hundred and fifty-eight angiographically confirmed SCAD-survivors (98% female) were phenotyped by CMR and compared in a case-control study with 59 (97% female) healthy controls (44.5 ± 8.4 vs. 45.0 ± 9.1 years). Spontaneous coronary artery dissection presentation was with non-ST-elevation myocardial infarction in 95 (60.3%), ST-elevation myocardial infarction (STEMI) in 52 (32.7%), and cardiac arrest in 11 (6.9%). Left ventricular function in SCAD-survivors was generally well preserved with small reductions in ejection fraction (57 ± 7.2% vs. 60 ± 4.9%, P < 0.01) and increases in left ventricular dimensions (end-diastolic volume: 85 ± 14 mL/m2 vs. 80 ± 11 mL/m2, P < 0.05; end-systolic volume: 37 ± 11 mL/m2 vs. 32 ± 7 mL/m2, P <0.01) compared to healthy controls. Infarcts were small with few large infarcts (median 4.06%; range 0-30.9%) and 39% having no detectable late gadolinium enhancement (LGE). Female SCAD patients presenting with STEMI had similar sized infarcts to female Type-1 STEMI patients age <75 years. Multivariate modelling demonstrated STEMI at presentation, initial TIMI 0/1 flow, multivessel SCAD, and a Beighton score >4 were associated with larger infarcts [>10% left ventricular (LV) mass]. CONCLUSION: The majority of patients presenting with SCAD have no or small infarctions and preserved ejection fraction. Patients presenting with STEMI, TIMI 0/1 flow, multivessel SCAD and those with features of connective tissue disorders are more likely to have larger infarcts.

Haemodynamic effects of pharmacologic stress with adenosine in patients with left ventricular systolic dysfunction.

BACKGROUND: In patients with heart failure, downregulation of adenosine receptor gene expression and impaired adenosine-related signal transduction may result in a diminished response to adenosine. This may have implications for cardiac stress testing. We evaluated the haemodynamic response to intravenous adenosine in patients with left ventricular systolic dysfunction (LVSD) undergoing stress cardiovascular magnetic resonance imaging (CMR). METHODS AND RESULTS: We retrospectively examined 497 consecutive patients referred for clinical stress CMR. Blood pressure and heart rate responses with intravenous adenosine were compared in patients with normal, mild-moderately impaired and severely impaired LV systolic function (ejection fraction [EF] > 55%, 36-55% and < 35%, respectively). Following 2 min of adenosine infusion, there was a significant difference between the groups in the heart rate change from baseline, with a diminished heart rate response in patients with LVSD (p < 0.001). An increase in the dose of adenosine (up to 210 µg/kg/min) was required to achieve a sufficient haemodynamic response in more patients with severe LVSD (41%) than those with mild-moderately impaired and normal LV systolic function (24% and 19%, respectively, p < 0.001). Even with increased doses of adenosine in subjects with severe LVSD, peak haemodynamic response remained blunted. With multivariate analysis age (p < 0.001) and LVEF (p = 0.031) were independent predictors of heart rate response to adenosine. CONCLUSION: Patients with reduced LVEF referred for stress CMR may have a blunted heart rate response to adenosine. Further study is warranted to determine whether this may be associated with reduced diagnostic accuracy and also the potential utility of further dose increases or alternative stressors.

Unusual Case of Three Total Occlusions.

This case highlights the value of extensive coronary collaterals in maintaining myocardial viability in severe coronary artery disease, and the role of cardiac MRI in guiding revascularization decisions.