Search

Invasive angiography is the reference standard for coronary artery disease (CAD) diagnosis but is expensive and associated with certain risks. Machine learning (ML) using clinical and noninvasive imaging parameters can be used for CAD diagnosis to avoid the side effects and cost of angiography. However, ML methods require labeled samples for efficient training. The labeled data scarcity and high labeling costs can be mitigated by active learning. This is achieved through selective query of challenging samples for labeling. To the best of our knowledge, active learning has not been used for CAD diagnosis yet. An Active Learning with Ensemble of Classifiers (ALEC) method is proposed for CAD diagnosis, consisting of four classifiers. Three of these classifiers determine whether a patient's three main coronary arteries are stenotic or not. The fourth classifier predicts whether the patient has CAD or not. ALEC is first trained using labeled samples. For each unlabeled sample, if the outputs of the classifiers are consistent, the sample along with its predicted label is added to the pool of labeled samples. Inconsistent samples are manually labeled by medical experts before being added to the pool. The training is performed once more using the samples labeled so far. The interleaved phases of labeling and training are repeated until all samples are labeled. Compared with 19 other active learning algorithms, ALEC combined with a support vector machine classifier attained superior performance with 97.01% accuracy. Our method is justified mathematically as well. We also comprehensively analyze the CAD dataset used in this paper. As part of dataset analysis, features pairwise correlation is computed. The top 15 features contributing to CAD and stenosis of the three main coronary arteries are determined. The relationship between stenosis of the main arteries is presented using conditional probabilities. The effect of considering the number of stenotic arteries on sample discrimination is investigated. The discrimination power over dataset samples is visualized, assuming each of the three main coronary arteries as a sample label and considering the two remaining arteries as sample features., Competing Interests: Declaration of competing interest None Declared., (Copyright © 2023 The Authors. Published by Elsevier Ltd.. All rights reserved.)

Coronary artery disease (CAD) is a prevalent disease with high morbidity and mortality rates. Invasive coronary angiography is the reference standard for diagnosing CAD but is costly and associated with risks. Noninvasive imaging like cardiac magnetic resonance (CMR) facilitates CAD assessment and can serve as a gatekeeper to downstream invasive testing. Machine learning methods are increasingly applied for automated interpretation of imaging and other clinical results for medical diagnosis. In this study, we proposed a novel CAD detection method based on CMR images by utilizing the feature extraction ability of deep neural networks and combining the features with the aid of a random forest for the very first time. It is necessary to convert image data to numeric features so that they can be used in the nodes of the decision trees. To this end, the predictions of multiple stand-alone convolutional neural networks (CNNs) were considered as input features for the decision trees. The capability of CNNs in representing image data renders our method a generic classification approach applicable to any image dataset. We named our method RF-CNN-F, which stands for Random Forest with CNN Features. We conducted experiments on a large CMR dataset that we have collected and made publicly accessible. Our method achieved excellent accuracy (99.18%) using Adam optimizer compared to a stand-alone CNN trained using fivefold cross validation (93.92%) tested on the same dataset., (© 2022. The Author(s).)

Introduction: Finite element (FE) mechanics models of the heart are becoming more sophisticated. However, there is lack of consensus about optimal element type and coupling of FE models to the circulation. We describe biventricular (left (LV) and right (RV) ventricles) FE mechanics model creation using hexahedral elements, airbags and a functional mockup interface (FMI) to lumped-parameter models of the circulation., Methods: Cardiac MRI (CMR) was performed in two healthy volunteers and a single patient with ischemic heart disease (IHD). CMR images were segmented and surfaced, meshing with hexahedral elements was performed with a "thin butterfly with septum" topology. LV and RV inflow and outflow airbags were coupled to lumped-parameter circulation models with an FMI interface. Pulmonary constriction (PAC) and vena cava occlusion (VCO) were simulated and end-systolic pressure-volume relations (ESPVR) were calculated., Results: Mesh construction was prompt with representative contouring and mesh adjustment requiring 32 and 26 min Respectively. The numbers of elements ranged from 4104 to 5184 with a representative Jacobian of 1.0026 ± 0.4531. Agreement between CMR-based surfaces and mesh was excellent with root-mean-squared error of 0.589 ± 0.321 mm. The LV ESPVR slope was 3.37 ± 0.09 in volunteers but 2.74 in the IHD patient. The effect of PAC and VCO on LV ESPVR was consistent with ventricular interaction (p = 0.0286)., Conclusion: Successful co-simulation using a biventricular FE mechanics model with hexahedral elements, airbags and an FMI interface to lumped-parameter model of the circulation was demonstrated. Future studies will include comparison of element type and study of cardiovascular pathologies and device therapies., (Copyright © 2021. Published by Elsevier Ltd.)

Purpose: Myocardial strain is increasingly used to assess left ventricular (LV) function. Incorporation of LV deformation into finite element (FE) modeling environment with subsequent strain calculation will allow analysis to reach its full potential. We describe a new kinematic model-based analysis framework (KMAF) to calculate strain from 3D cine-DENSE (displacement encoding with stimulated echoes) MRI., Methods: Cine-DENSE allows measurement of 3D myocardial displacement with high spatial accuracy. The KMAF framework uses cine cardiovascular magnetic resonance (CMR) to facilitate cine-DENSE segmentation, interpolates cine-DENSE displacement, and kinematically deforms an FE model to calculate strain. This framework was validated in an axially compressed gel phantom and applied in 10 healthy sheep and 5 sheep after myocardial infarction (MI)., Results: Excellent Bland-Altman agreement of peak circumferential (E cc ) and longitudinal (E ll ) strain (mean difference = 0.021 ± 0.04 and -0.006 ± 0.03, respectively), was found between KMAF estimates and idealized FE simulation. E rr had a mean difference of -0.014 but larger variation (±0.12). Cine-DENSE estimated end-systolic (ES) E cc , E ll and E rr exhibited significant spatial variation for healthy sheep. Displacement magnitude was reduced on average by 27%, 42%, and 56% after MI in the remote, adjacent and MI regions, respectively., Conclusions: The KMAF framework allows accurate calculation of 3D LV E cc and E ll from cine-DENSE., (© 2021 International Society for Magnetic Resonance in Medicine. This article has been contributed to by US Government employees and their work is in the public domain in the USA.)

Respiratory illnesses, such as bronchitis, emphysema, asthma, and COVID-19, substantially remodel lung tissue, deteriorate function, and culminate in a compromised breathing ability. Yet, the structural mechanics of the lung is significantly understudied. Classical pressure-volume air or saline inflation studies of the lung have attempted to characterize the organ's elasticity and compliance, measuring deviatory responses in diseased states; however, these investigations are exclusively limited to the bulk composite or global response of the entire lung and disregard local expansion and stretch phenomena within the lung lobes, overlooking potentially valuable physiological insights, as particularly related to mechanical ventilation. Here, we present a method to collect the first non-contact, full-field deformation measures of ex vivo porcine and murine lungs and interface with a pressure-volume ventilation system to investigate lung behavior in real time. We share preliminary observations of heterogeneous and anisotropic strain distributions of the parenchymal surface, associative pressure-volume-strain loading dependencies during continuous loading, and consider the influence of inflation rate and maximum volume. This study serves as a crucial basis for future works to comprehensively characterize the regional response of the lung across various species, link local strains to global lung mechanics, examine the effect of breathing frequencies and volumes, investigate deformation gradients and evolutionary behaviors during breathing, and contrast healthy and pathological states. Measurements collected in this framework ultimately aim to inform predictive computational models and enable the effective development of ventilators and early diagnostic strategies., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2020 Mariano, Sattari, Maghsoudi-Ganjeh, Tartibi, Lo and Eskandari.)

An analysis is presented of the equilibrium response of a radially deformed cylinder of isotropic, incompressible bio-elastic material swollen by an infused liquid satisfying a static diffusive balance law.

Background: Therapy for moderate ischemic mitral regurgitation remains unclear. Determination of myocardial viability, a necessary prerequisite for an improvement in regional contractility, is a likely key factor in determining response to revascularization alone. Myocardial strain has been proposed as a viability measure but has not been compared with late gadolinium enhancement (LGE) cardiac magnetic resonance imaging. We hypothesized that abnormal strain overestimates nonviable left ventricular (LV) segments measured with LGE and that ischemia and mechanical tethering by adjacent transmural myocardial infarction (TMI) also decreases strain in viable segments., Methods: Sixteen patients with mild or greater ischemic mitral regurgitation and 7 healthy volunteers underwent cardiac magnetic resonance imaging with noninvasive tags (complementary spatial modulation of magnetization [CSPAMM]), LGE, and stress perfusion. CSPAMM images were post-processed with harmonic phase and circumferential and longitudinal strains were calculated. Viability was defined as the absence of TMI on LGE (hyperenhancement >50% of wall thickness). The borderzone was defined as any segment bordering TMI. Abnormal strain thresholds (±1 to 2.5 SDs from normal mean) were compared with TMI, ischemia, and borderzone., Results: 7.4% of LV segments had TMI on LGE, and more than 14.5% of LV segments were nonviable by strain thresholds (p < 0.005). In viable segments, ischemia impaired longitudinal strain (least perfused one-third of LV segments: -0.18 ± 0.08 versus most perfused: -0.22 ± 0.1, p = 0.01) and circumferential strain (-0.12 ± 0.1 versus -0.16 ± 0.08, p < 0.05). In addition, infarct proximity impaired longitudinal strain (-0.16 ± 0.11 borderzone versus -0.18 ± 0.09 remote, p = 0.05)., Conclusions: Impaired LV strain overestimates nonviable myocardium compared with TMI on LGE. Ischemia and infarct proximity also decrease strain in viable segments., (Copyright © 2018 The Society of Thoracic Surgeons. Published by Elsevier Inc. All rights reserved.)

Background: Chronic ischemic mitral regurgitation (CIMR) is associated with poor outcome. Left ventricular (LV) strain after posterolateral myocardial infarction (MI) may drive LV remodeling. Although moderate CIMR has been previously shown to affect LV remodeling, the effect of CIMR on LV strain after posterolateral MI remains unknown. We tested the hypothesis that moderate CIMR alters LV strain after posterolateral MI., Methods: Posterolateral MI was created in 10 sheep. Cardiac magnetic resonance imaging with tags was performed 2 weeks before and 2, 8, and 16 weeks after MI. The left and right ventricular volumes were measured, and regurgitant volume indexed to body surface area (regurgitant volume index) was calculated as the difference between left ventricle and right ventricle stroke volumes divided by body surface area. Three-dimensional strain was calculated., Results: Circumferential strain (Ecc) and longitudinal strain (Ell) were reduced in the infarct proper, MI borderzone, and remote myocardium 16 weeks after MI. In addition, radial circumferential (Erc) and radial longitudinal (Erl) shear strains were reduced in remote myocardium but increased in the infarct and borderzone 16 weeks after MI. Of all strain components, however, only Erc was affected by regurgitant volume index (p = 0.0005). There was no statistically significant effect of regurgitant volume index on Ecc, Ell, Erl, or circumferential longitudinal shear strain (Ecl)., Conclusions: Moderate CIMR alters radial circumferential shear strain after posterolateral MI in sheep. Further studies are needed to determine the effect of shear strain on myocyte hypertrophy and the effect of mitral repair on myocardial strain., (Copyright © 2016 The Society of Thoracic Surgeons. Published by Elsevier Inc. All rights reserved.)

Background: Surgical ventricular restoration (Dor procedure) is generally thought to reduce left ventricular (LV) myofiber stress (FS) but to adversely affect pump function. However, the underlying mechanism is unclear. The goal of this study was to determine the effect of residual stress (RS) on LV FS and pump function after the Dor procedure., Methods: Previously described finite element models of the LV based on magnetic resonance imaging data obtained in 5 sheep 16 weeks after anteroapical myocardial infarction were used. Simulated polyethylene terephthalate fiber (Dacron) patches that were elliptical and 25% of the infarct opening area were implanted using a virtual suture technique (VIRTUAL-DOR). In each case, diastole and systole were simulated, and RS, FS, LV volumes, systolic and diastolic function, and pump (Starling) function were calculated., Results: VIRTUAL-DOR was associated with significant RS that was tensile (2.89 ± 1.31 kPa) in the remote myocardium and compressive (234.15 ± 65.53 kPa) in the border zone. VIRTUAL-DOR+RS (compared with VIRTUAL-DOR-NO-RS) was associated with further reduction in regional diastolic and systolic FS, with the greatest change in the border zone (43.5-fold and 7.1-fold, respectively; p < 0.0001). VIRTUAL-DOR+RS was also associated with further reduction in systolic and diastolic volumes (7.9%; p = 0.0606, and 10.6%; p = 0.0630, respectively). The resultant effect was a further reduction in pump function after VIRTUAL-DOR+RS., Conclusions: Residual stress that occurs after the Dor procedure is positive (tensile) in the remote myocardium and negative (compressive) in the border zone and associated with reductions in FS and LV volumes. The resultant effect is a further reduction in LV pump (Starling) function., (Copyright © 2015 The Society of Thoracic Surgeons. Published by Elsevier Inc. All rights reserved.)