Practical Guide to Evaluating Myocardial Disease by Cardiac MRI

Automated assessment of cardiac pathologies on cardiac MRI using T1-mapping and late gadolinium phase sensitive inversion recovery sequences with deep learning

BACKGROUND

A deep learning (DL) model that automatically detects cardiac pathologies on cardiac MRI may help streamline the diagnostic workflow. To develop a DL model to detect cardiac pathologies on cardiac MRI T1-mapping and late gadolinium phase sensitive inversion recovery (PSIR) sequences were used.

METHODS

Subjects in this study were either diagnosed with cardiac pathology (n = 137) including acute and chronic myocardial infarction, myocarditis, dilated cardiomyopathy, and hypertrophic cardiomyopathy or classified as normal (n = 63). Cardiac MR imaging included T1-mapping and PSIR sequences. Subjects were split 65/15/20% for training, validation, and hold-out testing. The DL models were based on an ImageNet pretrained DenseNet-161 and implemented using PyTorch and fastai. Data augmentation with random rotation and mixup was applied. Categorical cross entropy was used as the loss function with a cyclic learning rate (1e-3). DL models for both sequences were developed separately using similar training parameters. The final model was chosen based on its performance on the validation set. Gradient-weighted class activation maps (Grad-CAMs) visualized the decision-making process of the DL model.

RESULTS

The DL model achieved a sensitivity, specificity, and accuracy of 100%, 38%, and 88% on PSIR images and 78%, 54%, and 70% on T1-mapping images. Grad-CAMs demonstrated that the DL model focused its attention on myocardium and cardiac pathology when evaluating MR images.

CONCLUSIONS

The developed DL models were able to reliably detect cardiac pathologies on cardiac MR images. The diagnostic performance of T1 mapping alone is particularly of note since it does not require a contrast agent and can be acquired quickly.

Protein MRI Contrast Agents as an Effective Approach for Precision Molecular Imaging

ABSTRACT

Cancer and other acute and chronic diseases are results of perturbations of common molecular determinants in key biological and signaling processes. Imaging is critical for characterizing dynamic changes in tumors and metastases, the tumor microenvironment, tumor-stroma interactions, and drug targets, at multiscale levels. Magnetic resonance imaging (MRI) has emerged to be a primary imaging modality for both clinical and preclinical applications due to its advantages over other modalities, including sensitivity to soft tissues, nondepth limitations, and the use of nonionizing radiation. However, extending the application of MRI to achieve both qualitative and quantitative precise molecular imaging with the capability to quantify molecular biomarkers for early detection, staging, and monitoring therapeutic treatment requires the capacity to overcome several major challenges including the trade-off between metal-binding affinity and relaxivity, which is an issue frequently associated with small chelator contrast agents. In this review, we will introduce the criteria of ideal contrast agents for precision molecular imaging and discuss the relaxivity of current contrast agents with defined first shell coordination water molecules. We will then report our advances in creating a new class of protein-targeted MRI contrast agents (ProCAs) with contributions to relaxivity largely derived from the secondary sphere and correlation time. We will summarize our rationale, design strategy, and approaches to the development and optimization of our pioneering ProCAs with desired high relaxivity, metal stability, and molecular biomarker-targeting capability, for precision MRI. From first generation (ProCA1) to third generation (ProCA32), we have achieved dual high r1 and r2 values that are 6- to 10-fold higher than clinically approved contrast agents at magnetic fields of 1.5 T, and their relaxivity values at high field are also significantly higher, which enables high resolution during small animal imaging. Further engineering of multiple targeting moieties enables ProCA32 agents that have strong biomarker-binding affinity and specificity for an array of key molecular biomarkers associated with various chronic diseases, while maintaining relaxation and exceptional metal-binding and selectivity, serum stability, and resistance to transmetallation, which are critical in mitigating risks associated with metal toxicity. Our leading product ProCA32.collagen has enabled the first early detection of liver metastasis from multiple cancers at early stages by mapping the tumor environment and early stage of fibrosis from liver and lung in vivo, with strong translational potential to extend to precision MRI for preclinical and clinical applications for precision diagnosis and treatment.

Optimizing Deep Learning for Cardiac MRI Segmentation: The Impact of Automated Slice Range Classification

RATIONALE AND OBJECTIVES

Cardiac magnetic resonance imaging is crucial for diagnosing cardiovascular diseases, but lengthy postprocessing and manual segmentation can lead to observer bias. Deep learning (DL) has been proposed for automated cardiac segmentation; however, its effectiveness is limited by the slice range selection from base to apex.

MATERIALS AND METHODS

In this study, we integrated an automated slice range classification step to identify basal to apical short-axis slices before DL-based segmentation. We employed publicly available Multi-Disease, Multi-View & Multi-Center Right Ventricular Segmentation in Cardiac MRI data set with short-axis cine data from 160 training, 40 validation, and 160 testing cases. Three classification and seven segmentation DL models were studied. The top-performing segmentation model was assessed with and without the classification model. Model validation to compare automated and manual segmentation was performed using Dice score and Hausdorff distance and clinical indices (correlation score and Bland-Altman plots).

RESULTS

The combined classification (CBAM-integrated 2D-CNN) and segmentation model (2D-UNet with dilated convolution block) demonstrated superior performance, achieving Dice scores of 0.952 for left ventricle (LV), 0.933 for right ventricle (RV), and 0.875 for myocardium, compared to the stand-alone segmentation model (0.949 for LV, 0.925 for RV, and 0.867 for myocardium). Combined classification and segmentation model showed high correlation (0.92-0.99) with manual segmentation for biventricular volumes, ejection fraction, and myocardial mass. The mean absolute difference (2.8-8.3 mL) for clinical parameters between automated and manual segmentation was within the interobserver variability range, indicating comparable performance to manual annotation.

CONCLUSION

Integrating an initial automated slice range classification step into the segmentation process improves the performance of DL-based cardiac chamber segmentation.

Heart Failure Preserved Ejection Fraction in Women: Insights Learned from Imaging

While the prevalence of heart failure, in general, is similar in men and women, women experience a higher rate of HFpEF compared to HFrEF. Cardiovascular risk factors, parity, estrogen levels, cardiac physiology, and altered response to the immune system may be at the root of this difference. Studies have found that in response to increasing age and hypertension, women experience more concentric left ventricle remodeling, more ventricular and arterial stiffness, and less ventricular dilation compared to men, which predisposes women to developing more diastolic dysfunction. A multi-modality imaging approach is recommended to identify patients with HFpEF. Particularly, appreciation of sex-based differences as described in this review is important in optimizing the evaluation and care of women with HFpEF.

Diastology with Cardiac MRI: A Practical Guide

Diastolic filling of the ventricle is a complex interplay of volume and pressure, contingent on active energy-dependent myocardial relaxation and myocardial stiffness. Abnormal diastolic function is the hallmark of the clinical entity of heart failure with preserved ejection fraction (HFpEF), which is now the dominant type of heart failure and is associated with significant morbidity and mortality. Although echocardiography is the current first-line imaging modality used in evaluation of diastolic function, cardiac MRI (CMR) is emerging as an important technique. The principal role of CMR is to categorize the cause of diastolic dysfunction (DD) and distinguish other entities that manifest similarly to HFpEF, particularly infiltrative and pericardial disorders. CMR also provides prognostic information and risk stratification based on late gadolinium enhancement and parametric mapping techniques. Advances in hardware, sequences, and postprocessing software now enable CMR to diagnose and grade DD accurately, a role traditionally assigned to echocardiography. Two-dimensional or four-dimensional velocity-encoded phase-contrast sequences can measure flow and velocities at the mitral inflow, mitral annulus, and pulmonary veins to provide diastolic functional metrics analogous to those at echocardiography. The commonly used cine steady-state free-precession sequence can provide clues to DD including left ventricular mass, left ventricular filling curves, and left atrial size and function. MR strain imaging provides information on myocardial mechanics that further aids in diagnosis and prognosis of diastolic function. Research sequences such as MR elastography and MR spectroscopy can help evaluate myocardial stiffness and metabolism, respectively, providing additional insights on diastolic function. The authors review the physiology of diastolic function, mechanics of diastolic heart failure, and CMR techniques in the evaluation of diastolic function. ©RSNA, 2023 Quiz questions for this article are available in the supplemental material.

Artificial Intelligence Applications in Cardiovascular Magnetic Resonance Imaging: Are We on the Path to Avoiding the Administration of Contrast Media?

In recent years, cardiovascular imaging examinations have experienced exponential growth due to technological innovation, and this trend is consistent with the most recent chest pain guidelines. Contrast media have a crucial role in cardiovascular magnetic resonance (CMR) imaging, allowing for more precise characterization of different cardiovascular diseases. However, contrast media have contraindications and side effects that limit their clinical application in determinant patients. The application of artificial intelligence (AI)-based techniques to CMR imaging has led to the development of non-contrast models. These AI models utilize non-contrast imaging data, either independently or in combination with clinical and demographic data, as input to generate diagnostic or prognostic algorithms. In this review, we provide an overview of the main concepts pertaining to AI, review the existing literature on non-contrast AI models in CMR, and finally, discuss the strengths and limitations of these AI models and their possible future development.

Cardiac Magnetic Resonance Imaging in Appraising Myocardial Strain and Biomechanics: A Current Overview

Subclinical alterations in myocardial structure and function occur early during the natural disease course. In contrast, clinically overt signs and symptoms occur during late phases, being associated with worse outcomes. Identification of such subclinical changes is critical for timely diagnosis and accurate management. Hence, implementing cost-effective imaging techniques with accuracy and reproducibility may improve long-term prognosis. A growing body of evidence supports using cardiac magnetic resonance (CMR) to quantify deformation parameters. Tissue-tagging (TT-CMR) and feature-tracking CMR (FT-CMR) can measure longitudinal, circumferential, and radial strains and recent research emphasize their diagnostic and prognostic roles in ischemic heart disease and primary myocardial illnesses. Additionally, these methods can accurately determine LV wringing and functional dynamic geometry parameters, such as LV torsion, twist/untwist, LV sphericity index, and long-axis strain, and several studies have proved their utility in prognostic prediction in various cardiovascular patients. More recently, few yet important studies have suggested the superiority of fast strain-encoded imaging CMR-derived myocardial strain in terms of accuracy and significantly reduced acquisition time, however, more studies need to be carried out to establish its clinical impact. Herein, the current review aims to provide an overview of currently available data regarding the role of CMR in evaluating myocardial strain and biomechanics.

Utility of Exercise Stress Testing in Pediatric Patients with T-Wave Inversions

T-wave inversions (TWI) in inferolateral electrocardiogram (ECG) leads (II, III, aVF, V5-V6) are often suggestive of cardiac pathology in adult patients. They are a common reason of additional testing in pediatric patients. The goal of the study is to determine correlation of exercise stress test T-wave response with diagnosis of cardiac pathology. This is a retrospective review of pediatric patients < 21 years with the finding of baseline TWI in inferior (II, III, avF) and/or lateral (V5-V6) leads on 12-lead ECG. All patients underwent treadmill exercise stress test (EST) and an echocardiogram within 1 year of each other. Demographics, baseline ECG findings, echocardiogram results, and EST data were recorded. T-wave reversion was considered complete if T waves demonstrated normalization during exercise, partial if there was minimal improvement, and no response if there was no change or worsening of inversions. In our cohort of 72 patients with a mean age 14.6 (± 2.9) years and 61% males, 59 (82%) had a structurally normal heart. Thirteen patients had evidence of structural or functional heart disease. Of the 59 patients, 83% had either complete or partial T-wave response. Among the 13 patients with heart disease, two patients had genetic testing consistent with risk for hypertrophic cardiomyopathy and had complete and partial response on EST. Exercise stress testing for TWI in pediatric patients has low sensitivity and specificity for the diagnosis of cardiac disease and routine use in this patient population may not be indicated.

Quality of science and reporting for radiomics in cardiac magnetic resonance imaging studies: a systematic review

OBJECTIVES

To evaluate the quality of radiomics studies using cardiac magnetic resonance imaging (CMR) according to the radiomics quality score (RQS), Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) guidelines, and the standards defined by the Image Biomarker Standardization Initiative (IBSI) and identify areas needing improvement.

MATERIALS AND METHODS

PubMed and Embase were searched to identify radiomics studies using CMR until March 10, 2021. Of the 259 identified articles, 32 relevant original research articles were included. Studies were scored according to the RQS, TRIPOD guidelines, and IBSI standards by two cardiac radiologists.

RESULTS

The mean RQS was 14.3% of the maximum (5.16 out of 36). RQS were low for the demonstration of validation (-60.6%), calibration statistics (1.6%), potential clinical utility (3.1%), and open science (3.1%) items. No study conducted a phantom study or cost-effectiveness analysis. The adherence to TRIPOD guidelines was 55.9%. Studies were deficient in reporting title (3.1%), stating objective in abstract and introduction (6.3% and 9.4%), missing data (0%), discrimination/calibration (3.1%), and how to use the prediction model (3.1%). According to the IBSI standards, non-uniformity correction, image interpolation, grey-level discretization, and signal intensity normalization were performed in two (6.3%), four (12.5%), six (18.8%), and twelve (37.5%) studies, respectively.

CONCLUSION

The quality of radiomics studies using CMR is suboptimal. Improvements are needed in the areas of validation, calibration, clinical utility, and open science. Complete reporting of study objectives, missing data, discrimination/calibration, how to use the prediction model, and preprocessing steps are necessary.

KEY POINTS

• The quality of science in radiomics studies using CMR is currently inadequate. • RQS were low for validation, calibration, clinical utility, and open science; no study conducted a phantom study or cost-effectiveness analysis. • In stating the study objective, missing data, discrimination/calibration, how to use the prediction model, and preprocessing steps, improvements are needed.

Myocardial Extracellular Volume Fraction and T1 Mapping by Cardiac Magnetic Resonance Compared Between Patients With and Without Type 2 Diabetes, and the Effect of ECV and T2D on Cardiovascular Outcomes

Background: To investigate the difference in myocardial extracellular volume fraction (ECV) by cardiac magnetic resonance (CMR) T1 mapping between patients with and without type 2 diabetes (T2D), and the effect of ECV and T2D on cardiovascular (CV) outcomes. Methods: All patients aged > 18 years with known or suspected coronary artery disease who underwent CMR for assessment of myocardial ischemia or myocardial viability at the Department of Cardiology of the Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand from September 2017 to December 2018 were screened for inclusion eligibility. Left ventricular ejection fraction (LVEF), late gadolinium enhancement, and T1 mapping were performed. ECV values were derived from myocardial native T1 and contrast-enhanced T1 values that were obtained using modified Look-Locker inversion recovery at the septum of the mid-cavity short-axis map. Demographic data, clinical characteristics, and CV outcomes were collected by retrospective chart review. Composite CV outcomes included CV death, acute coronary syndrome, heart failure hospitalization, or ventricular tachycardia (VT)/ventricular fibrillation. Results: A total of 739 subjects (mean age: 69.5 ± 14.0 years, 49.3% men) were included. Of those, 188 subjects had T2D (25.4%). ECV was significantly higher in T2D than in non-T2D (30.0 ± 5.9% vs. 28.8 ± 4.7%, p = 0.004). During the mean follow-up duration of 26.2 ± 8.5 months, 43 patients (5.8%) had a clinical composite outcome, as follows: three CV death (0.4%), seven acute coronary syndrome (0.9%), 33 heart failure hospitalization (4.5%), and one VT (0.1%). T2D, low LVEF, and high ECV were all identified as independent predictors of CV events. Patients with T2D and high ECV had the highest risk of CV events. Conclusion: Among patients with known or suspected coronary artery disease, patients with T2D had a higher ECV. T2D and high ECV were both found to be independent risk factors for adverse CV outcomes.

Insights on the left ventricular thrombus in patients with ischemic dilated cardiomyopathy

Background

Left ventricular thrombus (LVT) formation represents a common complication of dilated cardiomyopathy (DCM). LVT is usually underestimated with transthoracic echocardiography (TTE) while cardiac magnetic resonance imaging (CMRI) is promising as an alternative imaging modality for cardiac thrombus detection. The study aims to compare TTE and CMRI in their ability to detect LVT, also, to assess the clinical and imaging parameters to determine variables that may predispose for thrombus formation. The study population includes seventy-six patients with ischemic DCM. They were divided into 2 groups based on the presence of LVT as detected by delayed-enhancement CMRI (DE-CMRI) [Group A included 20 patients with a LVT and Group B included 56 patients without].

Results

All of the current study population had ischemic DCM with left ventricular ejection fraction (LVEF) < 50%. DE-CMR detected thrombus in 20 cases of the studied population that represented group A. From group A, conventional TTE detected LVT only in 8 and cine-CMR detected 13 cases out of the out of 20 cases. The ejection fraction of the left ventricle as measured by functional CMRI was significantly lower in group A (P = 0.045). Interestingly, the myocardial scarring in group A was seen significantly more extensive than in group B (the P value is < 0.00001), paralleling the increased prevalence of thrombus.

Conclusions

DE-CMRI provides superiority for the detection of LVT compared with standard TTE or cine-CMRI and the amount of myocardial scarring detected by DE-CMRI can be considered an independent marker for thrombus presence.

The Radiologist as a Gatekeeper in Chest Pain

Chest pain is a symptom that can be found in life-threatening conditions such as acute coronary syndrome (ACS). Those patients requiring invasive coronary angiography treatment or surgery should be identified. Often the clinical setting and laboratory tests are not sufficient to rule out a coronary or aortic syndrome. Cardiac radiological imaging has evolved in recent years both in magnetic resonance (MR) and in computed tomography (CT). CT, in particular, due to its temporal and spatial resolution, the quickness of the examination, and the availability of scanners, is suitable for the evaluation of these patients. In particular, the latest-generation CT scanners allow the exclusion of diagnoses such as coronary artery disease and aortic pathology, thereby reducing the patient's stay in hospital and safely selecting patients by distinguishing those who do not need further treatment from those who will need more- or less-invasive therapies. CT additionally reduces costs by improving long-term patient outcome. The limitations related to patient characteristics and those related to radiation exposure are weakening with the improvement of CT technology.

Evaluation and Management of Aortic Stenosis in Chronic Kidney Disease: A Scientific Statement From the American Heart Association

Aortic stenosis with concomitant chronic kidney disease (CKD) represents a clinical challenge. Aortic stenosis is more prevalent and progresses more rapidly and unpredictably in CKD, and the presence of CKD is associated with worse short-term and long-term outcomes after aortic valve replacement. Because patients with advanced CKD and end-stage kidney disease have been excluded from randomized trials, clinicians need to make complex management decisions in this population that are based on retrospective and observational evidence. This statement summarizes the epidemiological and pathophysiological characteristics of aortic stenosis in the context of CKD, evaluates the nuances and prognostic information provided by noninvasive cardiovascular imaging with echocardiography and advanced imaging techniques, and outlines the special risks in this population. Furthermore, this statement provides a critical review of the existing literature pertaining to clinical outcomes of surgical versus transcatheter aortic valve replacement in this high-risk population to help guide clinical decision making in the choice of aortic valve replacement and specific prosthesis. Finally, this statement provides an approach to the perioperative management of these patients, with special attention to a multidisciplinary heart-kidney collaborative team-based approach.

Molecular MR Contrast Agents

Molecular magnetic resonance (MR) imaging utilizes molecular probes to provide added biochemical or cellular information to what can already be achieved with anatomical and functional MR imaging. This review provides an overview of molecular MR and focuses specifically on molecular MR contrast agents that provide contrast by shortening the T1 time. We describe the requirements for a successful molecular MR contrast agent and the challenges for clinical translation. The review highlights work from the last 5 years and places an emphasis on new contrast agents that have been validated in multiple preclinical models. Applications of molecular MR include imaging of inflammation, fibrosis, fibrogenesis, thromboembolic disease, and cancers. Molecular MR is positioned to move beyond detection of disease to the quantitative staging of disease and measurement of treatment response.