Cardiac Magnetic Resonance Imaging Applications in Chest Pain

Version	Summary	Created by	Modification	Content Size	Created at	Operation
1		Roohallah Alizadehsani	--	2846	2023-09-06 07:28:00	\|
2	format	Jason Zhu	Meta information modification	2846	2023-09-11 08:04:05	\|

This entry is adapted from the peer-reviewed paper 10.3390/diagnostics13162667

CMRI is the exclusive imaging technique capable of identifying myocardial edema, endomyocardial fibrosis, pericarditis accompanied by pericardial effusions, and apical thrombi within either the left or right ventricle.

cardiac magnetic resonance imaging emergency department

1. Introduction

Chest pain is a commonly encountered ailment in the ED. The prompt and precise differentiation between grade 1 acute myocardial infarction (AMI) and other causes of myocardial injury is crucial for individuals with abnormal troponin levels ^[1]. The notable capability of high-sensitivity troponin tests to predict the absence of AMI highlights the significance of diagnostic methods in identifying the root cause of AMI ^[2]. CMRI is a valuable technique for stratifying patients in the ED who experience chest discomfort. Unlike invasive coronary angiography, CMRI is a non-invasive procedure that is less expensive and results in a shorter hospitalization duration. Additionally, CMRI provides significant information about the structure, function, tissue edema, and the location and nature of tissue damage in the heart, all of which can assist in determining various etiologies of cardiac injury. CMRI may assist in discriminating between chest symptoms caused by type 1 AMI and supply–demand imbalances caused by acute cardiac non-coronary artery disease. After conducting a comprehensive research review, it was determined that utilizing CMRI with stress monitoring is a safe approach for patients who arrive at the ED with chest pain, regardless of the varying troponin levels ^[3]. Combined positron emission tomography (PET) and CMRI integrates the advances in functional and high-quality anatomical imaging from these systems for many clinical and scientific studies, including a wide use in cardiology. In this combined scanner, motion fields of the heart organ are estimated from concurrently acquired tagged magnetic resonance images ^[4]. CMRI is a helpful, secure, cost-efficient, and effective alternative to conventional diagnostic techniques utilized in this patient group. It is a helpful technique for risk stratifying individuals with suspected heart pathology and confirming diagnoses without the need for invasive testing ^[5].

CMRI can be used in a wide range of clinical situations. Cardiac cine imaging may be employed to measure global and regional operations using a variety of approaches. An example of this is T2-weighted imaging, which can evaluate stress-induced myocardial edema, rest perfusion, and the presence of myocardial ischemia following the administration of a vasodilator-like adenosine, dipyridamole, or regadenoson ^[6]. The utilization of delayed-enhancement CMRI allows for the identification and characterization of myocardial necrosis, inflammation, and infiltrative diseases by assessing their presence and pattern ^[7]^[8]^[9]^[10]. Techniques for tissue description can be utilized as appropriate, including tomographic images for chest morphology, velocity flow projection for valve appraisal, and tissue characterization techniques (such as T1 parametric projection for infiltrative disease and T2* imaging for iron overburden) ^[11]. CMRI plays a crucial role as a diagnostic tool in patients diagnosed with myocardial infarction with non-obstructive coronary arteries (MINOCA). CMRI is a critical diagnostic tool for assessment of the existence of regional wall motion anomalies, tissue edema, and subendocardial late gadolinium enhancement (LGE) in a typical coronary distribution area ^[12]^[13]^[14]. In a patient with acute chest discomfort and high-sensitivity troponin, the diagnosis of AMI could serve as a powerful tool in chest pain patients in the ED and improve the effectiveness of patient management. MINOCA can manifest in 5% to 6% of cases, even when all the universally recognized criteria for a type 1 AMI are satisfied ^[15]. Thrombosis, superimposed thromboembolism, vasospasm, plaque disruption, takotsubo cardiomyopathy, myocarditis or a composite of such occurrences are all possible causes of MINOCA ^[16].

Nevertheless, since the coronary arteries may seem ordinary or almost standard around the period of angiography, the ischemia system of myocardial necrosis cannot be correctly diagnosed, and consequently, essential secondary preventive treatments (as an example, statins and antiplatelet drugs) may not be implemented. Whilst CMRI cannot distinguish between different causes of ischemic damage in patients with MINOCA (such as spasm or emboli), it is capable of detecting ischemia and distinguishing it from myocarditis, takotsubo cardiomyopathy, or other causes of acute cardiac noncoronary artery diseases (CNCDs) ^[17]. Relying solely on the absence of epicardial stenosis of the internal carotid artery is insufficient to determine the underlying cause of damage in a patient with abnormal troponin levels. Additional information from CMRI is necessary to make a conclusive assessment. To differentiate between actual ischemic injury (referred to as True-MINOCA) and myocardial injury unrelated to ischemia caused by acute CNCDs (referred to as False-MINOCA), CMRI is employed ^[5].

While computed tomography (CT) is commonly favored in emergency departments due to its speed and versatility, the unique advantages of CMRI should not be overlooked. CMRI provides non-invasive imaging, cost-effectiveness, and valuable information about the cardiac structure and tissue damage, aiding in accurate diagnosis and patient management. By incorporating machine learning and combining CMRI with stress monitoring or other imaging techniques, like PET, clinicians can enhance risk stratification and confirm diagnoses without invasive procedures. Considering these benefits, CMRI plays a crucial role in evaluating cardiac conditions in emergency departments ^[5].

2. Application of CMRI in the Diagnosis of Different Types of Chest Pains

CMRI provides a wide range of different diagnoses unrelated to acute coronary syndrome (ACS) that can effectively explain the symptoms, and it also includes incidental findings. The identification of these unforeseen findings could have implications for patient management, leading to the establishment of new diagnoses or the need for further investigations. Between 2011 and 2015, adult patients with suspected ACS who visited an academic ED showed no indications of ischemia on initial electrocardiogram (ECG), had a minimum of one negative cardiac biomarker, and then had CMRI as a component of their diagnostic assessment were prospectively recruited. This finding suggests that CMRI can be used to diagnose symptomatic coronary artery disease (CAD) and potentially non-CAD severe cardiac abnormalities. These considerations may influence its usage in ACS workups in the emergency department ^[18]. In the investigation of the prognostic and diagnostic utility of CMRI in the diagnosis of ischemic heart disease (IHD), researchers have explored the current improvements, limitations, and future directions. For example, Fagiry et al. examined these aspects to enhance the effectiveness of CMRI in clinical practice.

An inherent problem associated with cardiac catheterization and CT coronary angiography is the considerable radiation exposure endured by the patient during the procedure. Consequently, employing CMRI technology to solve such problems is highly useful to patients ^[19]. Jalnapurkar et al. ^[20] conducted a study to examine the diagnostic importance of stress CMRI in women presenting with suspected ischemia. The study focused on 113 female patients who underwent stress CMRI, encompassing anatomic, functional, adenosine stress perfusion, and delayed-enhancement photography. Prior to this, these patients had undergone assessment for indications and manifestations of ischemia; however, there was no indication of obstructive CAD detected. From 113 patients, 65 were diagnosed with coronary microvascular dysfunction (CMD) on the basis of subendocardial perfusion abnormalities consistent with myocardial ischemia on stress CMRI, 10 with CAD, 2 with left ventricular (LV) hypertrophy, and 3 patients were diagnosed with congenital coronary anomalies or cardiomyopathy that had not been detected in prior cardiac evaluations. The rest (33 patients) were normal. These findings indicate that stress CMRI often reveals abnormalities and offers diagnostic value in identifying CMD in women who display symptoms and indications of ischemia but do not show any signs of obstructive CAD. Stress CMRI appears helpful for diagnostic assessment in these diagnostically challenging people.

To detect LGE patterns via cardiac MRI in high-risk patients with right ventricular dysfunction following the placement of a left ventricular assist device (LVAD), Simkowski et al. ^[21] proposed an unsupervised machine learning (ML) method. They utilized the 17-segment model to extract LGE patterns from CMRI scans performed on patients who had received an LVAD at a medical facility within a 12-month timeframe. Employing an unsupervised ML technique for hierarchical agglomerative clustering, the patients were subsequently classified based on similarities in the LGE patterns. The clusters which resulted from this were then statistically compared. Based on the findings, the application of unsupervised ML to analyze the LGE patterns observed on CMRI has the capability to identify groups of patients who are prone to developing right ventricular failure (RVF). Patients diagnosed with non-ischemic and mixed etiologies of heart failure may face an increased risk of developing RVF compared to those with purely ischemic causes. This heightened risk can be attributed to the extensive involvement of biventricular myocardium indicated by the observed LGE patterns on CMRI. Alsunbuli ^[22] evaluated several imaging modalities by using their inherent features to benchmark against a simulated ideal test, utilizing a qualitative approach to the comparison, as well as the various societies’ guidelines. According to the findings, CMRI poses no danger of radiation exposure but provides lesser resolution than CT. It requires more time from physicians and patients and hence is more demanding. It requires fewer operators than echocardiography and enables the identification of small changes in serial follow-up evaluations, particularly for the LV volume and function. CMRI can also be used during pregnancy. In terms of the drawbacks, it cannot be used intra-procedurally and is contraindicated in the presence of certain pacemakers. CMRI is also susceptible to artifacts that may be detected using a chest X-ray, such as a retrocardiac surgical needle in one instance.

Grober et al. ^[23] conducted a comparison between diffusion-weighted MRI (DMRI) and conventional MRI techniques to detect microadenomas in patients with Cushing’s disease. They further evaluated the efficacy of a 3D volumetric interpolated breath-hold examination, a 3D T1 sequence known as a spoiled gradient echo (SGE), which offers enhanced soft-tissue contrast and improved resolution. SGE has better sensitivity for identifying and localizing pituitary microadenomas than DMRI. However, DMRI is rarely used to diagnose adenoma. SGE should be included in the routine MRI procedure for Cushing’s disease patients. Moonen et al. ^[24] used CMRI to determine the frequency of Fabry disease in a group of individuals with inexplicable LGE. Fabry disease is a rare X-linked genetic condition with cardiac symptoms such as LVH, contractile failure, and fibrosis, which can be seen as LGE of the myocardium on CMRI. Fabry disease is a critical diagnosis to establish, since the missing enzyme can be replaced for the rest of one’s life. In terms of detecting and pinpointing pituitary microadenomas, the SGE technique demonstrates greater sensitivity compared to DMRI. It is uncommon for an adenoma to be solely detected using DMRI. Therefore, it is recommended to include SGE as a standard component of the MRI protocol for patients diagnosed with Cushing’s disease. In a study conducted by Moonen et al. ^[24], the presence of Fabry disease was examined in a cohort of patients displaying unexplained LGE on CMRI. Diagnosing Fabry disease holds significant importance due to the availability of lifelong enzyme replacement therapy as a treatment option for the deficient enzyme. According to the findings, the presence of unexplained LGE on CMRI could potentially indicate the presence of late-onset Fabry disease.

3. Utilizing Machine Learning for the Diagnosis of Chest Pain through CMRI

Artificial intelligence (AI) and machine learning (ML) are quickly gaining traction in medicine ^[19]^[25]. In the coming years, they are anticipated to profoundly change clinical practice, notably in the field of medical imaging ^[4]^[26]. AI is a broad term that refers to using robots to do activities that are common to human intellect, such as inferring conclusions through deduction or induction. On the other hand, ML is a more limited kind of computer processing that learns how to generate predictions using a mathematical model and training data. By being subjected to more instances, ML learns parameters from examples and can perform better at tasks like identifying and distinguishing patterns in data. The most sophisticated ML techniques, also known as DL, are particularly well-suited for this task. DL segmentation methods have recently been proven to outperform classic methods such as cardiac atlases, level set, statistical models, deformable models, and graph cuts. Nevertheless, a recent study of a number of automated techniques revealed that in more than 80% of CMRIs, even the top performing algorithms produced anatomically implausible segmentations ^[27]. When specialists perform segmentation, such mistakes do not occur. To gain acceptability in clinical practice, the automated methods’ flaws must be addressed through continued research. This can be accomplished by producing more accurate segmentation results or developing techniques that automatically detect segmentation errors.

By combining automated segmentation and evaluating segmentation uncertainty, Sander et al. employed CMRI to identify regions in the images where local segmentation failures occur. They utilized a convolutional neural networks (CNNs) uncertainty to discover local segmentation problems that may require expert repair. To compare the performance of manual and (corrected) automatic segmentation, the Dice coefficient, 3D Hausdorff distance, and clinical markers were utilized. The findings suggest that combining automated segmentation with manual correction of identified segmentation errors results in enhanced segmentation accuracy and a significant 10-fold decrease in the time required by experts for segmentation compared to manual segmentation alone, as demonstrated in the studies ^[28]. During segmentation training, Oktay et al. ^[29] devised an auto-encoder-based anatomically restricted neural network (NN) that utilizes constraints to make inferences about limitations. In a study, Duan et al. ^[30] incorporated atlas propagation to explicitly enforce shape refinement in a DL-based segmentation approach for CMRIs. This was extremely convenient when there were photo capture artifacts present. By employing cardiac anatomical metrics, Painchaud et al. ^[31] devised a post-processing technique to identify and transform anatomically questionable heart segmentations into accurate ones. Employing an ML-based method, Park et al. ^[32] predicted AMI. The occlusion of coronary arteries is responsible for the occurrence of AMI, and prompt revascularization is necessary to improve the prognosis. However, AMI has been misdiagnosed as other illnesses, and reperfusion delay has been linked to a poor outcome in patients. The authors used ML algorithms to anticipate AMI in patients with acute chest discomfort based on data collected at admittance.

Controlling the quality of cardiovascular images is so critical. Varied manufacturers, various machine types, and different MCE scanning parameters all influence ML. At the moment, the ML system based on DL suffers from a lack of explainability. After much training, an ML model could identify myocardial fibrosis based on a picture. However, it might not explain what effective characteristics it learned to reach such a conclusion. As a result, explainability is a vital study topic concerning medical ML ^[33]^[34]. To ensure high accuracy and optimize the algorithm, it is crucial to utilize a substantial amount of high-quality data during the initial learning phase of ML, leveraging its inherent capabilities. Furthermore, the acquisition cost and time required to cardiovascular imaging, particularly MCE data from CMRI, are significant. The critical task at hand is to establish a model that can effectively learn the best optimal solution even when provided with limited samples. Through the utilization of migration learning, it becomes feasible to transfer valuable information from prior ML models to novel models, resulting in a reduction in the required data resources for DL ^[35].

5. Implementation Results

Researchers offer a tiny example of implementation in this section of the study to provide a clearer overview of CMRI use in the ED. Late enhancement imaging tests were performed 15 min following gadolinium–DTPA injection utilizing a 3D-gradient faulty turbo fast-field echo (FFE) sequence that includes an individually designed 180° inversion pre-pulse (Look-Locker) to provide appropriate myocardium suppression ^[36]. A series of images were acquired using a 2D-sequence approach, which included short-axis images with a 5 mm slice thickness encompassing the entire left ventricle, along with two to three long-axis views. The presence of dark patches within the enlarged myocardium supplied by the infarct artery indicated the presence of persistent microvascular obstruction. Various patterns of late enhancement, including subendocardial, transmural, intramural, subepicardial, and diffuse patterns, were detected. To visualize cardiac edema, a T2-weighted turbo spin-echo sequence was employed, along with a fat saturation pulse. Images were acquired in a continuous short-axis orientation, covering the entire left ventricle, with a slice thickness of 15 mm. Myocardial edema was defined as a relative myocardial EPCs intensity exceeding 2.0 times that of skeletal muscle.

Coronary artery disease leads to myocardial damage, which can be identified through subendocardial or transmural late enhancement patterns. In contrast, acute myocarditis is often associated with the presence of late enhancement, as characterized by a diffuse, intramural, or subepicardial pattern. Patients with ST-segment elevation myocardial infarction (STEMI) had the greatest levels of creatine kinase (CK), troponin-I, and leukocytes. They gradually dropped from individuals with non–ST-segment elevation myocardial infarction (NSTEMI), acute myocarditis, and takotsubo cardiomyopathy to takotsubo cardiomyopathy. In terms of the C-reactive protein (CRP) levels, patients with acute myocarditis exhibited the highest initial and peak values. There were statistically significant differences in the levels of CK, troponin-I, and the first CRP among the different groups.

The volumes and ejection percentages of the ventricles were considerably different. Acute myocarditis patients had the greatest LV volumes. The LV ejection fraction of STEMI patients was considerably lower than that of NSTEMI patients (p = 0.006). Acute myocarditis patients had substantially greater RV volumes than other categories (p = 0.03). In the group of patients experiencing their first episode of severe chest pain, wall motion abnormalities were detected in all 95/95 (100%) cases of STEMI, 51/68 (75%) cases of NSTEMI, 18/27 (66.7%) cases of acute myocarditis, and 12/12 (100%) cases of takotsubo cardiomyopathy. The observed differences were statistically significant (p < 0.001). A random distribution of wall motion anomalies was seen in individuals with acute myocarditis. The aberrant wall motion in individuals with takotsubo cardiomyopathy was concentrated in the midventricular–apical regions.

References

Nowak, R.M.; Jacobsen, G.; Christenson, R.H.; Moyer, M.; Hudson, M.; McCord, J. Differentiating type 1 and 2 acute myocardial infarctions using the N-terminal pro B-type natriuretic peptide/cardiac troponin T ratio. Am. J. Emerg. Med. 2018, 36, 1849–1854.
Vasile, V.C.; Jaffe, A.S. High-Sensitivity Cardiac Troponin for the Diagnosis of Patients with Acute Coronary Syndromes. Curr. Cardiol. Rep. 2017, 19, 92.
DeFilippis, A.P.; Chapman, A.R.; Mills, N.L.; de Lemos, J.A.; Arbab-Zadeh, A.; Newby, L.K.; Morrow, D.A. Assessment and Treatment of Patients With Type 2 Myocardial Infarction and Acute Nonischemic Myocardial Injury. Circulation 2019, 140, 1661–1678.
Miller, D.D.; Brown, E.W. Artificial Intelligence in Medical Practice: The Question to the Answer? Am. J. Med. 2018, 131, 129–133.
Cavalier, J.S.; Klem, I. Using Cardiac Magnetic Resonance Imaging to Evaluate Patients with Chest Pain in the Emergency Department. J. Cardiovasc. Imaging 2021, 29, 91.
Nickander, J.; Themudo, R.; Thalén, S.; Sigfridsson, A.; Xue, H.; Kellman, P.; Ugander, M. The relative contributions of myocardial perfusion, blood volume and extracellular volume to native T1 and native T2 at rest and during adenosine stress in normal physiology. J. Cardiovasc. Magn. Reson. 2019, 21, 73.
Sirajuddin, A.; Mirmomen, S.M.; Kligerman, S.J.; Groves, D.W.; Burke, A.P.; Kureshi, F.; White, C.S.; Arai, A.E. Ischemic Heart Disease: Noninvasive Imaging Techniques and Findings. RadioGraphics 2021, 41, E990–E1021.
De Carvalho, F.S.; Mukai, K.; Clayton, J.; Ordovas, K. Cardiac MRI: A preferred method for assessing myocardial ischemia and infarct burden. Appl. Radiol. 2017, 46, 21–29.
Burrage, M.K.; Shanmuganathan, M.; Masi, A.; Hann, E.; Zhang, Q.; Popescu, I.A.; Soundarajan, R.; Pelado, J.L.; Chow, K.; Neubauer, S. Cardiovascular magnetic resonance stress and rest T1-mapping using regadenoson for detection of ischemic heart disease compared to healthy controls. Int. J. Cardiol. 2021, 333, 239–245.
Bajaj, N.S.; Singh, S.; Farag, A.; EL-Hajj, S.; Heo, J.; Iskandrian, A.E.; Hage, F.G. The prognostic value of non-perfusion variables obtained during vasodilator stress myocardial perfusion imaging. J. Nucl. Cardiol. 2016, 23, 390–413.
Ferreira, V.M.; Piechnik, S.K. CMR parametric mapping as a tool for myocardial tissue characterization. Korean Circ. J. 2020, 50, 658–676.
Montone, R.A.; Jang, I.-K.; Beltrame, J.F.; Sicari, R.; Meucci, M.C.; Bode, M.; Gaibazzi, N.; Niccoli, G.; Bucciarelli-Ducci, C.; Crea, F. The evolving role of cardiac imaging in patients with myocardial infarction and non-obstructive coronary arteries. Prog. Cardiovasc. Dis. 2021, 68, 78–87.
Mark, P.B.; Johnston, N.; Groenning, B.A.; Foster, J.E.; Blyth, K.G.; Martin, T.N.; Steedman, T.; Dargie, H.J.; Jardine, A.G. Redefinition of uremic cardiomyopathy by contrast-enhanced cardiac magnetic resonance imaging. Kidney Int. 2006, 69, 1839–1845.
Satoh, H. Distribution of late gadolinium enhancement in various types of cardiomyopathies: Significance in differential diagnosis, clinical features and prognosis. World J. Cardiol. 2014, 6, 585.
He, C.; Zhu, C.; Han, B.; Hu, H.; Wang, S.; Zhai, C.; Hu, H. Association between anxiety and clinical outcomes in Chinese patients with myocardial infarction in the absence of obstructive coronary artery disease. Clin. Cardiol. 2020, 43, 659–665.
Agewall, S.; Beltrame, J.F.; Reynolds, H.R.; Niessner, A.; Rosano, G.; Caforio, A.L.P.; De Caterina, R.; Zimarino, M.; Roffi, M.; Kjeldsen, K.; et al. ESC working group position paper on myocardial infarction with non-obstructive coronary arteries. Eur. Heart J. 2016, 38, ehw149.
Ciliberti, G.; Coiro, S.; Tritto, I.; Benedetti, M.; Guerra, F.; Del Pinto, M.; Finocchiaro, G.; Cavallini, C.; Capucci, A.; Kaski, J.C.; et al. Predictors of poor clinical outcomes in patients with acute myocardial infarction and non-obstructed coronary arteries (MINOCA). Int. J. Cardiol. 2018, 267, 41–45.
Ziegler, C.E.; Painter, D.M.; Borawski, J.B.; Kim, R.J.; Kim, H.W.; Limkakeng, A.T. Unexpected Cardiac MRI Findings in Patients Presenting to the Emergency Department for Possible Acute Coronary Syndrome. Crit. Pathw. Cardiol. A J. Evid.-Based Med. 2018, 17, 167–171.
Wong, N.D. Epidemiological studies of CHD and the evolution of preventive cardiology. Nat. Rev. Cardiol. 2014, 11, 276–289.
Jalnapurkar, S.; Zarrini, P.; Mehta, P.K.; Thomson, L.E.J.; Agarwal, M.; Samuels, B.A.; Shufelt, C.L.; Eastwood, J.-A.; Berman, D.; Merz, N.B.; et al. Role of Stress Cardiac Magnetic Resonance Imaging in Women With Suspected Ischemia But No Obstructive Coronary Artery Disease. J. Radiol. Nurs. 2017, 36, 180–183.
Simkowski, J.M.; Wehbe, R.M.; Goergen, J.; Anderson, A.S.; Ghafourian, K.; Okwuosa, I.; Vorovich, E.E.; Ahmad, F.S.; Tibrewala, A.; Pham, D.T.; et al. Unsupervised Machine Learning of LGE Patterns on Cardiac MRI Identifies Patients at Risk for Right Ventricular Failure After LVAD. J. Card. Fail. 2020, 26, S146.
Alsunbuli, A. The use of cardiac magnetic resonance imaging (CMRI) for adult congenital heart disease patients: Qualitative comparative review. Clin. Med. 2020, 20, s6–s7.
Grober, Y.; Grober, H.; Wintermark, M.; Jane, J.A.; Oldfield, E.H. Comparison of MRI techniques for detecting microadenomas in Cushing’s disease. J. Neurosurg. 2018, 128, 1051–1057.
Moonen, A.; Lal, S.; Ingles, J.; Yeates, L.; Semsarian, C.; Puranik, R. Prevalence of Anderson-Fabry disease in a cohort with unexplained late gadolinium enhancement on cardiac MRI. Int. J. Cardiol. 2020, 304, 122–124.
Fagiry, M.A.; Abdelaziz, I.; Davidson, R.; Mahmoud, M.Z. The recent advances, drawbacks, and the future directions of CMRI in the diagnosis of IHD. Sci. Rep. 2021, 11, 14958.
Chan, B. The rise of artificial intelligence and the crisis of moral passivity. AI Soc. 2020, 35, 991–993.
Bernard, O.; Lalande, A.; Zotti, C.; Cervenansky, F.; Yang, X.; Heng, P.-A.; Cetin, I.; Lekadir, K.; Camara, O.; Gonzalez Ballester, M.A.; et al. Deep Learning Techniques for Automatic MRI Cardiac Multi-Structures Segmentation and Diagnosis: Is the Problem Solved? IEEE Trans. Med. Imaging 2018, 37, 2514–2525.
Sander, J.; de Vos, B.D.; Išgum, I. Automatic segmentation with detection of local segmentation failures in cardiac MRI. Sci. Rep. 2020, 10, 21769.
Oktay, O.; Ferrante, E.; Kamnitsas, K.; Heinrich, M.; Bai, W.; Caballero, J.; Cook, S.A.; de Marvao, A.; Dawes, T.; O’Regan, D.P.; et al. Anatomically Constrained Neural Networks (ACNNs): Application to Cardiac Image Enhancement and Segmentation. IEEE Trans. Med. Imaging 2018, 37, 384–395.
Duan, J.; Bello, G.; Schlemper, J.; Bai, W.; Dawes, T.J.W.; Biffi, C.; de Marvao, A.; Doumoud, G.; O’Regan, D.P.; Rueckert, D. Automatic 3D Bi-Ventricular Segmentation of Cardiac Images by a Shape-Refined Multi-Task Deep Learning Approach. IEEE Trans. Med. Imaging 2019, 38, 2151–2164.
Painchaud, N.; Skandarani, Y.; Judge, T.; Bernard, O.; Lalande, A.; Jodoin, P.-M. Cardiac Segmentation With Strong Anatomical Guarantees. IEEE Trans. Med. Imaging 2020, 39, 3703–3713.
Park, J.Y.; Noh, Y.; Choi, B.G.; Rha, S.W. CRT-100.48 Prediction of Acute Myocardial Infarction Using a Machine Learning-Based Approach From Data at Admission. JACC Cardiovasc. Interv. 2020, 13, S13.
Pang, S.; Yu, Z.; Orgun, M.A. A novel end-to-end classifier using domain transferred deep convolutional neural networks for biomedical images. Comput. Methods Programs Biomed. 2017, 140, 283–293.
Tsubaki, M.; Tomii, K.; Sese, J. Compound–protein interaction prediction with end-to-end learning of neural networks for graphs and sequences. Bioinformatics 2019, 35, 309–318.
Tang, F.; Bai, C.; Zhao, X.-X.; Yuan, W.-F. Artificial Intelligence and Myocardial Contrast Enhancement Pattern. Curr. Cardiol. Rep. 2020, 22, 77.
Hombach, V.; Merkle, N.; Kestler, H.A.; Torzewski, J.; Kochs, M.; Marx, N.; Nusser, T.; Burgstahler, C.; Rasche, V.; Bernhardt, P.; et al. Characterization of patients with acute chest pain using cardiac magnetic resonance imaging. Clin. Res. Cardiol. Suppl. 2010, 5, 63–69.

© Text is available under the terms and conditions of the Creative Commons Attribution (CC BY) license; additional terms may apply. By using this site, you agree to the Terms and Conditions and Privacy Policy.

Upload a video for this entry

Information

Subjects: Computer Science, Interdisciplinary Applications

Contributors MDPI registered users' name will be linked to their SciProfiles pages. To register with us, please refer to https://encyclopedia.pub/register :

Hossein Zareiamand

Amin Darroudi

Iraj Mohammadi

Seyed Vahid Moravvej

Saba Danaei

Roohallah Alizadehsani

View Times: 300

Update Date: 11 Sep 2023

Table of Contents

Video Upload Options

Confirm