Machine learning analysis of complex late gadolinium enhancement patterns to improve risk prediction of major arrhythmic events.

Background: Machine learning analysis of complex myocardial scar patterns affords the potential to enhance risk prediction of life-threatening arrhythmia in stable coronary artery disease (CAD). Objective: To assess the utility of computational image analysis, alongside a machine learning (ML) approach, to identify scar microstructure features on late gadolinium enhancement cardiovascular magnetic resonance (LGE-CMR) that predict major arrhythmic events in patients with CAD. Methods: Patients with stable CAD were prospectively recruited into a CMR registry. Shape-based scar microstructure features characterizing heterogeneous ('peri-infarct') and homogeneous ('core') fibrosis were extracted. An ensemble of machine learning approaches were used for risk stratification, in addition to conventional analysis using Cox modeling. Results: Of 397 patients (mean LVEF 45.4 ± 16.0) followed for a median of 6 years, 55 patients (14%) experienced a major arrhythmic event. When applied within an ML model for binary classification, peri-infarct zone (PIZ) entropy, peri-infarct components and core interface area outperformed a model representative of the current standard of care (LVEF<35% and NYHA>Class I): AUROC (95%CI) 0.81 (0.81-0.82) vs. 0.64 (0.63-0.65), p = 0.002. In multivariate cox regression analysis, these features again remained significant after adjusting for LVEF<35% and NYHA>Class I: PIZ entropy hazard ratio (HR) 1.88, 95% confidence interval (CI) 1.38-2.56, p < 0.001; number of PIZ components HR 1.34, 95% CI 1.08-1.67, p = 0.009; core interface area HR 1.6, 95% CI 1.29-1.99, p = <0.001. Conclusion: Machine learning models using LGE-CMR scar microstructure improved arrhythmic risk stratification as compared to guideline-based clinical parameters; highlighting a potential novel approach to identifying candidates for implantable cardioverter defibrillators in stable CAD.

Comprehensive Phenotypic Characterization of Late Gadolinium Enhancement Predicts Sudden Cardiac Death in Coronary Artery Disease.

BACKGROUND: Late gadolinium enhancement (LGE) cardiac magnetic resonance (CMR) offers the potential to noninvasively characterize the phenotypic substrate for sudden cardiac death (SCD). OBJECTIVES: The authors assessed the utility of infarct characterization by CMR, including scar microstructure analysis, to predict SCD in patients with coronary artery disease (CAD). METHODS: Patients with stable CAD were prospectively recruited into a CMR registry. LGE quantification of core infarction and the peri-infarct zone (PIZ) was performed alongside computational image analysis to extract morphologic and texture scar microstructure features. The primary outcome was SCD or aborted SCD. RESULTS: Of 437 patients (mean age: 64 years; mean left ventricular ejection fraction [LVEF]: 47%) followed for a median of 6.3 years, 49 patients (11.2%) experienced the primary outcome. On multivariable analysis, PIZ mass and core infarct mass were independently associated with the primary outcome (per gram: HR: 1.07 [95% CI: 1.02-1.12]; P = 0.002 and HR: 1.03 [95% CI: 1.01-1.05]; P = 0.01, respectively), and the addition of both parameters improved discrimination of the model (Harrell's C-statistic: 0.64-0.79). PIZ mass, however, did not provide incremental prognostic value over core infarct mass based on Harrell's C-statistic or risk reclassification analysis. Severely reduced LVEF did not predict the primary endpoint after adjustment for scar mass. On scar microstructure analysis, the number of LGE islands in addition to scar transmurality, radiality, interface area, and entropy were all associated with the primary outcome after adjustment for severely reduced LVEF and New York Heart Association functional class of >1. No scar microstructure feature remained associated with the primary endpoint when PIZ mass and core infarct mass were added to the regression models. CONCLUSIONS: Comprehensive LGE characterization independently predicted SCD risk beyond conventional predictors used in implantable cardioverter-defibrillator (ICD) insertion guidelines. These results signify the potential for a more personalized approach to determining ICD candidacy in CAD.

TCRpower: quantifying the detection power of T-cell receptor sequencing with a novel computational pipeline calibrated by spike-in sequences.

T-cell receptor (TCR) sequencing has enabled the development of innovative diagnostic tests for cancers, autoimmune diseases and other applications. However, the rarity of many T-cell clonotypes presents a detection challenge, which may lead to misdiagnosis if diagnostically relevant TCRs remain undetected. To address this issue, we developed TCRpower, a novel computational pipeline for quantifying the statistical detection power of TCR sequencing methods. TCRpower calculates the probability of detecting a TCR sequence as a function of several key parameters: in-vivo TCR frequency, T-cell sample count, read sequencing depth and read cutoff. To calibrate TCRpower, we selected unique TCRs of 45 T-cell clones (TCCs) as spike-in TCRs. We sequenced the spike-in TCRs from TCCs, together with TCRs from peripheral blood, using a 5' RACE protocol. The 45 spike-in TCRs covered a wide range of sample frequencies, ranging from 5 per 100 to 1 per 1 million. The resulting spike-in TCR read counts and ground truth frequencies allowed us to calibrate TCRpower. In our TCR sequencing data, we observed a consistent linear relationship between sample and sequencing read frequencies. We were also able to reliably detect spike-in TCRs with frequencies as low as one per million. By implementing an optimized read cutoff, we eliminated most of the falsely detected sequences in our data (TCR α-chain 99.0% and TCR ß-chain 92.4%), thereby improving diagnostic specificity. TCRpower is publicly available and can be used to optimize future TCR sequencing experiments, and thereby enable reliable detection of disease-relevant TCRs for diagnostic applications.

Left ventricular shape predicts arrhythmic risk in fibrotic dilated cardiomyopathy.

AIMS: Remodelling of the left ventricular (LV) shape is one of the hallmarks of non-ischaemic dilated cardiomyopathy (DCM) and may contribute to ventricular arrhythmias and sudden cardiac death. We sought to investigate a novel three dimensional (3D) shape analysis approach to quantify LV remodelling for arrhythmia prediction in DCM. METHODS AND RESULTS: We created 3D LV shape models from end-diastolic cardiac magnetic resonance images of 156 patients with DCM and late gadolinium enhancement (LGE). Using the shape models, principle component analysis, and Cox-Lasso regression, we derived a prognostic LV arrhythmic shape (LVAS) score which identified patients who reached a composite arrhythmic endpoint of sudden cardiac death, aborted sudden cardiac death, and sustained ventricular tachycardia. We also extracted geometrical metrics to look for potential prognostic markers. During a follow-up period of up to 16 years (median 7.7, interquartile range: 3.9), 25 patients met the arrhythmic endpoint. The optimally prognostic LV shape for predicting the time-to arrhythmic event was a paraboloidal longitudinal profile, with a relatively wide base. The corresponding LVAS was associated with arrhythmic events in univariate Cox regression (hazard ratio = 2.0 per quartile; 95% confidence interval: 1.3-2.9), in univariate Cox regression with propensity score adjustment, and in three multivariate models; with LV ejection fraction, New York Heart Association Class III/IV (Model 1), implantable cardioverter-defibrillator receipt (Model 2), and cardiac resynchronization therapy (Model 3). CONCLUSION: Biomarkers of LV shape remodelling in DCM can help to identify the patients at greatest risk of lethal ventricular arrhythmias.

Late-Gadolinium Enhancement Interface Area and Electrophysiological Simulations Predict Arrhythmic Events in Patients With Nonischemic Dilated Cardiomyopathy.

OBJECTIVES: This study sought to investigate whether shape-based late gadolinium enhancement (LGE) metrics and simulations of re-entrant electrical activity are associated with arrhythmic events in patients with nonischemic dilated cardiomyopathy (NIDCM). BACKGROUND: The presence of LGE predicts life-threatening ventricular arrhythmias in NIDCM; however, risk stratification remains imprecise. LGE shape and simulations of electrical activity may be able to provide additional prognostic information. METHODS: Cardiac magnetic resonance (CMR)-LGE shape metrics were computed for a cohort of 156 patients with NIDCM and visible LGE and tested retrospectively for an association with an arrhythmic composite endpoint of sudden cardiac death and ventricular tachycardia. Computational models were created from images and used in conjunction with simulated stimulation protocols to assess the potential for re-entry induction in each patient's scar morphology. A mechanistic analysis of the simulations was carried out to explain the associations. RESULTS: During a median follow-up of 1,611 (interquartile range: 881 to 2,341) days, 16 patients (10.3%) met the primary endpoint. In an inverse probability weighted Cox regression, the LGE-myocardial interface area (hazard ratio [HR]: 1.75; 95% confidence interval [CI]: 1.24 to 2.47; p = 0.001), number of simulated re-entries (HR: 1.40; 95% CI: 1.23 to 1.59; p < 0.01) and LGE volume (HR: 1.44; 95% CI: 1.07 to 1.94; p = 0.02) were associated with arrhythmic events. Computational modeling revealed repolarization heterogeneity and rate-dependent block of electrical wavefronts at the LGE-myocardial interface as putative arrhythmogenic mechanisms directly related to the LGE interface area. CONCLUSIONS: The area of interface between scar and surviving myocardium, as well as simulated re-entrant activity, are associated with an elevated risk of major arrhythmic events in patients with NIDCM and LGE and represent novel risk predictors.

The immuneML ecosystem for machine learning analysis of adaptive immune receptor repertoires.

Adaptive immune receptor repertoires (AIRR) are key targets for biomedical research as they record past and ongoing adaptive immune responses. The capacity of machine learning (ML) to identify complex discriminative sequence patterns renders it an ideal approach for AIRR-based diagnostic and therapeutic discovery. To date, widespread adoption of AIRR ML has been inhibited by a lack of reproducibility, transparency, and interoperability. immuneML (immuneml.uio.no) addresses these concerns by implementing each step of the AIRR ML process in an extensible, open-source software ecosystem that is based on fully specified and shareable workflows. To facilitate widespread user adoption, immuneML is available as a command-line tool and through an intuitive Galaxy web interface, and extensive documentation of workflows is provided. We demonstrate the broad applicability of immuneML by (i) reproducing a large-scale study on immune state prediction, (ii) developing, integrating, and applying a novel deep learning method for antigen specificity prediction, and (iii) showcasing streamlined interpretability-focused benchmarking of AIRR ML.

T cell receptor repertoire as a potential diagnostic marker for celiac disease.

An individual's T cell repertoire is skewed towards some specificities as a result of past antigen exposure and subsequent clonal expansion. Identifying T cell receptor signatures associated with a disease is challenging due to the overall complexity of antigens and polymorphic HLA allotypes. In celiac disease, the antigen epitopes are well characterised and the specific HLA-DQ2-restricted T-cell repertoire associated with the disease has been explored in depth. By investigating T cell receptor repertoires of unsorted lamina propria T cells from 15 individuals, we provide the first proof-of-concept study showing that it could be possible to infer disease state by matching against a priori known disease-associated T cell receptor sequences.

A computational investigation into rate-dependant vectorcardiogram changes due to specific fibrosis patterns in non-ischæmic dilated cardiomyopathy.

Patients with scar-associated fibrotic tissue remodelling are at greater risk of ventricular arrhythmic events, but current methods to detect the presence of such remodelling require invasive procedures. We present here a potential method to detect the presence, location and dimensions of scar using pacing-dependent changes in the vectorcardiogram (VCG). Using a clinically-derived whole-torso computational model, simulations were conducted at both slow and rapid pacing for a variety of scar patterns within the myocardium, with various VCG-derived metrics being calculated, with changes in these metrics being assessed for their ability to discern the presence and size of scar. Our results indicate that differences in the dipole angle at the end of the QRS complex and differences in the QRS area and duration may be used to predict scar properties. Using machine learning techniques, we were also able to predict the location of the scar to high accuracy, using only these VCG-derived rate-dependent changes as input. Such a non-invasive predictive tool for the presence of scar represents a potentially useful clinical tool for identifying patients at arrhythmic risk.

3D Electrophysiological Modeling of Interstitial Fibrosis Networks and Their Role in Ventricular Arrhythmias in Non-Ischemic Cardiomyopathy.

OBJECTIVE: Interstitial fibrosis is a pathological expansion of the heart's inter-cellular collagen matrix. It is a potential complication of nonischemic cardiomyopathy (NICM), a class of diseases involving electrical and or mechanical dysfunction of cardiac tissue not caused by atherosclerosis. Patients with NICM and interstitial fibrosis often suffer from life threatening arrhythmias, which we aim to simulate in this study. METHODS: Our methodology builds on an efficient discrete finite element (DFE) method which allows for the representation of fibrosis as infinitesimal splits in a mesh. We update the DFE method with a local connectivity analysis which creates a consistent topology in the fibrosis network. This is particularly important in nonischemic disease due to the potential presence of large and contiguous fibrotic regions and therefore potentially complex fibrosis networks. RESULTS: In experiments with an image-based model, we demonstrate that our methodology is able to simulate reentrant electrical events associated with cardiac arrhythmias. These reentries depended crucially upon sufficient fibrosis density, which was marked by conduction slowing at high pacing rates. We also created a 2D test-case which demonstrated that fibrosis topologies can modulate transient conduction block, and thereby reentrant activations. CONCLUSION: Ventricular arrhythmias due to interstitial fibrosis in NICM can be efficiently simulated using our methods in medical image based geometries. Furthermore, fibrosis topology modulates transient conduction block, and should be accounted for in electrophysiological simulations with interstitial fibrosis. SIGNIFICANCE: Our study provides methodology which has the potential to predict arrhythmias and to optimize treatments non-invasively for nonischemic cardiomyopathies.

Scar shape analysis and simulated electrical instabilities in a non-ischemic dilated cardiomyopathy patient cohort.

This paper presents a morphological analysis of fibrotic scarring in non-ischemic dilated cardiomyopathy, and its relationship to electrical instabilities which underlie reentrant arrhythmias. Two dimensional electrophysiological simulation models were constructed from a set of 699 late gadolinium enhanced cardiac magnetic resonance images originating from 157 patients. Areas of late gadolinium enhancement (LGE) in each image were assigned one of 10 possible microstructures, which modelled the details of fibrotic scarring an order of magnitude below the MRI scan resolution. A simulated programmed electrical stimulation protocol tested each model for the possibility of generating either a transmural block or a transmural reentry. The outcomes of the simulations were compared against morphological LGE features extracted from the images. Models which blocked or reentered, grouped by microstructure, were significantly different from one another in myocardial-LGE interface length, number of components and entropy, but not in relative area and transmurality. With an unknown microstructure, transmurality alone was the best predictor of block, whereas a combination of interface length, transmurality and number of components was the best predictor of reentry in linear discriminant analysis.

In vivo estimation of elastic heterogeneity in an infarcted human heart.

In myocardial infarction, muscle tissue of the heart is damaged as a result of ceased or severely impaired blood flow. Survivors have an increased risk of further complications, possibly leading to heart failure. Material properties play an important role in determining post-infarction outcome. Due to spatial variation in scarring, material properties can be expected to vary throughout the tissue of a heart after an infarction. In this study we propose a data assimilation technique that can efficiently estimate heterogeneous elastic material properties in a personalized model of cardiac mechanics. The proposed data assimilation is tested on a clinical dataset consisting of regional left ventricular strains and in vivo pressures during atrial systole from a human with a myocardial infarction. Good matches to regional strains are obtained, and simulated equi-biaxial tests are carried out to demonstrate regional heterogeneities in stress-strain relationships. A synthetic data test shows a good match of estimated versus ground truth material parameter fields in the presence of no to low levels of noise. This study is the first to apply adjoint-based data assimilation to the important problem of estimating cardiac elastic heterogeneities in 3-D from medical images.

Fibrosis Microstructure Modulates Reentry in Non-ischemic Dilated Cardiomyopathy: Insights From Imaged Guided 2D Computational Modeling.

Aims: Patients who present with non-ischemic dilated cardiomyopathy (NIDCM) and enhancement on late gadolinium magnetic resonance imaging (LGE-CMR), are at high risk of sudden cardiac death (SCD). Further risk stratification of these patients based on LGE-CMR may be improved through better understanding of fibrosis microstructure. Our aim is to examine variations in fibrosis microstructure based on LGE imaging, and quantify the effect on reentry inducibility and mechanism. Furthermore, we examine the relationship between transmural activation time differences and reentry. Methods and Results: 2D Computational models were created from a single short axis LGE-CMR image, with 401 variations in fibrosis type (interstitial, replacement) and density, as well as presence or absence of reduced conductivity (RC). Transmural activation times (TAT) were measured, as well as reentry incidence and mechanism. Reentries were inducible above specific density thresholds (0.8, 0.6 for interstitial, replacement fibrosis). RC reduced these thresholds (0.3, 0.4 for interstitial, replacement fibrosis) and increased reentry incidence (48 no RC vs. 133 with RC). Reentries were classified as rotor, micro-reentry, or macro-reentry and depended on fibrosis micro-structure. Differences in TAT at coupling intervals 210 and 500ms predicted reentry in the models (sensitivity 89%, specificity 93%). A sensitivity analysis of TAT and reentry incidence showed that these quantities were robust to small changes in the pacing location. Conclusion: Computational models of fibrosis micro-structure underlying areas of LGE in NIDCM provide insight into the mechanisms and inducibility of reentry, and their dependence upon the type and density of fibrosis. Transmural activation times, measured at the central extent of the scar, can potentially differentiate microstructures which support reentry.

High-resolution data assimilation of cardiac mechanics applied to a dyssynchronous ventricle.

Computational models of cardiac mechanics, personalized to a patient, offer access to mechanical information above and beyond direct medical imaging. Additionally, such models can be used to optimize and plan therapies in-silico, thereby reducing risks and improving patient outcome. Model personalization has traditionally been achieved by data assimilation, which is the tuning or optimization of model parameters to match patient observations. Current data assimilation procedures for cardiac mechanics are limited in their ability to efficiently handle high-dimensional parameters. This restricts parameter spatial resolution, and thereby the ability of a personalized model to account for heterogeneities that are often present in a diseased or injured heart. In this paper, we address this limitation by proposing an adjoint gradient-based data assimilation method that can efficiently handle high-dimensional parameters. We test this procedure on a synthetic data set and provide a clinical example with a dyssynchronous left ventricle with highly irregular motion. Our results show that the method efficiently handles a high-dimensional optimization parameter and produces an excellent agreement for personalized models to both synthetic and clinical data.

Adjoint multi-start-based estimation of cardiac hyperelastic material parameters using shear data.

Cardiac muscle tissue during relaxation is commonly modeled as a hyperelastic material with strongly nonlinear and anisotropic stress response. Adapting the behavior of such a model to experimental or patient data gives rise to a parameter estimation problem which involves a significant number of parameters. Gradient-based optimization algorithms provide a way to solve such nonlinear parameter estimation problems with relatively few iterations, but require the gradient of the objective functional with respect to the model parameters. This gradient has traditionally been obtained using finite differences, the calculation of which scales linearly with the number of model parameters, and introduces a differencing error. By using an automatically derived adjoint equation, we are able to calculate this gradient more efficiently, and with minimal implementation effort. We test this adjoint framework on a least squares fitting problem involving data from simple shear tests on cardiac tissue samples. A second challenge which arises in gradient-based optimization is the dependency of the algorithm on a suitable initial guess. We show how a multi-start procedure can alleviate this dependency. Finally, we provide estimates for the material parameters of the Holzapfel and Ogden strain energy law using finite element models together with experimental shear data.