Search

BACKGROUND: Screening for islet autoantibodies in children and adolescents identifies individuals who will later develop type 1 diabetes, allowing patient and family education to prevent diabetic ketoacidosis at onset and to enable consideration of preventive therapies. We aimed to assess whether islet autoantibody screening is effective for predicting type 1 diabetes in adolescents aged 10−18 years with an increased risk of developing type 1 diabetes. METHODS: Data were harmonised from prospective studies from Finland (the Diabetes Prediction and Prevention study), Germany (the BABYDIAB study), and the USA (Diabetes Autoimmunity Study in the Young and the Diabetes Evaluation in Washington study). Autoantibodies against insulin, glutamic acid decarboxylase, and insulinoma-associated protein 2 were measured at each follow-up visit. Children who were lost to follow-up or diagnosed with type 1 diabetes before 10 years of age were excluded. Inverse probability censoring weighting was used to include data from remaining participants. Sensitivity and the positive predictive value of these autoantibodies, tested at one or two ages, to predict type 1 diabetes by the age of 18 years were the main outcomes. FINDINGS: Of 20 303 children with an increased type 1 diabetes risk, 8682 were included for the analysis with inverse probability censoring weighting. 1890 were followed up to 18 years of age or developed type 1 diabetes between the ages of 10 years and 18 years, and their median follow-up was 18·3 years (IQR 14·5–20·3). 442 (23·4%) of 1890 adolescents were positive for at least one islet autoantibody, and 262 (13·9%) developed type 1 diabetes. Time from seroconversion to diabetes diagnosis increased by 0·64 years (95% CI 0·34–0·95) for each 1-year increment of diagnosis age (Pearson’s correlation coefficient 0·88, 95% CI 0·50–0·97, p=0·0020). The median interval between the last prediagnostic sample and diagnosis was 0·3 years (IQR 0·1–1·3) in the 227 participants who were autoantibody positive and 6·8 years (1·6–9·9) for the 35 who were autoantibody negative. Single screening at the age of 10 years was 90% (95% CI 86–95) sensitive, with a positive predictive value of 66% (60–72) for clinical diabetes. Screening at two ages (10 years and 14 years) increased sensitivity to 93% (95% CI 89–97) but lowered the positive predictive value to 55% (49–60). INTERPRETATION: Screening of adolescents at risk for type 1 diabetes only once at 10 years of age for islet autoantibodies was highly effective to detect type 1 diabetes by the age of 18 years, which in turn could enable prevention of diabetic ketoacidosis and participation in secondary prevention trials. FUNDING: JDRF International.

Aims/hypothesis The aim of this study was to explore the utility of islet autoantibody (IAb) levels for the prediction of type 1 diabetes in autoantibody-positive children. Methods Prospective cohort studies in Finland, Germany, Sweden and the USA followed 24,662 children at increased genetic or familial risk of developing islet autoimmunity and diabetes. For the 1403 who developed IAbs (523 of whom developed diabetes), levels of autoantibodies against insulin (IAA), glutamic acid decarboxylase (GADA) and insulinoma-associated antigen-2 (IA-2A) were harmonised for analysis. Diabetes prediction models using multivariate logistic regression with inverse probability censored weighting (IPCW) were trained using 10-fold cross-validation. Discriminative power for disease was estimated using the IPCW concordance index (C index) with 95% CI estimated via bootstrap. Results A baseline model with covariates for data source, sex, diabetes family history, HLA risk group and age at seroconversion with a 10-year follow-up period yielded a C index of 0.61 (95% CI 0.58, 0.63). The performance improved after adding the IAb positivity status for IAA, GADA and IA-2A at seroconversion: C index 0.72 (95% CI 0.71, 0.74). Using the IAb levels instead of positivity indicators resulted in even better performance: C index 0.76 (95% CI 0.74, 0.77). The predictive power was maintained when using the IAb levels alone: C index 0.76 (95% CI 0.75, 0.76). The prediction was better for shorter follow-up periods, with a C index of 0.82 (95% CI 0.81, 0.83) at 2 years, and remained reasonable for longer follow-up periods, with a C index of 0.76 (95% CI 0.75, 0.76) at 11 years. Inclusion of the results of a third IAb test added to the predictive power, and a suitable interval between seroconversion and the third test was approximately 1.5 years, with a C index of 0.78 (95% CI 0.77, 0.78) at 10 years follow-up. Conclusions/interpretation Consideration of quantitative patterns of IAb levels improved the predictive power for type 1 diabetes in IAb-positive children beyond qualitative IAb positivity status. Graphical abstract

The left ventricular outflow tract (LVOT) and ascending aorta are spatially complex, with distinct pathologies and embryologic origins. Prior work examined the genetics of thoracic aortic diameter in a single plane.We sought to elucidate the genetic basis for the diameter of the LVOT, aortic root, and ascending aorta.Using deep learning, we analyzed 2.3 million cardiac magnetic resonance images from 43,317 UK Biobank participants. We computed the diameters of the LVOT, the aortic root, and at 6 locations of ascending aorta. For each diameter, we conducted a genome-wide association study and generated a polygenic score. Finally, we investigated associations between these scores and disease incidence.A total of 79 loci were significantly associated with at least 1 diameter. Of these, 35 were novel, and most were associated with 1 or 2 diameters. A polygenic score of aortic diameter approximately 13 mm from the sinotubular junction most strongly predicted thoracic aortic aneurysm (n = 427,016; mean HR: 1.42 per SD; 95% CI: 1.34-1.50; P = 6.67 × 10We detected distinct genetic loci underpinning the diameters of the LVOT, aortic root, and at several segments of ascending aorta. We spatially defined a region of aorta whose genetics may be most relevant to predicting thoracic aortic aneurysm. We further described a genetic signature that may predispose to aortic stenosis. Understanding genetic contributions to proximal aortic diameter may enable identification of individuals at risk for aortic disease and facilitate prioritization of therapeutic targets.

A fundamental challenge in diagnostics is integrating multiple modalities to develop a joint characterization of physiological state. Using the heart as a model system, we develop a cross-modal autoencoder framework for integrating distinct data modalities and constructing a holistic representation of cardiovascular state. In particular, we use our framework to construct such cross-modal representations from cardiac magnetic resonance images (MRIs), containing structural information, and electrocardiograms (ECGs), containing myoelectric information. We leverage the learned cross-modal representation to (1) improve phenotype prediction from a single, accessible phenotype such as ECGs; (2) enable imputation of hard-to-acquire cardiac MRIs from easy-to-acquire ECGs; and (3) develop a framework for performing genome-wide association studies in an unsupervised manner. Our results systematically integrate distinct diagnostic modalities into a common representation that better characterizes physiologic state.

OBJECTIVE: To estimate risk of progression to stage 3 type 1 diabetes based on varying definition of multiple islet autoantibody positivity (mIA). RESEARCH DESIGN AND METHODS: T1DI is a combined prospective dataset of children from Finland, Germany, Sweden and USA who are at increased genetic risk for type 1 diabetes. Analysis included 16,709 infants-toddlers enrolled by age 2.5 years and comparison between groups using Kaplan-Meier survival analysis. RESULTS: Of 865 (5%) with mIA, 537 (62%) progressed to type 1 diabetes. The 15-year cumulative incidence of diabetes varied from most stringent definition (mIA/Persistent/2: ≥2 islet autoantibodies positive at the same visit with ≥2 antibodies persistent at next visit; 88%, [95% CI: 85-92%]) to least stringent (mIA/Any: positivity for two islet autoantibodies without co-occurring positivity or persistence; 18%, [5-40%]). Progression in mIA/Persistent/2 was significantly higher than all other groups (P CONCLUSIONS: The 15-year risk of progression to type 1 diabetes risk varies markedly from 18-88% based on stringency of mIA definition. While initial categorization identifies highest risk individuals, short-term follow-up over 2 years may help stratify evolving risk especially for those with less stringent definitions of mIA.

The role of race in medical decision-making has been a contentious issue. Insights from history and population genetics suggest considering race as a differentiating marker for medical practices can be influenced by systemic bias, leading to serious errors. This may negatively impact treatment of complex diseases such as cardiovascular disease (CVD). We seek to identify instrumental variables and independently verifiable epidemiological tests of whether diagnoses and treatments impacting severe cardiovascular conditions are racially linked. Using data from the UK Biobank (UKB), we found minimal, non-significant racial differences in log odds ratio (OR) between a range of cardiovascular outcomes such as atrial fibrillation, coronary artery disease, coronary thrombosis, heart failure and cardiac fatality. Genetics classification with respect to principal components vs. racial identification of Black British showed no significant differences in diagnoses or therapeutics for CVD related diseases and their associated comorbidities. However, Black British had significant risk of association with genetically predisposed risk of CVD as captured by polygenic risk scores (PRS) of CVD (OR=1.12; 95%CI:1.034-1.223;p <0.006) as well as in 14 related traits. We used a sub-population based feature selection method to find Townsend Deprivation Index, smoking history, hypertension, PRS for ischemic stroke, low density lipoprotein cholesterol, and type II diabetes as the top features predicting the ethnographic category of Black British with an AUC of 79.5%. Therefore, PRS can be used to understand racial disparities in disease outcome which is otherwise not reflected in clinical factors such as diagnoses outcome status or therapeutics in large observational cohorts such as UKB. PRS yield better predictive power with underrepresented minorities and can improve clinical decision-making.

For any given body mass index (BMI), individuals vary substantially in fat distribution, and this variation may have important implications for cardiometabolic risk. Here, we study disease associations with BMI-independent variation in visceral (VAT), abdominal subcutaneous (ASAT), and gluteofemoral (GFAT) fat depots in 40,032 individuals of the UK Biobank with body MRI. We apply deep learning models based on two-dimensional body MRI projections to enable near-perfect estimation of fat depot volumes (R2 in heldout dataset = 0.978-0.991 for VAT, ASAT, and GFAT). Next, we derive BMI-adjusted metrics for each fat depot (e.g. VAT adjusted for BMI, VATadjBMI) to quantify local adiposity burden. VATadjBMI is associated with increased risk of type 2 diabetes and coronary artery disease, ASATadjBMI is largely neutral, and GFATadjBMI is associated with reduced risk. These results – describing three metabolically distinct fat depots at scale – clarify the cardiometabolic impact of BMI-independent differences in body fat distribution.

Summary Background Parkinson's disease is heterogeneous in symptom presentation and progression. Increased understanding of both aspects can enable better patient management and improve clinical trial design. Previous approaches to modelling Parkinson's disease progression assumed static progression trajectories within subgroups and have not adequately accounted for complex medication effects. Our objective was to develop a statistical progression model of Parkinson's disease that accounts for intra-individual and inter-individual variability and medication effects. Methods In this longitudinal data study, data were collected for up to 7-years on 423 patients with early Parkinson's disease and 196 healthy controls from the Parkinson's Progression Markers Initiative (PPMI) longitudinal observational study. A contrastive latent variable model was applied followed by a novel personalised input-output hidden Markov model to define disease states. Clinical significance of the states was assessed using statistical tests on seven key motor or cognitive outcomes (mild cognitive impairment, dementia, dyskinesia, presence of motor fluctuations, functional impairment from motor fluctuations, Hoehn and Yahr score, and death) not used in the learning phase. The results were validated in an independent sample of 610 patients with Parkinson's disease from the National Institute of Neurological Disorders and Stroke Parkinson's Disease Biomarker Program (PDBP). Findings PPMI data were download July 25, 2018, medication information was downloaded on Sept 24, 2018, and PDBP data were downloaded between June 15 and June 24, 2020. The model discovered eight disease states, which are primarily differentiated by functional impairment, tremor, bradykinesia, and neuropsychiatric measures. State 8, the terminal state, had the highest prevalence of key clinical outcomes including 18 (95%) of 19 recorded instances of dementia. At study outset 4 (1%) of 333 patients were in state 8 and 138 (41%) of 333 patients reached stage 8 by year 5. However, the ranking of the starting state did not match the ranking of reaching state 8 within 5 years. Overall, patients starting in state 5 had the shortest time to terminal state (median 2·75 [95% CI 1·75–4·25] years). Interpretation We developed a statistical progression model of early Parkinson's disease that accounts for intra-individual and inter-individual variability and medication effects. Our predictive model discovered non-sequential, overlapping disease progression trajectories, supporting the use of non-deterministic disease progression models, and suggesting static subtype assignment might be ineffective at capturing the full spectrum of Parkinson's disease progression. Funding Michael J Fox Foundation.

State-of-the-art genetic risk interpretation for a common complex disease such as coronary artery disease (CAD) requires assessment for both monogenic variants-such as those related to familial hypercholesterolemia-as well as the cumulative impact of many common variants, as quantified by a polygenic score.The objective of the study was to describe a combined monogenic and polygenic CAD risk assessment program and examine its impact on patient understanding and changes to clinical management.Study participants attended an initial visit in a preventive genomics clinic and a disclosure visit to discuss results and recommendations, primarily via telemedicine. Digital postdisclosure surveys and chart review evaluated the impact of disclosure.There were 60 participants (mean age 51 years, 37% women, 72% with no known CAD), including 30 (50%) referred by their cardiologists and 30 (50%) self-referred. Two (3%) participants had a monogenic variant pathogenic for familial hypercholesterolemia, and 19 (32%) had a high polygenic score in the top quintile of the population distribution. In a postdisclosure survey, both the genetic test report (in 80% of participants) and the discussion with the clinician (in 89% of participants) were ranked as very or extremely helpful in understanding the result. Of the 42 participants without CAD, 17 or 40% had a change in management, including statin initiation, statin intensification, or coronary imaging.Combined monogenic and polygenic assessments for CAD risk provided by preventive genomics clinics are beneficial for patients and result in changes in management in a significant portion of patients.

Inter-individual variation in fat distribution is increasingly recognized as clinically important but is not routinely assessed in clinical practice, in part because medical imaging has not been practical to deploy at scale for this task. Here, we report a deep learning model trained on an individual’s body shape outline—or “silhouette” —that enables accurate estimation of specific fat depots of interest, including visceral (VAT), abdominal subcutaneous (ASAT), and gluteofemoral (GFAT) adipose tissue volumes, and VAT/ASAT ratio. Two-dimensional coronal and sagittal silhouettes are constructed from whole-body magnetic resonance images in 40,032 participants of the UK Biobank and used as inputs for a convolutional neural network to predict each of these quantities. Mean age of the study participants is 65 years and 51% are female. A cross-validated deep learning model trained on silhouettes enables accurate estimation of VAT, ASAT, and GFAT volumes (R2: 0.88, 0.93, and 0.93, respectively), outperforming a comparator model combining anthropometric and bioimpedance measures (ΔR2 = 0.05–0.13). Next, we study VAT/ASAT ratio, a nearly body-mass index (BMI)—and waist circumference-independent marker of metabolically unhealthy fat distribution. While the comparator model poorly predicts VAT/ASAT ratio (R2: 0.17–0.26), a silhouette-based model enables significant improvement (R2: 0.50–0.55). Increased silhouette-predicted VAT/ASAT ratio is associated with increased risk of prevalent and incident type 2 diabetes and coronary artery disease independent of BMI and waist circumference. These results demonstrate that body silhouette images can estimate important measures of fat distribution, laying the scientific foundation for scalable population-based assessment.

A fundamental challenge in diagnostics is integrating multiple modalities to develop a joint characterization of physiological state. Using the heart as a model system, we develop a cross-modal autoencoder framework for integrating distinct data modalities and constructing a holistic representation of cardio-vascular state. In particular, we use our framework to construct such cross-modal representations from cardiac magnetic resonance images (MRIs), containing structural information, and electrocardiograms (ECGs), containing myoelectric information. We leverage the learned cross-modal representation to (1) improve phenotype prediction from a single, accessible phenotype such as ECGs; (2) enable imputation of hard-to-acquire cardiac MRIs from easy-to-acquire ECGs; and (3) develop a framework for performing genome-wide association studies in an unsupervised manner. Our results provide a framework for integrating distinct diagnostic modalities into a common representation that better characterizes physiologic state.

With availability of voluminous sets of observational data, an empirical paradigm to screen for drug repurposing opportunities (i.e., beneficial effects of drugs on nonindicated outcomes) is feasible. In this article, we use a linked claims and electronic health record database to comprehensively explore repurposing effects of antihypertensive drugs. We follow a target trial emulation framework for causal inference to emulate randomized controlled trials estimating confounding adjusted effects of antihypertensives on each of 262 outcomes of interest. We then fit hierarchical models to the results as a form of postprocessing to account for multiple comparisons and to sift through the results in a principled way. Our motivation is twofold. We seek both to surface genuinely intriguing drug repurposing opportunities and to elucidate through a real application some study design decisions and potential biases that arise in this context.

The central task of causal inference is to remove (via statistical adjustment) confounding bias that would be present in naive unadjusted comparisons of outcomes in different treatment groups. Statistical adjustment can roughly be broken down into two steps. In the first step, the researcher selects some set of variables to adjust for. In the second step, the researcher implements a causal inference algorithm to adjust for the selected variables and estimate the average treatment effect. In this paper, we use a simulation study to explore the operating characteristics and robustness of state-of-the-art methods for step two (statistical adjustment for selected variables) when step one (variable selection) is performed in a realistically sub-optimal manner. More specifically, we study the robustness of a cross-fit machine learning based causal effect estimator to the presence of extraneous variables in the adjustment set. The take-away for practitioners is that there is value to, if possible, identifying a small sufficient adjustment set using subject matter knowledge even when using machine learning methods for adjustment.

To date, there have been 180 million confirmed cases of COVID-19, with more than 3.8 million deaths, reported to WHO worldwide. In this paper we address the problem of understanding the host genome’s influence, in concert with clinical variables, on the severity of COVID-19 manifestation in the patient. Leveraging positive-unlabeled machine learning algorithms coupled with RubricOE, a state-of-the-art genomic analysis framework, on UK BioBank data we extract novel insights on the complex interplay. The algorithm is also sensitive enough to detect the changing influence of the emergent B.1.1.7 SARS-CoV-2 (alpha) variant on disease severity, and, changing treatment protocols. The genomic component also implicates biological pathways that can help in understanding the disease etiology. Our work demonstrates that it is possible to build a robust and sensitive model despite significant bias, noise and incompleteness in both clinical and genomic data by a careful interleaving of clinical and genomic methodologies.

Prediction models are commonly used to estimate risk for cardiovascular diseases, to inform diagnosis and management. However, performance may vary substantially across relevant subgroups of the population. Here we investigated heterogeneity of accuracy and fairness metrics across a variety of subgroups for risk prediction of two common diseases: atrial fibrillation (AF) and atherosclerotic cardiovascular disease (ASCVD). We calculated the Cohorts for Heart and Aging in Genomic Epidemiology Atrial Fibrillation (CHARGE-AF) score for AF and the Pooled Cohort Equations (PCE) score for ASCVD in three large datasets: Explorys Life Sciences Dataset (Explorys, n = 21,809,334), Mass General Brigham (MGB, n = 520,868), and the UK Biobank (UKBB, n = 502,521). Our results demonstrate important performance heterogeneity across subpopulations defined by age, sex, and presence of preexisting disease, with fairly consistent patterns across both scores. For example, using CHARGE-AF, discrimination declined with increasing age, with a concordance index of 0.72 [95% CI 0.72–0.73] for the youngest (45–54 years) subgroup to 0.57 [0.56–0.58] for the oldest (85–90 years) subgroup in Explorys. Even though sex is not included in CHARGE-AF, the statistical parity difference (i.e., likelihood of being classified as high risk) was considerable between males and females within the 65–74 years subgroup with a value of − 0.33 [95% CI − 0.33 to − 0.33]. We also observed weak discrimination (i.e.

Diabetes mellitus, commonly known as diabetes, is a chronic disease that often results in multiple complications. Risk prediction of diabetes complications is critical for healthcare professionals to design personalized treatment plans for patients in diabetes care for improved outcomes. In this paper, focusing on Type 2 diabetes mellitus (T2DM), we study the risk of developing complications after the initial T2DM diagnosis from longitudinal patient records. We propose a novel multi-task learning approach to simultaneously model multiple complications where each task corresponds to the risk modeling of one complication. Specifically, the proposed method strategically captures the relationships (1) between the risks of multiple T2DM complications, (2) between different risk factors, and (3) between the risk factor selection patterns, which assumes similar complications have similar contributing risk factors. The method uses coefficient shrinkage to identify an informative subset of risk factors from high-dimensional data, and uses a hierarchical Bayesian framework to allow domain knowledge to be incorporated as priors. The proposed method is favorable for healthcare applications because in addition to improved prediction performance, relationships among the different risks and among risk factors are also identified. Extensive experimental results on a large electronic medical claims database show that the proposed method outperforms state-of-the-art models by a significant margin. Furthermore, we show that the risk associations learned and the risk factors identified lead to meaningful clinical insights.

BackgroundInter-individual variation in fat distribution is increasingly recognized as clinically important but is not routinely assessed in clinical practice because quantification requires medical imaging.ObjectivesWe hypothesized that a deep learning model trained on an individual’s body shape outline – or “silhouette” – would enable accurate estimation of specific fat depots, including visceral (VAT), abdominal subcutaneous (ASAT), and gluteofemoral (GFAT) adipose tissue volumes, and VAT/ASAT ratio. We additionally set out to study whether silhouette-estimated VAT/ASAT ratio may stratify risk of cardiometabolic diseases independent of body mass index (BMI) and waist circumference.MethodsTwo-dimensional coronal and sagittal silhouettes were constructed from whole-body magnetic resonance images in 40,032 participants of the UK Biobank and used to train a convolutional neural network to predict VAT, ASAT, and GFAT volumes, and VAT/ASAT ratio. Logistic and Cox regressions were used to determine the independent association of silhouette-predicted VAT/ASAT ratio with type 2 diabetes and coronary artery disease.ResultsMean age of the study participants was 65 years and 51% were female. A deep learning model trained on silhouettes enabled accurate estimation of VAT, ASAT, and GFAT volumes (R2: 0.88, 0.93, and 0.93, respectively), outperforming a comparator model combining anthropometric and bioimpedance measures (ΔR2 = 0.05-0.13). Next, we studied VAT/ASAT ratio, a nearly BMI- and waist circumference-independent marker of unhealthy fat distribution. While the comparator model poorly predicted VAT/ASAT ratio (R2: 0.17-0.26), a silhouette-based model enabled significant improvement (R2: 0.50-0.55). Silhouette-predicted VAT/ASAT ratio was associated with increased prevalence of type 2 diabetes and coronary artery disease.ConclusionsBody silhouette images can estimate important measures of fat distribution, laying the scientific foundation for population-based assessment.

Disease risk models can identify high-risk patients and help clinicians provide more personalized care. However, risk models developed on one dataset may not generalize across diverse subpopulations of patients in different datasets and may have unexpected performance. It is challenging for clinical researchers to inspect risk models across different subgroups without any tools. Therefore, we developed an interactive visualization system called RMExplorer (Risk Model Explorer) to enable interactive risk model assessment. Specifically, the system allows users to define subgroups of patients by selecting clinical, demographic, or other characteristics, to explore the performance and fairness of risk models on the subgroups, and to understand the feature contributions to risk scores. To demonstrate the usefulness of the tool, we conduct a case study, where we use RMExplorer to explore three atrial fibrillation risk models by applying them to the UK Biobank dataset of 445,329 individuals. RMExplorer can help researchers to evaluate the performance and biases of risk models on subpopulations of interest in their data., Comment: IEEE VIS 2022 Short

The first surge of the COVID-19 pandemic entirely altered healthcare delivery. Whether this also altered the receipt of high- and low-value care is unknown.To test the association between the April through June 2020 surge of COVID-19 and various high- and low-value care measures to determine how the delivery of care changed.Difference in differences analysis, examining the difference in quality measures between the April through June 2020 surge quarter and the January through March 2020 quarter with the same 2 quarters' difference the year prior.Adults in the MarketScan® Commercial Database and Medicare Supplemental Database.Fifteen low-value and 16 high-value quality measures aggregated into 8 clinical quality composites (4 of these low-value).We analyzed 9,352,569 adults. Mean age was 44 years (SD, 15.03), 52% were female, and 75% were employed. Receipt of nearly every type of low-value care decreased during the surge. For example, low-value cancer screening decreased 0.86% (95% CI, -1.03 to -0.69). Use of opioid medications for back and neck pain (DiD +0.94 [95% CI, +0.82 to +1.07]) and use of opioid medications for headache (DiD +0.38 [95% CI, 0.07 to 0.69]) were the only two measures to increase. Nearly all high-value care measures also decreased. For example, high-value diabetes care decreased 9.75% (95% CI, -10.79 to -8.71).The first COVID-19 surge was associated with receipt of less low-value care and substantially less high-value care for most measures, with the notable exception of increases in low-value opioid use.

Introduction: Pathogenic DNA variants associated with inherited cardiomyopathies are recognized as clinically important and actionable when identified, leading some clinicians to recommend population-wide genomic screening. The prevalence and clinical importance of such variants within the context of contemporary clinical care warrant further study. Methods: Using whole exome sequencing data from participants in Atherosclerosis Risk in Communities (ARIC) and the UK Biobank (UKBB), observed DNA sequence variants in any of 12 genes ( ACTC1, GLA, LMNA, MYBPC3, MYH7, MYL2, MYL3, PRKAG2, TNNI3, TNNT2, TPM1, and TTN ) known to be causative of inherited cardiomyopathies with accepted preventative or therapeutic interventions were classified as pathogenic or likely pathogenic by a clinical laboratory geneticist blinded to case status per recommendations from the American College of Medical Genetics. The relationship of carrier status to risk of death or cardiovascular disease was assessed. Results: Among 9,667 ARIC participants (mean [SD] age 54 [5.7] years; 43.8% female) and 49,744 UKBB participants (mean [SD] age 57.1 [8.0] years; 54.6% female), 68 (0.70%) and 362 (0.73%) harbored an actionable pathogenic or likely pathogenic variant associated with cardiomyopathy, respectively. Carriers of these variants were not reliably identifiable by imaging or electrocardiography signatures. Presence of these variants was associated with 1.7- to 1.8-fold increased risk of heart failure, 2.1- to 2.9-fold increased risk of atrial fibrillation, and 1.6- to 1.8-fold increased risk of all-cause mortality across studies. Conclusions: The findings suggest that approximately 0.7% of middle-aged adult population in ARIC and the UKBB harbored a pathogenic variant associated with cardiomyopathy. Carriers of these variants would be difficult to identify within routine clinical practice but suffer from increased risk of cardiovascular disease and all-cause mortality.

OBJECTIVE To use islet autoantibody titers to improve the estimation of future type 1 diabetes risk in children. RESEARCH DESIGN AND METHODS Prospective cohort studies in Finland, Germany, Sweden, and the U.S. followed 24,662 children at increased genetic or familial risk to develop islet autoimmunity and diabetes. For 1,604 children with confirmed positivity, titers of autoantibodies against insulin (IAA), GAD antibodies (GADA), and insulinoma-associated antigen 2 (IA-2A) were harmonized for diabetes risk analyses. RESULTS Survival analysis from time of confirmed positivity revealed markedly different 5-year diabetes risks associated with IAA (n = 909), GADA (n = 1,076), and IA-2A (n = 714), when stratified by quartiles of titer, ranging from 19% (GADA 1st quartile) to 60% (IA-2A 4th quartile). The minimum titer associated with a maximum difference in 5-year risk differed for each autoantibody, corresponding to the 58.6th, 52.4th, and 10.2nd percentile of children specifically positive for each of IAA, GADA, and IA-2A, respectively. Using these autoantibody type-specific titer thresholds in the 1,481 children with all autoantibodies tested, the 5-year risk conferred by single (n = 954) and multiple (n = 527) autoantibodies could be stratified from 6 to 75% (P < 0.0001). The thresholds effectively identified children with a ≥50% 5-year risk when considering age-specific autoantibody screening (57–65% positive predictive value and 56–74% sensitivity for ages 1–5 years). Multivariable analysis confirmed the significance of associations between the three autoantibody titers and diabetes risk, informing a childhood risk surveillance strategy. CONCLUSIONS This study defined islet autoantibody type-specific titer thresholds that significantly improved type 1 diabetes risk stratification in children.

Myocardial interstitial fibrosis is a common thread in multiple cardiovascular diseases including heart failure, atrial fibrillation, conduction disease and sudden cardiac death. To investigate the biologic pathways that underlie interstitial fibrosis in the human heart, we developed a machine learning model to measure myocardial T1 time, a marker of myocardial interstitial fibrosis, in 42,654 UK Biobank participants. Greater T1 time was associated with impaired glucose metabolism, systemic inflammation, renal disease, aortic stenosis, cardiomyopathy, heart failure, atrial fibrillation and conduction disease. In genome-wide association analysis, we identified 12 independent loci associated with native myocardial T1 time with evidence of high genetic correlation between the interventricular septum and left ventricle free wall (r2g = 0.82). The identified loci implicated genes involved in glucose homeostasis (SLC2A12), iron homeostasis (HFE, TMPRSS6), tissue repair (ADAMTSL1, VEGFC), oxidative stress (SOD2), cardiac hypertrophy (MYH7B) and calcium signaling (CAMK2D). Transcriptome-wide association studies highlighted the role of expression of ADAMTSL1 and SLC2A12 in human cardiac tissue in modulating myocardial tissue characteristics and interstitial fibrosis. Harnessing machine learning to perform large-scale phenotyping of interstitial fibrosis in the human heart, our results yield novel insights into biologically relevant pathways for myocardial fibrosis and prioritize investigation of pathways for the development of anti-fibrotic therapies.

BackgroundThe left ventricular outflow tract (LVOT) and ascending aorta are spatially complex, with distinct pathologies and embryologic origins. Prior work examined genetics of thoracic aortic diameter in a single plane. We sought to elucidate the genetic basis for the diameter of the LVOT, the aortic root, and the ascending aorta.MethodsWe used deep learning to analyze 2.3 million cardiac magnetic resonance images from 43,317 UK Biobank participants. We computed the diameters of the LVOT, the aortic root, and at six locations in the ascending aorta. For each diameter, we conducted a genome-wide association study and generated a polygenic score. Finally, we investigated associations between these polygenic scores and disease incidence.Results79 loci were significantly associated with at least one diameter. Of these, 35 were novel, and a majority were associated with one or two diameters. A polygenic score of aortic diameter approximately 13mm from the sinotubular junction most strongly predicted thoracic aortic aneurysm in UK Biobank participants (n=427,016; HR=1.42 per standard deviation; CI=1.34-1.50, P=6.67×10−21). A polygenic score predicting a smaller aortic root was predictive of aortic stenosis (n=426,502; HR=1.08 per standard deviation; CI=1.03-1.12, P=5×10−6).ConclusionsWe detected distinct common genetic loci underpinning the diameters of the LVOT, the aortic root, and at several segments in the ascending aorta. We spatially defined a region of aorta whose genetics may be most relevant to predicting thoracic aortic aneurysm. We further described a genetic signature that may predispose to aortic stenosis. Understanding the genetic contributions to the diameter of the proximal aorta may enable identification of individuals at risk for life-threatening aortic disease and facilitate prioritization of therapeutic targets.

As the largest conduit vessel, the aorta is responsible for the conversion of phasic systolic inflow from ventricular ejection into more continuous blood delivery to peripheral arteries. Distension during systole and recoil during diastole conserves ventricular energy and is enabled by the specialized composition of the aortic extracellular matrix. Aortic distensibility decreases with age and prematurely in vascular disease. To discover genetic determinants of aortic distensibility we trained a deep learning model to quantify aortic size throughout the cardiac cycle and calculate aortic distensibility and aortic strain in 42,342 participants in the UK Biobank with available cardiac magnetic resonance imaging. In up to 40,028 participants with genetic data, common variant analysis identified 12 and 26 loci for ascending and 11 and 21 loci for descending aortic distensibility and strain, respectively. Of the newly identified loci, 22 were specific to strain or distensibility and were not identified in a thoracic aortic diameter GWAS within the same samples. Loci associated with both aortic diameter and aortic strain or distensibility demonstrated a consistent, inverse directionality. Transcriptome-wide analyses, rare-variant burden tests, and analyses of gene expression in single nucleus RNA sequencing of human aorta were performed to prioritize genes at individual loci. Loci highlighted multiple genes involved in elastogenesis, matrix degradation, and extracellular polysaccharide generation. Characterization of the genetic determinants of aortic function may provide novel targets for medical intervention in aortic disease.

Prediction models are commonly used to estimate risk for cardiovascular diseases; however, performance may vary substantially across relevant subgroups of the population. Here we investigated the variability of performance and fairness across a variety of subgroups for risk prediction of two common diseases, atherosclerotic cardiovascular disease (ASCVD) and atrial fibrillation (AF). We calculated the Cohorts for Heart and Aging in Genomic Epidemiology Atrial Fibrillation (CHARGE-AF) for AF and the Pooled Cohort Equations (PCE) score for ASCVD in three large data sets: Explorys Life Sciences Dataset (Explorys, n = 21,809,334), Mass General Brigham (MGB, n = 520,868), and the UK Biobank (UKBB, n = 502,521). Our results demonstrate important performance heterogeneity of established cardiovascular risk scores across subpopulations defined by age, sex, and presence of preexisting disease. For example, in CHARGE-AF, discrimination declined with increasing age, with concordance index of 0.72 [ 95% CI, 0.72–0.73 ] for the youngest (45–54y) subgroup to 0.57 [ 0.56–0.58 ], for the oldest (85–90y) subgroup in Explorys. The statistical parity difference (i.e., likelihood of being classified as high risk) was considerable between males and females within the 65–74y subgroup with a value of -0.33 [ 95% CI, -0.33–-0.33 ]. We observed also that large segments of the population suffered from both decreased discrimination (i.e.

For any given level of overall adiposity – as commonly quantified by body mass index (BMI) within clinical practice – individuals vary considerably in fat distribution. We and others have noted that increased visceral fat (VAT) is associated with increased cardiometabolic risk, while gluteofemoral fat (GFAT) may be protective. Familial partial lipodystrophy (FPLD) – often caused by rare variants in the LMNA gene – represents an extreme example of this paradigm, leading to a severe shift to visceral fat with subsequent insulin resistance and adverse metabolic profile. By contrast, the inherited basis of body fat distribution in the broader population is not fully understood. Here, we studied up to 38,965 UK Biobank participants with VAT, abdominal subcutaneous (ASAT), and GFAT volumes precisely quantified using abdominal MRI. Because genetic associations with these raw depot volumes were largely driven by variants known to affect BMI, we next studied six phenotypes of local adiposity: VAT adjusted for BMI (VATadjBMI), ASATadjBMI, GFATadjBMI, VAT/ASAT, VAT/GFAT, and ASAT/GFAT. We identify 178 unique loci associated with at least one adiposity trait, including 29 newly-identified loci. Rare variant association studies extend prior evidence of association for PDE3B as an important modulator of fat distribution. Sex-specific analyses of local adiposity traits noted overall higher estimated heritability in females, increased effect sizes for identified loci, and 25 female-specific associations. Individuals in the extreme tails of fat distribution phenotypes were highly enriched for predisposing common variants, as quantified using polygenic scores. Taking GFATadjBMI as an example, individuals with extreme values were 3.8-fold (95%CI 2.8 to 5.2) more likely to have a polygenic score within the top 5% of the distribution. These results – using more precise and BMI-independent measures of local adiposity – confirm fat distribution as a highly heritable trait with important implications for cardiometabolic health outcomes.

Background: Individuals of South Asian ancestry represent 23% of the global population, corresponding to 1.8 billion people, and have substantially higher risk of atherosclerotic cardiovascular disease compared with most other ethnicities. US practice guidelines now recognize South Asian ancestry as an important risk-enhancing factor. The magnitude of enhanced risk within the context of contemporary clinical care, the extent to which it is captured by existing risk estimators, and its potential mechanisms warrant additional study. Methods: Within the UK Biobank prospective cohort study, 8124 middle-aged participants of South Asian ancestry and 449 349 participants of European ancestry who were free of atherosclerotic cardiovascular disease at the time of enrollment were examined. The relationship of ancestry to risk of incident atherosclerotic cardiovascular disease—defined as myocardial infarction, coronary revascularization, or ischemic stroke—was assessed with Cox proportional hazards regression, along with examination of a broad range of clinical, anthropometric, and lifestyle mediators. Results: The mean age at study enrollment was 57 years, and 202 405 (44%) were male. Over a median follow-up of 11 years, 554 of 8124 (6.8%) individuals of South Asian ancestry experienced an atherosclerotic cardiovascular disease event compared with 19 756 of 449 349 (4.4%) individuals of European ancestry, corresponding to an adjusted hazard ratio of 2.03 (95% CI, 1.86–2.22; P 2-fold higher observed risk, the predicted 10-year risk of cardiovascular disease according to the American Heart Association/American College of Cardiology Pooled Cohort equations and QRISK3 equations was nearly identical for individuals of South Asian and European ancestry. Adjustment for a broad range of clinical, anthropometric, and lifestyle risk factors led to only modest attenuation of the observed hazard ratio to 1.45 (95% CI, 1.28–1.65, P Conclusions: Within a large prospective study, South Asian individuals had substantially higher risk of atherosclerotic cardiovascular disease compared with individuals of European ancestry, and this risk was not captured by the Pooled Cohort Equations.

While associations between the type and number of islet autoantibodies and progression to type 1 diabetes (T1D) have been reported, the effect of titer values is less well understood. We aim to quantify the ability of autoantibody titers at seroconversion to improve T1D onset prediction. Prospective cohorts in Finland, Germany, Sweden, and the US have followed 24662 children at increased genetic risk for development of islet autoantibodies and T1D. For the 1400 who seroconverted (523 developed T1D), the titers of insulin autoantibodies (IAA), glutamic acid decarboxylase autoantibodies (GADA), and insulinoma antigen-2 autoantibodies (IA2A) at time of initial and confirmed seroconversion, i.e., the respective first and consecutive second autoantibody-positive serum sample, were normalized (to log multiples of upper limit of normal) and analyzed. Prediction models using multivariate logistic regression with inverse probability censored weighting (IPCW) were trained using 10-fold cross validation. Discriminative power for disease was estimated using the IPCW concordance index (c-index) with 95% confidence intervals estimated via bootstrap. Multivariate Cox proportional hazards models were used to quantify the impact of the autoantibody titers. A baseline model with covariates for data source, sex, HLA-DR/DQ genotype, and age at initial and confirmed seroconversion had a performance of 64 c-index; 95% CI 62-65. Significant improvement was observed after adding IAA, GADA, IA2A positivity indicators at initial and confirmed seroconversion (74; 72-75). Adding the corresponding autoantibody titers resulted in significant additional gains (76; 75-76). Adjusted hazard ratios (HR) from the Cox model showed that autoantibody titers at confirmed seroconversion were predictive of diabetes (HR 1.29; 95% CI 1.20-1.38 for IAA, 1.12; 1.07-1.18 for GADA, and 1.19; 1.15-1.25 for IA2A, all p Islet autoantibody titers at seroconversion improve T1D prediction. Disclosure K. Ng: Employee; Self; IBM. V. Anand: None. R. Veijola: None. M. Maziarz: None. K. Waugh: None. W. Hagopian: None. P. Achenbach: None. T1di study group: n/a. Funding JDRF (1-IND-2019-717-I-X)

Truncating variants in the gene encoding titin (TTNtv) are the most commonly identified pathogenic variants in cross-sectional studies of patients with dilated cardiomyopathy or atrial fibrillation ([1][1],[2][2]). In principle, gene sequencing to identify individuals who harbor a TTNtv prior to

IMPORTANCE: Pathogenic variants associated with inherited cardiomyopathy are recognized as important and clinically actionable when identified, leading some clinicians to recommend population-wide genomic screening. OBJECTIVE: To determine the prevalence and clinical importance of pathogenic variants associated with inherited cardiomyopathy within the context of contemporary clinical care. DESIGN, SETTING, AND PARTICIPANTS: This was a genetic association study of participants in Atherosclerosis in Risk Communities (ARIC), recruited from 1987 to 1989, with median follow-up of 27 years, and the UK Biobank, recruited from 2006 to 2010, with median follow-up of 10 years. ARIC participants were recruited from 4 sites across the US. UK Biobank participants were recruited from 22 sites across the UK. Participants in the US were of African and European ancestry; those in the UK were of African, East Asian, South Asian, and European ancestry. Statistical analyses were performed between August 1, 2021, and February 9, 2022. EXPOSURES: Rare genetic variants predisposing to inherited cardiomyopathy. MAIN OUTCOMES AND MEASURES: Pathogenicity of observed DNA sequence variants in sequenced exomes of 13 genes (ACTC1, FLNC, GLA, LMNA, MYBPC3, MYH7, MYL2, MYL3, PRKAG2, TNNI3, TNNT2, TPM1, and TTN) associated with inherited cardiomyopathies were classified by a blinded clinical geneticist per American College of Medical Genetics recommendations. Incidence of all-cause mortality, heart failure, and atrial fibrillation were determined. Cardiac magnetic resonance imaging, echocardiography, and electrocardiogram measures were assessed in a subset of participants. RESULTS: A total of 9667 ARIC participants (mean [SD] age, 54.0 [5.7] years; 4232 women [43.8%]; 2658 African [27.5%] and 7009 European [72.5%] ancestry) and 49 744 UK Biobank participants (mean [SD] age, 57.1 [8.0] years; 27 142 women [54.5%]; 1006 African [2.0%], 173 East Asian [0.3%], 939 South Asian [1.9%], and 46 449 European [93.4%] European ancestry) were included in the study. Of those, 59 participants (0.61%) in ARIC and 364 participants (0.73%) in UK Biobank harbored an actionable pathogenic or likely pathogenic variant associated with dilated or hypertrophic cardiomyopathy. Carriers of these variants were not reliably identifiable by imaging. However, the presence of these variants was associated with increased risk of heart failure (hazard ratio [HR], 1.7; 95% CI, 1.1-2.8), atrial fibrillation (HR, 2.9; 95% CI, 1.9-4.5), and all-cause mortality (HR, 1.5; 95% CI, 1.1-2.2) in ARIC. Similar risk patterns were observed in the UK Biobank. CONCLUSIONS AND RELEVANCE: Results of this genetic association study suggest that approximately 0.7% of study participants harbored a pathogenic variant associated with inherited cardiomyopathy. These variant carriers would be challenging to identify within clinical practice without genetic testing but are at increased risk for cardiovascular disease and all-cause mortality.

Amyotrophic Lateral Sclerosis (ALS) is a neurodegenerative disease that is complex in its onset, pattern of spread, and disease progression. The heterogeneity of ALS makes it extremely challenging to determine if a disease modifying therapy is effectively slowing progression. While accurately modeling ALS progression is critical to developing therapeutics, current computational methods fail to capture the complexity of disease progression. We aimed to robustly characterize disease progression patterns in ALS.We obtained data from four clinical cohorts that cover more than 3,500 patients and include both observational and clinical trial studies. To determine whether there were common patterns of disease progression, we developed an approach based on a Mixture of Gaussian Processes (MoGP) to model longitudinal clinical data. Our approach automatically identifies clusters of patients who show similar disease progression patterns, modeling their average trajectory and the spread of the distribution in each cluster. Importantly, the method does not require any prior knowledge of the expected number of clusters.The MoGP approach revealed that ALS progression, as measured using the ALS functional rating scale (ALSFRS-R) or forced vital capacity, is often non-linear with periods of stable disease preceded or followed by rapid decline. Patterns of progression in ALSFRS-R were robust to sparse data. When at least one year of longitudinal data were available, MoGP predictions were significantly more accurate than linear models, which are commonly used in clinical trials. Progression patterns were consistent across different cohorts despite differences in the frequency of data collection and the lengths of follow-up periods. We further showed that clusters identified from one large, publicly available study population could be used to stratify unseen participants in other studies. We also showed that these progression trajectories correspond with survival outcomes.This work highlights the importance of modeling nonlinear disease progression for developing more advanced clinical trial endpoint analysis models. In ALS, sporadic, rapid decline (“functional cliffs”) and sigmoidal patterns in disease progression in untreated patients may obscure detection of therapeutic efficacy if linear models are used. We provide a pre-trained computational model of observed clinical patterns that can be used by others to analyze new ALS patient cohorts. We expect that the MoGP approach can also be applied to additional ALS outcome measures and to other progressive diseases. Our results provide a critical advance in characterizing the complex disease progression patterns of ALS.

BackgroundThe clinical implications of BMI-independent variation in fat distribution are not fully understood.MethodsWe studied MRI imaging data of 40,032 UK Biobank participants. Using previously quantified visceral (VAT), abdominal subcutaneous (ASAT), and gluteofemoral (GFAT) adipose tissue volume in up to 9,041 to train convolutional neural networks (CNNs), we quantified these depots in the remainder of the participants. We derived new metrics for each adipose depot – fully independent of BMI – by quantifying deviation from values predicted by BMI (e.g. VAT adjusted for BMI, VATadjBMI) and determined associations with cardiometabolic diseases.ResultsCNNs based on two-dimensional projection images enabled near-perfect estimation of VAT, ASAT, and GFAT, with r2 in a holdout testing dataset (r2 = 0.978-0.991). Using the newly derived measures of local adiposity – residualized based on BMI – we note marked heterogeneity in associations with cardiometabolic diseases. Taking presence of type 2 diabetes as an example, VATadjBMI was associated with significantly increased risk (odds ratio per standard deviation increase (OR/SD) 1.49; 95%CI: 1.43-1.55), while ASATadjBMI was largely neutral (OR/SD 1.08; 95%CI: 1.03-1.14) and GFATadjBMI conferred protection (OR/SD 0.75; 95%CI: 0.71-0.79). Similar patterns were observed for coronary artery disease.ConclusionsDeep learning models trained on a simplified MRI input enable near perfect quantification of VAT, ASAT, and GFAT. For any given BMI, measures of local adiposity have variable and divergent associations with cardiometabolic diseases.

Type 1 diabetes (T1D) is a chronic autoimmune disease that affects about 1 in 300 children and up to 1 in 100 adults during their life-time(1). Improvements in early prediction of T1D onset may help prevent diagnosis for diabetic ketoacidosis, a serious complication often associated with a missed or delayed T1D diagnosis. In addition to genetic factors, progression to T1D is strongly associated with immunologic factors that can be measured during clinical visits. We developed a T1D-specific ontology that captures the dynamic patterns of these biomarkers and used it together with a survival model, RankSvx, proposed in our prior work(2). We applied this approach to a T1D dataset harmonized from three birth cohort studies from the United States, Finland, and Sweden. Results show that the dynamic biomarker patterns captured in the proposed ontology are able to improve prediction performance (in concordance index) by 5.3%, 3.3%, 2.8%, and 1.0% over baseline for 3, 6, 9, and 12 month duration windows, respectively.

BackgroundPolygenic scores – which quantify inherited risk by integrating information from many common sites of DNA variation – may enable a tailored approach to clinical medicine. However, alongside considerable enthusiasm, we and others have highlighted a lack of systematic approaches for score disclosure. Here, we review the landscape of polygenic score reporting and describe a generalizable approach for development of polygenic score disclosure tools for coronary artery disease.MethodsFirst, we assembled a working group of clinicians, geneticists, data visualization specialists, and software developers. The group reviewed existing polygenic score reports and then designed a two-page mock polygenic score report for coronary artery disease. We then conducted a qualitative user-experience study with this report and an interview guide focused on comprehension, experience, and attitudes. Interviews were transcribed and thematically analyzed for themes identification.ResultsWe conducted interviews with ten adult individuals (50% females, 70% without prior genetic testing experience, age range 20 to 70 years) recruited via an online platform. We identified three themes from interviews: (1) visual elements, such as color and simple graphics, enable participants to interpret, relate to, and contextualize their polygenic score, (2) word-based descriptions of risk and polygenic scores presented as percentiles were most often recognized and understood, (3) participants had varying levels of interest in understanding complex genomic information and therefore would benefit from additional resources that can adapt to their individual needs in real time. In response to user feedback, colors used for communicating risk were modified to minimize unintended color associations and odds ratios were removed. Of note, all 10 participants expressed interest in receiving this report based on their personal genomic information.ConclusionsOur findings describe a generalizable approach to develop and test a polygenic score disclosure tool that is desired by the general public. These results are likely to inform ongoing efforts related to polygenic score disclosure within clinical practice.

The heart evolved hundreds of millions of years ago. During mammalian evolution, the cardiovascular system developed with complete separation between pulmonary and systemic circulations incorporated into a single pump with chambers dedicated to each circulation. A lower pressure right heart chamber supplies deoxygenated blood to the lungs, while a high pressure left heart chamber supplies oxygenated blood to the rest of the body. Due to the complexity of morphogenic cardiac looping and septation required to form these two chambers, congenital heart diseases often involve maldevelopment of the evolutionarily recent right heart chamber. Additionally, some diseases predominantly affect structures of the right heart, including arrhythmogenic right ventricular cardiomyopathy (ARVC) and pulmonary hypertension. To gain insight into right heart structure and function, we fine-tuned deep learning models to recognize the right atrium, the right ventricle, and the pulmonary artery, and then used those models to measure right heart structures in over 40,000 individuals from the UK Biobank with magnetic resonance imaging. We found associations between these measurements and clinical disease including pulmonary hypertension and dilated cardiomyopathy. We then conducted genome-wide association studies, identifying 104 distinct loci associated with at least one right heart measurement. Several of these loci were found near genes previously linked with congenital heart disease, such asNKX2-5, TBX3, WNT9B, andGATA4. We also observed interesting commonalities and differences in association patterns at genetic loci linked with both right and left ventricular measurements. Finally, we found that a polygenic predictor of right ventricular end systolic volume was associated with incident dilated cardiomyopathy (HR 1.28 per standard deviation; P = 2.4E-10), and remained a significant predictor of disease even after accounting for a left ventricular polygenic score. Harnessing deep learning to perform large-scale cardiac phenotyping, our results yield insights into the genetic and clinical determinants of right heart structure and function.

To support point-of-care decision making by presenting outcomes of past treatment choices for cohorts of similar patients based on observational data from electronic health records (EHRs), a machine-learning precision cohort treatment option (PCTO) workflow consisting of (1) data extraction, (2) similarity model training, (3) precision cohort identification, and (4) treatment options analysis was developed. The similarity model is used to dynamically create a cohort of similar patients, to inform clinical decisions about an individual patient. The workflow was implemented using EHR data from a large health care provider for three different highly prevalent chronic diseases: hypertension (HTN), type 2 diabetes mellitus (T2DM), and hyperlipidemia (HL). A retrospective analysis demonstrated that treatment options with better outcomes were available for a majority of cases (75%, 74%, 85% for HTN, T2DM, HL, respectively). The models for HTN and T2DM were deployed in a pilot study with primary care physicians using it during clinic visits. A novel data-analytic workflow was developed to create patient-similarity models that dynamically generate personalized treatment insights at the point-of-care. By leveraging both knowledge-driven treatment guidelines and data-driven EHR data, physicians can incorporate real-world evidence in their medical decision-making process when considering treatment options for individual patients.

Diabetes is a chronic diseases that affects nearly half a billion people around the globe, and is almost always associated with a number of complications, including kidney failure, blindness, stroke, and heart attack. An important step towards improved diabetes care is to accurately predict the risk of diabetes complications and to identify the corresponding risk factors associated with the onset of each complication. In this paper, we study the problem of risk prediction and outcome-specific risk factor identification from readily available patient medical record data. We adopt a private-shared multi-task learning (MTL) model, which jointly models multiple complications with each task corresponding to the risk modeling of one complication. The MTL formulation not only boosts prediction performance but also enables identification of outcome-specific risk factors. Specifically, we decompose the coefficient matrix, in which each column (vector) corresponds to the coefficient of one complication risk model, into a shared component and an outcome-specific private component. The shared component is assumed to be low-rank to capture the relationships among complications in terms of overall diabetes health condition. The private component is assumed to be non-overlapping and sparse so that they are discriminative among the different complication outcomes. Further, the shared component and the private component for the same complication are assumed to be orthogonal. Extensive experimental results on a type 2 diabetes cohort extracted from a large electronic medical claims database show that the proposed method outperforms baseline models by a significant margin. Also the identified outcome-specific risk factors provide meaningful clinical insights. The results demonstrate that simultaneously modeling multiple risks through MTL not only improves prediction performance but also enables identification of outcome-specific risk factors.

Background: Atrial fibrillation (AF) is associated with increased risks of stroke and heart failure. Electronic health record (EHR)–based AF risk prediction may facilitate efficient deployment of interventions to diagnose or prevent AF altogether. Methods: We externally validated an electronic health record AF (EHR-AF) score in IBM Explorys Life Sciences, a multi-institutional dataset containing statistically deidentified EHR data for over 21 million individuals (Explorys Dataset). We included individuals with complete AF risk data, ≥2 office visits within 2 years, and no prevalent AF. We compared EHR-AF to existing scores including CHARGE-AF (Cohorts for Heart and Aging Research in Genomic Epidemiology Atrial Fibrillation), C 2 HEST (coronary artery disease or chronic obstructive pulmonary disease, hypertension, elderly, systolic heart failure, thyroid disease), and CHA 2 DS 2 -VASc. We assessed association between AF risk scores and 5-year incident AF, stroke, and heart failure using Cox proportional hazards modeling, 5-year AF discrimination using C indices, and calibration of predicted AF risk to observed AF incidence. Results: Of 21 825 853 individuals in the Explorys Dataset, 4 508 180 comprised the analysis (age 62.5, 56.3% female). AF risk scores were strongly associated with 5-year incident AF (hazard ratio per SD increase 1.85 using CHA 2 DS 2 -VASc to 2.88 using EHR-AF), stroke (1.61 using C 2 HEST to 1.92 using CHARGE-AF), and heart failure (1.91 using CHA 2 DS 2 -VASc to 2.58 using EHR-AF). EHR-AF (C index, 0.808 [95% CI, 0.807–0.809]) demonstrated favorable AF discrimination compared to CHARGE-AF (0.806 [95% CI, 0.805–0.807]), C 2 HEST (0.683 [95% CI, 0.682–0.684]), and CHA 2 DS 2 -VASc (0.720 [95% CI, 0.719–0.722]). Of the scores, EHR-AF demonstrated the best calibration to incident AF (calibration slope, 1.002 [95% CI, 0.997–1.007]). In subgroup analyses, AF discrimination using EHR-AF was lower in individuals with stroke (C index, 0.696 [95% CI, 0.692–0.700]) and heart failure (0.621 [95% CI, 0.617–0.625]). Conclusions: EHR-AF demonstrates predictive accuracy for incident AF using readily ascertained EHR data. AF risk is associated with incident stroke and heart failure. Use of such risk scores may facilitate decision support and population health management efforts focused on minimizing AF-related morbidity.

Rapid advances in artificial intelligence (AI) and availability of biological, medical, and healthcare data have enabled the development of a wide variety of models. Significant success has been achieved in a wide range of fields, such as genomics, protein folding, disease diagnosis, imaging, and clinical tasks. Although widely used, the inherent opacity of deep AI models has brought criticism from the research field and little adoption in clinical practice. Concurrently, there has been a significant amount of research focused on making such methods more interpretable, reviewed here, but inherent critiques of such explainability in AI (XAI), its requirements, and concerns with fairness/robustness have hampered their real-world adoption. We here discuss how user-driven XAI can be made more useful for different healthcare stakeholders through the definition of three key personas-data scientists, clinical researchers, and clinicians-and present an overview of how different XAI approaches can address their needs. For illustration, we also walk through several research and clinical examples that take advantage of XAI open-source tools, including those that help enhance the explanation of the results through visualization. This perspective thus aims to provide a guidance tool for developing explainability solutions for healthcare by empowering both subject matter experts, providing them with a survey of available tools, and explainability developers, by providing examples of how such methods can influence in practice adoption of solutions.

In recent years, the rapid advances in Artificial Intelligence (AI) techniques along with an ever-increasing availability of healthcare data have made many novel analyses possible. Significant successes have been observed in a wide range of tasks such as next diagnosis prediction, AKI prediction, adverse event predictions including mortality and unexpected hospital re-admissions. However, there has been limited adoption and use in the clinical practice of these methods due to their black-box nature. A significant amount of research is currently focused on making such methods more interpretable or to make post-hoc explanations more accessible. However, most of this work is done at a very low level and as a result, may not have a direct impact at the point-of-care. This tutorial will provide an overview of the landscape of different approaches that have been developed for explainability in healthcare. Specifically, we will present the problem of explainability as it pertains to various personas involved in healthcare viz. data scientists, clinical researchers, and clinicians. We will chart out the requirements for such personas and present an overview of the different approaches that can address such needs. We will also walk-through several use-cases for such approaches. In this process, we will provide a brief introduction to explainability, charting its different dimensions as well as covering some relevant interpretability methods spanning such dimensions. We will touch upon some practical guides for explainability and provide a brief survey of open source tools such as the IBM AI Explainability 360 Open Source Toolkit.

Disease progression models are important computational tools in healthcare and are used for tasks such as improving disease understanding, informing drug discovery, and aiding in patient management. Although many algorithms for time series modeling exist, healthcare applications face particular challenges such as small datasets, medication effects, disease heterogeneity, and a desire for personalized predictions. In this work, we present a disease progression model that addresses these needs by proposing a probabilistic time-series model that captures individualized disease states, personalized medication effects, disease-state medication effects, or any combination thereof. The model builds on the framework of an input-output hidden Markov model where the parameters are learned using a structured variational approximation. To demonstrate the utility of the algorithm, we apply it to both synthetic and real-world datasets. In the synthetic case, we demonstrate the benefits afforded by the proposed model as compared to standard techniques. In the real-world cases, we use two Parkinson’s disease datasets to show improved predictive performance when ground truth is available and clinically relevant insights that are not revealed via classic Markov models when ground truth is not available.

Objective To present clinicians at the point-of-care with real-world data on the effectiveness of various treatment options in a precision cohort of patients closely matched to the index patient. Materials and Methods We developed disease-specific, machine-learning, patient-similarity models for hypertension (HTN), type II diabetes mellitus (T2DM), and hyperlipidemia (HL) using data on approximately 2.5 million patients in a large medical group practice. For each identified decision point, an encounter during which the patient’s condition was not controlled, we compared the actual outcome of the treatment decision administered to that of the best-achieved outcome for similar patients in similar clinical situations. Results For the majority of decision points (66.8%, 59.0%, and 83.5% for HTN, T2DM, and HL, respectively), there were alternative treatment options administered to patients in the precision cohort that resulted in a significantly increased proportion of patients under control than the treatment option chosen for the index patient. The expected percentage of patients whose condition would have been controlled if the best-practice treatment option had been chosen would have been better than the actual percentage by: 36% (65.1% vs 48.0%, HTN), 68% (37.7% vs 22.5%, T2DM), and 138% (75.3% vs 31.7%, HL). Conclusion Clinical guidelines are primarily based on the results of randomized controlled trials, which apply to a homogeneous subject population. Providing the effectiveness of various treatment options used in a precision cohort of patients similar to the index patient can provide complementary information to tailor guideline recommendations for individual patients and potentially improve outcomes.

Importance Pathogenic DNA variants associated with familial hypercholesterolemia, hereditary breast and ovarian cancer syndrome, and Lynch syndrome are widely recognized as clinically important and actionable when identified, leading some clinicians to recommend population-wide genomic screening. Objectives To assess the prevalence and clinical importance of pathogenic or likely pathogenic variants associated with each of 3 genomic conditions (familial hypercholesterolemia, hereditary breast and ovarian cancer syndrome, and Lynch syndrome) within the context of contemporary clinical care. Design, Setting, and Participants This cohort study used gene-sequencing data from 49 738 participants in the UK Biobank who were recruited from 22 sites across the UK between March 21, 2006, and October 1, 2010. Inpatient hospital data date back to 1977; cancer registry data, to 1957; and death registry data, to 2006. Statistical analysis was performed from July 22, 2019, to November 15, 2019. Exposures Pathogenic or likely pathogenic DNA variants classified by a clinical laboratory geneticist. Main Outcomes and Measures Composite end point specific to each genomic condition based on atherosclerotic cardiovascular disease events for familial hypercholesterolemia, breast or ovarian cancer for hereditary breast and ovarian cancer syndrome, and colorectal or uterine cancer for Lynch syndrome. Results Among 49 738 participants (mean [SD] age, 57 [8] years; 27 144 female [55%]), 441 (0.9%) harbored a pathogenic or likely pathogenic variant associated with any of 3 genomic conditions, including 131 (0.3%) for familial hypercholesterolemia, 235 (0.5%) for hereditary breast and ovarian cancer syndrome, and 76 (0.2%) for Lynch syndrome. Presence of these variants was associated with increased risk of disease: for familial hypercholesterolemia, 28 of 131 carriers (21.4%) vs 4663 of 49 607 noncarriers (9.4%) developed atherosclerotic cardiovascular disease; for hereditary breast and ovarian cancer syndrome, 32 of 116 female carriers (27.6%) vs 2080 of 27 028 female noncarriers (7.7%) developed associated cancers; and for Lynch syndrome, 17 of 76 carriers (22.4%) vs 929 of 49 662 noncarriers (1.9%) developed colorectal or uterine cancer. The predicted probability of disease at age 75 years despite contemporary clinical care was 45.3% for carriers of familial hypercholesterolemia, 41.1% for hereditary breast and ovarian cancer syndrome, and 38.3% for Lynch syndrome. Across the 3 conditions, 39.7% (175 of 441) of the carriers reported a family history of disease vs 23.2% (34 517 of 148 772) of noncarriers. Conclusions and Relevance The findings suggest that approximately 1% of the middle-aged adult population in the UK Biobank harbored a pathogenic variant associated with any of 3 genomic conditions. These variants were associated with an increased risk of disease despite contemporary clinical care and were not reliably detected by family history.

Background: Nonalcoholic fatty liver disease (NAFLD) is a highly prevalent yet underdiagnosed and under-discussed disease. Given that NAFLD has not been explored sufficiently compared with other diseases, opportunities abound for scientists to discover new biomarkers (such as laboratory observations, current comorbidities, and behavioral descriptors) that can be linked to the development of conditions and complications that may develop at a later stage of the patient’s life.Methods: We analyzed IBM Explorys, a repository that contains electronic medical records (EMRs) of more than 60 million individuals. We used a classification algorithm that members of our group have previously validated to identify patients at a high probability for NAFLD. The algorithm identified more than 80,000 patients with a high probability for NAFLD who had at least 5 years of follow-up. We followed an unbiased approach for prediction modeling and applied standard statistical methods (such as logistic regression and bootstrapping) as well as Clinical Classifications Software (CCS) definitions to identify associations between a variety of covariates and disease outcomes.Results: Our methodology identified several thousand strongly statistically significant associations between covariates and outcomes in NAFLD. Most of the associations are known, but others may be new and require further investigation in subsequent studies.Conclusions: A discovery mechanism composed of standard statistical methods and applied on a large collection of EMRs confirmed known associations and identified potentially new associations that can act as biomarkers that might merit further research.

Electronic Health Records (EHRs) have been widely used in healthcare studies recently, such as the analyses for patient diagnostic outcome and understanding of disease progression. EHR is a treasure for researchers who conduct the Real-World study to discovering Real-World Evidence (RWE). In this paper, we design an end-to-end learning system for disease states discovery based on a data-driven strategy. A large-scale proprietary EHR data mart containing about 55 million patients with over 20 billion data records is used for data extraction and analysis. Given a disease of interest, researchers could easily obtain the hidden disease states. Once our system were operational, biomedical researchers could get the results for downstream analyses such as disease prediction, drug design and outcome analyses.

Clustering, the process of grouping together similar items into distinct partitions, is a common type of unsupervised machine learning that can be useful for summarizing and aggregating complex multi-dimensional data. However, data can be clustered in many ways, and there exist a large body of algorithms designed to reveal different patterns. While having access to a wide variety of algorithms is helpful, in practice, it is quite difficult for data scientists to choose and parameterize algorithms to get the clustering results relevant for their dataset and analytical tasks. To alleviate this problem, we built Clustervision , a visual analytics tool that helps ensure data scientists find the right clustering among the large amount of techniques and parameters available. Our system clusters data using a variety of clustering techniques and parameters and then ranks clustering results utilizing five quality metrics. In addition, users can guide the system to produce more relevant results by providing task-relevant constraints on the data. Our visual user interface allows users to find high quality clustering results, explore the clusters using several coordinated visualization techniques, and select the cluster result that best suits their task. We demonstrate this novel approach using a case study with a team of researchers in the medical domain and showcase that our system empowers users to choose an effective representation of their complex data.