Search

Postexercise heart rate recovery (HRR) is an important indicator of cardiac autonomic function and abnormal HRR is associated with adverse outcomes. We hypothesized that deep learning on resting electrocardiogram (ECG) tracings may identify individuals with impaired HRR.We trained a deep learning model (convolutional neural network) to infer HRR based on resting ECG waveforms (HRRAmong 56,793 individuals (mean age 57 years, 51% women), the HRRDeep learning-derived estimates of HRR using resting ECG independently associated with future clinical outcomes, including new-onset DM and all-cause mortality. Inferring postexercise heart rate response from a resting ECG may have potential clinical implications and impact on preventive strategies warrants future study.

Background: Artificial intelligence (AI)–enabled analysis of 12-lead ECGs may facilitate efficient estimation of incident atrial fibrillation (AF) risk. However, it remains unclear whether AI provides meaningful and generalizable improvement in predictive accuracy beyond clinical risk factors for AF. Methods: We trained a convolutional neural network (ECG-AI) to infer 5-year incident AF risk using 12-lead ECGs in patients receiving longitudinal primary care at Massachusetts General Hospital (MGH). We then fit 3 Cox proportional hazards models, composed of ECG-AI 5-year AF probability, CHARGE-AF clinical risk score (Cohorts for Heart and Aging in Genomic Epidemiology–Atrial Fibrillation), and terms for both ECG-AI and CHARGE-AF (CH-AI), respectively. We assessed model performance by calculating discrimination (area under the receiver operating characteristic curve) and calibration in an internal test set and 2 external test sets (Brigham and Women’s Hospital [BWH] and UK Biobank). Models were recalibrated to estimate 2-year AF risk in the UK Biobank given limited available follow-up. We used saliency mapping to identify ECG features most influential on ECG-AI risk predictions and assessed correlation between ECG-AI and CHARGE-AF linear predictors. Results: The training set comprised 45 770 individuals (age 55±17 years, 53% women, 2171 AF events) and the test sets comprised 83 162 individuals (age 59±13 years, 56% women, 2424 AF events). Area under the receiver operating characteristic curve was comparable using CHARGE-AF (MGH, 0.802 [95% CI, 0.767–0.836]; BWH, 0.752 [95% CI, 0.741–0.763]; UK Biobank, 0.732 [95% CI, 0.704–0.759]) and ECG-AI (MGH, 0.823 [95% CI, 0.790–0.856]; BWH, 0.747 [95% CI, 0.736–0.759]; UK Biobank, 0.705 [95% CI, 0.673–0.737]). Area under the receiver operating characteristic curve was highest using CH-AI (MGH, 0.838 [95% CI, 0.807 to 0.869]; BWH, 0.777 [95% CI, 0.766 to 0.788]; UK Biobank, 0.746 [95% CI, 0.716 to 0.776]). Calibration error was low using ECG-AI (MGH, 0.0212; BWH, 0.0129; UK Biobank, 0.0035) and CH-AI (MGH, 0.012; BWH, 0.0108; UK Biobank, 0.0001). In saliency analyses, the ECG P-wave had the greatest influence on AI model predictions. ECG-AI and CHARGE-AF linear predictors were correlated (Pearson r : MGH, 0.61; BWH, 0.66; UK Biobank, 0.41). Conclusions: AI-based analysis of 12-lead ECGs has similar predictive usefulness to a clinical risk factor model for incident AF and the approaches are complementary. ECG-AI may enable efficient quantification of future AF risk.

With novel hallucinogens poised to enter psychiatry, a unified framework for quantifying which changes in consciousness are optimal for treatment is needed. Using transformers (i.e. BERT) and 11,816 publicly-available drug testimonials, we first predicted 28-dimensions of sentiment across each narrative, and then validated these predictions with adjudication by a clinical psychiatrist. Secondly, we fine-tuned BERT to predict biochemical and demographic information from natural language testimonials of drug experiences. Thirdly, canonical correlation analysis (CCA) linked 52 drugs' receptor affinities with word usage, revealing 11 statistically-significant latent receptor-experience factors, each mapped to a 3D cortical atlas. Together, these machine learning methods elucidate a neurobiologically-informed, temporally-sensitive portrait of drug-induced subjective experiences. The models’ results converged, revealing a pervasive distinction between the universal psychedelic heights of feeling in contrast to the grim, mundane, and personal experiences with addiction and mental illness. MDMA was linked to "Love", DMT and 5-MeO-DMT to "Mystical Experiences" and “Entities and Beings”, and other tryptamines to "Surprise", "Curiosity" and "Realization". Applying these models to real-time biofeedback, practitioners could delicately calibrate the course of therapeutic sessions.

Multiobjective optimization is often a difficult task owing to the need to balance competing objectives. A typical approach to handling this is to estimate a Pareto frontier in objective space by i...

Background: Cardiac magnetic resonance (CMR) is the gold standard for left ventricular hypertrophy (LVH) diagnosis. CMR-derived LV mass can be estimated using proprietary algorithms (eg, InlineVF), but their accuracy and availability may be limited. Objective: To develop an open-source deep learning model to estimate CMR-derived LV mass. Methods: Within participants of the UK Biobank prospective cohort undergoing CMR, we trained 2 convolutional neural networks to estimate LV mass. The first (ML4Hreg) performed regression informed by manually labeled LV mass (available in 5065 individuals), while the second (ML4Hseg) performed LV segmentation informed by InlineVF (version D13A) contours. We compared ML4Hreg, ML4Hseg, and InlineVF against manually labeled LV mass within an independent holdout set using Pearson correlation and mean absolute error (MAE). We assessed associations between CMR-derived LVH and prevalent cardiovascular disease using logistic regression adjusted for age and sex. Results: We generated CMR-derived LV mass estimates within 38,574 individuals. Among 891 individuals in the holdout set, ML4Hseg reproduced manually labeled LV mass more accurately (r = 0.864, 95% confidence interval [CI] 0.847–0.880; MAE 10.41 g, 95% CI 9.82–10.99) than ML4Hreg (r = 0.843, 95% CI 0.823–0.861; MAE 10.51, 95% CI 9.86–11.15, P = .01) and InlineVF (r = 0.795, 95% CI 0.770–0.818; MAE 14.30, 95% CI 13.46–11.01, P < .01). LVH defined using ML4Hseg demonstrated the strongest associations with hypertension (odds ratio 2.76, 95% CI 2.51–3.04), atrial fibrillation (1.75, 95% CI 1.37–2.20), and heart failure (4.67, 95% CI 3.28–6.49). Conclusions: ML4Hseg is an open-source deep learning model providing automated quantification of CMR-derived LV mass. Deep learning models characterizing cardiac structure may facilitate broad cardiovascular discovery.

Objective: To assess whether a videogame-like digital treatment is superior to a control in improving processing speed in adults with multiple sclerosis (MS). Methods: Adults with MS and baseline Symbol Digit Modalities Test (SDMT) z-scores between −2 and 0 were enrolled in a double-blind randomized controlled clinical trial. After completing a baseline in-clinic evaluation (Visit 1), they were randomized to complete an in-home, tablet-based videogame-like digital treatment (AKL-T03) or control word game (AKL-T09) for up to 25 minutes/day, 5 days/week, for 6 weeks. A repeat in-clinic evaluation occurred at 6 weeks (Visit 2), and again 8 weeks later to determine persistence of effects (Visit 3). The pre-specified primary outcome was change in SDMT score between Visits 1 and 2. Results: SDMT increased at Visit 2 for participants randomized to both AKL-T03 ( p Conclusion: This in-home digital intervention resulted in substantial and durable improvements in processing speed. A larger randomized controlled clinical trial is planned. Trial Registration: This trial is registered on ClinicalTrials.gov under “NCT03569618,” https://clinicaltrials.gov/ct2/show/NCT03569618 .

Introduction: Deep learning-derived representations of 12-lead electrocardiograms (ECGs) may allow for atrial fibrillation (AF) risk prediction. However, it remains unclear whether ECG-based artificial intelligence improves prediction beyond established clinical risk factors for AF and whether predictions are generalizable. Methods: Within a dataset comprising over 500,000 individuals receiving regular primary care at a multi-institutional network, we trained a convolutional neural network to predict incident AF using 12-lead ECGs (“ECG-AI”). ECG-AI was trained in individuals with ≥1 ECG performed at Massachusetts General Hospital (MGH) within 3 years prior to start of follow-up. We then fit a Cox proportional hazards model with incident AF as the outcome and a) logit-transformed ECG-AI AF probability, and b) the Cohorts for Aging and Genomic Epidemiology AF (CHARGE-AF) score, as covariates (“CH-AI”). We compared the discrimination and calibration of CHARGE-AF versus CH-AI in three independent samples: MGH (n=4,166), Brigham and Women’s Hospital (BWH, n=37,963) and the UK Biobank (n=41,034). Based on available follow-up, AF was evaluated at 5 years in MGH and BWH, and 2 years in the UK Biobank. Results: ECG-AI was trained in 36,081 individuals with an ECG performed at MGH (mean age 55±17, 53% female). CH-AI had substantially better discrimination (area under the receiver operating characteristic curve [AUROC]: MGH 0.838, BWH 0.777, UK Biobank 0.746; average precision [AP] 0.30, 0.21, 0.06) versus CHARGE-AF (AUROC: 0.802, 0.752, 0.732; AP 0.21, 0.17, 0.02, Figure ). CH-AI was well-calibrated in MGH (calibration error 0.012) and BWH (0.019), but overestimated AF risk in the UK Biobank (0.068). Calibration in the UK Biobank was excellent after recalibration to the sample-level 2-year AF hazard (error 7.1x10 -5 ). Conclusions: A model combining clinical AF risk factors with deep learning-derived ECG-based AF risk is favorable for predicting 5-year risk of AF.

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.

Background: Classical methods for detecting left ventricular (LV) hypertrophy (LVH) using 12-lead ECGs are insensitive. Deep learning models using ECG to infer cardiac magnetic resonance (CMR)-derived LV mass may improve LVH detection. Methods: Within 32 239 individuals of the UK Biobank prospective cohort who underwent CMR and 12-lead ECG, we trained a convolutional neural network to predict CMR-derived LV mass using 12-lead ECGs (left ventricular mass-artificial intelligence [LVM-AI]). In independent test sets (UK Biobank [n=4903] and Mass General Brigham [MGB, n=1371]), we assessed correlation between LVM-AI predicted and CMR-derived LV mass and compared LVH discrimination using LVM-AI versus traditional ECG-based rules (ie, Sokolow-Lyon, Cornell, lead aVL rule, or any ECG rule). In the UK Biobank and an ambulatory MGB cohort (MGB outcomes, n=28 612), we assessed associations between LVM-AI predicted LVH and incident cardiovascular outcomes using age- and sex-adjusted Cox regression. Results: LVM-AI predicted LV mass correlated with CMR-derived LV mass in both test sets, although correlation was greater in the UK Biobank (r=0.79) versus MGB (r=0.60, P Conclusions: Deep learning-inferred LV mass estimates from 12-lead ECGs correlate with CMR-derived LV mass, associate with incident cardiovascular disease, and may improve LVH discrimination compared to traditional ECG rules.

BackgroundElectronic health records (EHRs) promise to enable broad-ranging discovery with power exceeding that of conventional research cohort studies. However, research using EHR datasets may be subject to selection bias, which can be compounded by missing data, limiting the generalizability of derived insights.MethodsMass General Brigham (MGB) is a large New England-based healthcare network comprising seven tertiary care and community hospitals with associated outpatient practices. Within an MGB-based EHR warehouse of >3.5 million individuals with at least one ambulatory care visit, we approximated a community-based cohort study by selectively sampling individuals longitudinally attending primary care practices between 2001-2018 (n=520,868), which we named the Community Care Cohort Project (C3PO). We also utilized pre-trained deep natural language processing (NLP) models to recover vital signs (i.e., height, weight, and blood pressure) from unstructured notes in the EHR. We assessed the validity of C3PO by deploying established risk models including the Pooled Cohort Equations (PCE) and the Cohorts for Aging and Genomic Epidemiology Atrial Fibrillation (CHARGE-AF) score, and compared model performance in C3PO to that observed within typical EHR Convenience Samples which included all individuals from the same parent EHR with sufficient data to calculate each score but without a requirement for longitudinal primary care. All analyses were facilitated by the JEDI Extractive Data Infrastructure pipeline which we designed to efficiently aggregate EHR data within a unified framework conducive to regular updates.ResultsC3PO includes 520,868 individuals (mean age 48 years, 61% women, median follow-up 7.2 years, median primary care visits per individual 13). Estimated using reports, C3PO contains over 2.9 million electrocardiograms, 450,000 echocardiograms, 12,000 cardiac magnetic resonance images, and 75 million narrative notes. Using tabular data alone, 286,009 individuals (54.9%) had all vital signs available at baseline, which increased to 358,411 (68.8%) after NLP recovery (31% reduction in missingness). Among individuals with both NLP and tabular data available, NLP-extracted and tabular vital signs obtained on the same day were highly correlated (e.g., Pearson r range 0.95-0.99, pConclusionsIntentional sampling of individuals receiving regular ambulatory care and use of NLP to recover missing data have the potential to reduce bias in EHR research and maximize generalizability of insights.

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.

BackgroundCardiac magnetic resonance (CMR) is the gold standard for left ventricular hypertrophy (LVH) diagnosis. CMR-derived LV mass can be estimated using proprietary algorithms (e.g., inlineVF), but their accuracy and availability may be limited.ObjectiveTo develop an open-source deep learning model to estimate CMR-derived LV mass.MethodsWithin participants of the UK Biobank prospective cohort undergoing CMR, we trained two convolutional neural networks to estimate LV mass. The first (ML4Hreg) performed regression informed by manually labeled LV mass (available in 5,065 individuals), while the second (ML4Hseg) performed LV segmentation informed by inlineVF contours. We compared ML4Hreg, ML4Hseg, and inlineVF against manually labeled LV mass within an independent holdout set using Pearson correlation and mean absolute error (MAE). We assessed associations between CMR-derived LVH and prevalent cardiovascular disease using logistic regression adjusted for age and sex.ResultsWe generated CMR-derived LV mass estimates within 38,574 individuals. Among 891 individuals in the holdout set, ML4Hseg reproduced manually labeled LV mass more accurately (r=0.864, 95% CI 0.847-0.880; MAE 10.41g, 95% CI 9.82-10.99) than ML4Hreg (r=0.843, 95% CI 0.823-0.861; MAE 10.51, 95% CI 9.86-11.15, p=0.01) and inlineVF (r=0.795, 95% CI 0.770-0.818; MAE 14.30, 95% CI 13.46-11.01, pseg demonstrated the strongest associations with hypertension (odds ratio 2.76, 95% CI 2.51-3.04), atrial fibrillation (1.75, 95% CI 1.37-2.20), and heart failure (4.53, 95% CI 3.16-6.33).ConclusionsML4Hseg is an open-source deep learning model providing automated quantification of CMR-derived LV mass. Deep learning models characterizing cardiac structure may facilitate broad cardiovascular discovery.

BACKGROUND Cognitive impairment (CI) is one of the most prevalent symptoms of multiple sclerosis (MS). However, it is difficult to include cognitive assessment as part of MS standard care since the comprehensive neuropsychological examinations are usually time-consuming and extensive. OBJECTIVE To improve access to CI assessment, we evaluated the feasibility and potential assessment sensitivity of a tablet-based cognitive battery in patients with MS. METHODS In total, 53 participants with MS (24 [45%] with CI and 29 [55%] without CI) and 24 non-MS participants were assessed with a tablet-based cognitive battery (Adaptive Cognitive Evaluation [ACE]) and standard cognitive measures, including the Symbol Digit Modalities Test (SDMT) and the Paced Auditory Serial Addition Test (PASAT). Associations between performance in ACE and the SDMT/PASAT were explored, with group comparisons to evaluate whether ACE modules can capture group-level differences. RESULTS Correlations between performance in ACE and the SDMT (R=–0.57, PP=.01), were observed. Compared to non-MS and non-CI MS groups, the CI MS group showed a slower reaction time (CI MS vs non-MS: PP=.004) and a higher attention cost (CI MS vs non-MS: P=.02; CI MS vs non-CI MS: P CONCLUSIONS These results provide preliminary evidence that ACE, a tablet-based cognitive assessment battery, provides modules that could potentially serve as a digital cognitive assessment for people with MS. CLINICALTRIAL ClinicalTrials.gov NCT03569618; https://clinicaltrials.gov/ct2/show/NCT03569618

The aorta is the largest blood vessel in the body, and enlargement or aneurysm of the aorta can predispose to dissection, an important cause of sudden death. While rare syndromes have been identified that predispose to aortic aneurysm, the common genetic basis for the size of the aorta remains largely unknown. By leveraging a deep learning architecture that was originally developed to recognize natural images, we trained a model to evaluate the dimensions of the ascending and descending thoracic aorta in cardiac magnetic resonance imaging. After manual annotation of just 116 samples, we applied this model to 3,840,140 images from the UK Biobank. We then conducted a genome-wide association study in 33,420 individuals, revealing 68 loci associated with ascending and 35 with descending thoracic aortic diameter, of which 10 loci overlapped. Integration of common variation with transcriptome-wide analyses, rare-variant burden tests, and single nucleus RNA sequencing prioritized SVIL, a gene highly expressed in vascular smooth muscle, that was significantly associated with the diameter of the ascending and descending aorta. A polygenic score for ascending aortic diameter was associated with a diagnosis of thoracic aortic aneurysm in the remaining 391,251 UK Biobank participants who did not undergo imaging (HR = 1.44 per standard deviation; P = 3.7·10−12). Defining the genetic basis of the diameter of the aorta may enable the identification of asymptomatic individuals at risk for aneurysm or dissection and facilitate the prioritization of potential therapeutic targets for the prevention or treatment of aortic aneurysm. Finally, our results illustrate the potential for rapidly defining novel quantitative traits derived from a deep learning model, an approach that can be more broadly applied to biomedical imaging data.

The intersection of medicine and machine learning (ML) has the potential to transform healthcare. We describe how physiology, a foundational discipline of medical training and practice with a rich quantitative history, could serve as a starting point for the development of a common language between clinicians and ML experts, thereby accelerating real-world impact.

Excess accumulation of liver fat – termed hepatic steatosis when fat accounts for > 5.5% of liver content – is a leading risk factor for end-stage liver disease and is strongly associated with important cardiometabolic disorders. Using a truth dataset of 4,511 UK Biobank participants with liver fat previously quantified via abdominal MRI imaging, we developed a machine learning algorithm to quantify liver fat with correlation coefficients of 0.97 and 0.99 in hold-out testing datasets and applied this algorithm to raw imaging data from an additional 32,192 participants. Among all 36,703 individuals with abdominal MRI imaging, median liver fat was 2.2%, with 6,250 (17%) meeting criteria for hepatic steatosis. Although individuals afflicted with hepatic steatosis were more likely to have been diagnosed with conditions such as obesity or diabetes, a prediction model based on clinical data alone without imaging could not reliably estimate liver fat content. To identify genetic drivers of variation in liver fat, we first conducted a common variant association study of 9.8 million variants, confirming three known associations for variants in theTM6SF2, APOE, andPNPLA3genes and identifying five new variants associated with increased hepatic fat in or near theMARC1, ADH1B, TRIB1, GPAMandMAST3genes. A polygenic score that integrated information from each of these eight variants was strongly associated with future clinical diagnosis of liver diseases. Next, we performed a rare variant association study in a subset of 11,021 participants with gene sequencing data available, identifying an association between inactivating variants in theAPOBgene and substantially lower LDL cholesterol, but more than 10-fold increased risk of steatosis. Taken together, these results provide proof of principle for the use of machine learning algorithms on raw imaging data to enable epidemiologic studies and genetic discovery.

BackgroundCross-sectional studies of various forms of dilated cardiomyopathy have noted a truncating mutation in the gene encoding titin (‘TTNtv’) in 7-30% of patients, but the clinical importance of identifying a TTNtv in an asymptomatic adult is largely unknown. In contrast to cross-sectional studies, prospective cohort studies allow for unbiased estimates of the disease risks associated with a genotype exposure.ObjectivesTo determine the prevalence of cardiac imaging abnormalities and risk of incident disease among middle-aged TTNtv carriers without known congestive heart failure.MethodsWe analyze exome sequencing data of 45,747 participants of the UK Biobank without known congestive heart failure to identify TTNtv carriers. Among 10,552 with cardiac magnetic resonance imaging (MRI), we determine the relationship between TTNtv carrier status and left ventricular ejection fraction. In this prospective cohort, we quantify the absolute and relative risks of incident disease in TTNtv carriers versus noncarriers.ResultsAmong 45,747 middle-aged participants without known congestive heart failure, 196 (0.43%) harbored a TTNtv. The average ejection fraction was 61% in TTNtv carriers versus 65% in noncarriers (P = 1.8 × 10−8), with a 9.3-fold increase (95% CI 3.9 – 22.2) in odds of subnormal ejection fraction (P = 5.7 × 10−5). Over a median follow-up of 6.9 years, a composite endpoint of incident dilated cardiomyopathy, congestive heart failure, or all-cause mortality was observed in 6.6% of TTNtv carriers versus 2.9% of non-carriers (adjusted hazard ratio 2.5; 95% CI 1.4 – 4.3; p = 1.1 × 10−3).ConclusionsApproximately 1 in 230 middle-aged adults without known congestive heart failure harbored a TTNtv. These carriers had a substantially increased relative risk—but modest absolute risk—of having a subnormal ejection fraction or manifesting clinical disease during prospective follow-up.Condensed AbstractCross-sectional studies of dilated cardiomyopathy have noted a truncating mutation in the gene encoding titin (‘TTNtv’) in up to 30% of patients—but the clinical importance of TTNtv in asymptomatic adults is largely unknown. Here, we observe a TTNtv in 0.43% of 45,747 middle-aged adults. Average ejection fraction was 61% in TTNtv carriers versus 65% in non-carriers (p

Available computational models for many engineering design applications are both expensive and and of a black-box nature. This renders traditional optimization techniques difficult to apply, including gradient-based optimization and expensive heuristic approaches. For such situations, Bayesian global optimization approaches, that both explore and exploit a true function while building a metamodel of it, are applied. These methods often rely on a set of alternative candidate designs over which a querying policy is designed to search. For even modestly high-dimensional problems, such an alternative set approach can be computationally intractable, due to the reliance on excessive exploration of the design space. To overcome this, we have developed a framework for the optimization of expensive black-box models, which is based on active subspace exploitation and a two-step knowledge gradient policy. We demonstrate our approach on three benchmark problems and a practical aerostructural wing design problem, where our method performs well against traditional direct application of Bayesian global optimization techniques.

A "seek-test-treat" intervention (ARISTOTLE) was implemented in response to an outbreak of human immunodeficiency virus (HIV) infection among persons who inject drugs (PWID) in Athens. We assess trends in HIV incidence, prevalence, risk behaviors and access to prevention/treatment.Methods included behavioral data collection, provision of injection equipment, HIV testing, linkage to opioid substitution treatment (OST) programs and HIV care during 5 rounds of respondent-driven sampling (2012-2013). HIV incidence was estimated from observed seroconversions.Estimated coverage of the target population was 88% (71%-100%; 7113 questionnaires/blood samples from 3320 PWID). The prevalence of HIV infection was 16.5%. The incidence per 100 person-years decreased from 7.8 (95% confidence interval, 4.6-13.1) (2012) to 1.7 (0.55-5.31) (2013; P for trend = .001). Risk factors for seroconversion were frequency of injection, homelessness, and history of imprisonment. Injection at least once daily declined from 45.2% to 18.8% (P.001) and from 36.8% to 26.0% (P = .007) for sharing syringes, and the proportion of undiagnosed HIV infection declined from 84.3% to 15.0% (P.001). Current OST increased from 12.2% to 27.7% (P.001), and 48.4% of unlinked seropositive participants were linked to HIV care through 2013. Repeat participants reported higher rates of adequate syringe coverage, linkage to HIV care and OST.Multiple evidence-based interventions delivered through rapid recruitment in a large proportion of the population of PWID are likely to have helped mitigate this HIV outbreak.

Current design strategies for multi-physics systems seek to exploit synergistic interactions among disciplines in the system. However, when dealing with a multidisciplinary system with multiple physics represented, the use of high-fidelity computational models is often prohibitive. In this situation, recourse is often made to lower fidelity models that have potentially significant uncertainty associated with them. We present here a novel approach to quantifying the discipline level uncertainty in coupled multi-physics models, so that these individual models may later be used in isolation or coupled within other systems. Our approach is based off of a Gibbs sampling strategy and the identification of a necessary detailed balance condition that constrains the possible characteristics of individual model discrepancy distributions. We demonstrate our methodology on both a linear and nonlinear example problem.Copyright © 2016 by ASME

Indomethacin was utilized in 89 patients in preterm labor; the mean gestational age at the start of indomethacin therapy was 25.4 weeks (± 4.1), mean duration of indomethacin treatment was 43.9 days (± 31.3), mean gestational age at delivery was 33.6 weeks (± 3.2). Long-term ( > 48 h) indomethacin tocolysis evoked serious maternal side effects in 30.3% of patients studied. These side effects included exaggerated hypertensive responses in patients treated concomitantly with beta-blockers, gastric erosion and bleeding, fever, interstitial nephritis and postpartum hemorrhage.

Corynebacterium pseudotuberculosis infection in an Israeli dairy cattle herd is described. The disease was characterized by ulcerative granulomatous lesions, which occurred in an epidemic form. Thirty-two cows and two heifers were affected, the ratio of the number affected to number at risk being 17.5 : 1 and 9.5 : 1, respectively. The culling rate was 50% of the affected animals. Most of the affected animals were cows (91.2%), with one first-calving cow (2.9%) and two heifers (5.9%) also affected. The infection occurred during the summer to autumn months (August-December), and lasted 118 days. The incubation period is about 2 months. The disease appeared in two clinical forms - cutaneous and mastitic - or as a mixed form. C. pseudotuberculosis organisms that were isolated from the ulcerative granulomatous lesions and from milk samples failed to reduce nitrate. A decrease in milk production (4%) and an increase in the bulk-milk somatic cell count from a herd mean of 240 x 10(3) mL(-1) to 460 x 10(3) mL(-1) were noted during the morbidity period. The organism was isolated from milk samples of eight animals (25%). Clinical, epizootiological and microbiological aspects of the infection are described.

Thirty-four cases of soft tissue tumors (STT) other than rhabdomyosarcoma in infants under 1 year of age were seen in our institution between 1955 and 1985. All were diagnosed initially as malignant tumors except for three cases of fibromatosis, and, thus, they received therapy appropriate at that time. During a recent pathologic review, four were seen to have had hemangioma, six hemangiopericytoma, one hamartoma, seven fibromatosis, eight fibrosarcoma, and eight unclassified sarcomas. Of these 34 cases, the initial histological diagnosis was confirmed in only 17. Sixteen patients were believed to have received inappropriately aggressive therapy; indeed, four of these cases diagnosed prior to 1975 had been benign. Major long-term side effects were seen in 7/16 treated patients (six radiotherapy, one surgery), who, in retrospect, should have received less aggressive treatment. The reasons for these initial errors in diagnosis included doubtful histologic features associated in some cases with alarming clinical presentation. We believe that difficulties still remain in distinguishing benign from malignant STT in this age group by histological examination. Therefore, we urge extreme caution in using aggressive therapy before the diagnosis is certain.

We report the largest series of induction chemotherapy for inflammatory breast carcinoma (IBC). Results of two chemotherapy protocols with radiation therapy (RT) (170 patients) are compared with results with radiation alone (60 patients) in the treatment of this disease. From 1973 to 1975, 60 patients (control, group C) received RT (45 Gy and 20 to 30 Gy boost) and hormonal manipulation. From 1976 to 1980, 91 patients (group A) were treated with induction chemotherapy: Adriamycin (Adria Laboratories, Columbus, Ohio), vincristine, and methotrexate (AVM) and RT on a cyclical schedule; and maintenance chemotherapy: vincristine, cyclophosphamide, and 5-fluorouracil (5-FU) (VCF). From 1980 to 1982, 79 patients (group B) received induction chemotherapy, Adriamycin, vincristine, cyclophosphamide, methotrexate, and 5-FU (AVCMF) and RT on a cyclical schedule and VCF maintenance. Hormonal manipulation was performed in all groups. Disease-free survival at 4 years was 15% for group C, 32% for group A, and 54% for group B (P less than .005 group C v group A, less than .00001 group C v group B, and less than .01 group A v group B). Total survival at 4 years was 42% for group C, 53% for group A, and 74% for group B (P = .17 group C v group A, less than .00001 group C v group B, and less than .001 group A v group B). Clinical assessment of tumor aggressiveness, nodal status, type of chemotherapy administered, and early response to chemotherapy (by third course) were all prognostic factors. There is an important, highly statistically significant benefit in terms of both disease-free survival and total survival observed in patients treated with the addition of chemotherapy compared with radiation alone in IBC.

Evidence that supports the use of systemic, presurgical induction chemotherapy to render soft tissue sarcomas resectable or to minimize the extent of surgical excision is presented. Induction chemotherapy was administered in 34 cases of nonmetastatic soft tissue sarcomas. All patients had large tumors for which only mutilating surgery, if any, was possible. In 21 patients, a combination of Adriamycin (doxorubicin), cyclophosphamide, cisplatin, vindesine, and DTIC (DCPAV) produced two complete remissions (CR) and 6 partial remissions (PR). A combination of cyclophosphamide, vincristine, Adriamycin, and DTIC (CYVADIC) produced three PR in eight patients, and a combination of Adriamycin and ifosfamide (AI) produced two PR in five patients. After two to seven cycles of chemotherapy, 24 patients underwent surgery. In 19, gross tumor excision was performed; 12 proved to be microscopically fully resected. Disease in two patients entered CR with chemotherapy alone, but surgery was performed in both patients as well. Irradiation was administered in ten patients to produce or insure CR (eight cases of residual disease postoperatively), and in two patients with unresectable disease. Four patients with disease in CR after surgery also received radiation due to the initial massive tumor size. The Kaplan-Meier survival curves at 2 years showed 18% total survival in the patients in whom CR of disease was not achieved, and 80% survival in patients with disease in CR. Of the 22 patients with disease in CR (by all means), disease-free survival was 1 to 44 months (mean, 13.7). Disease currently remains in CR in ten patients with a mean follow-up of 13.6 months (3 to 34 months from end of therapy). Ten patients had a local recurrence following a CR after 3 to 44 months (mean, 15.3 months).

Forty-three infants under 1 year of age (median 4 months) with rhabdomyosarcoma (RMS), representing 11% of all childhood RMS seen over the same period, were treated and followed between 1955 and 1984 at our institute. Compared to the 340 older children (1-16 years of age), there were no significant differences in male/female ratio, primary site, or clinical stage, but there was a higher frequency of alveolar and poorly differentiated histologic subtypes in infants aged under 1 year (P = 0.05 and P less than 0.001, respectively) and increased frequency of botryoid RMS in older children (P less than 0.02). Although children older than 1 year of age had a significantly higher complete remission rate (83% vs. 51%, P less than 0.001), their 5 year overall survival (43%) was no different from that of infants under 1 year of age (35%). In the older group there was no survival difference between alveolar (38%) and nonalveolar (44%) histologic subtypes, but in infants younger than 1 year, alveolar subtype had a significantly poorer prognosis (13% vs. 47%, P = 0.01). Management of patients at this age group is difficult and major advances are needed to increase survival rates with a minimum of morbidity.

Current therapeutic modalities for childhood acute lymphoblastic leukemia (ALL) are associated with a high cure rate, and recurrences more than 4 years after therapy cessation are very unusual. We report three cases of exceptionally late recurrences of childhood ALL after cessation of chemotherapy (CT) given for respective periods of 8, 7, and 24 months. CT was followed by maintenance immunotherapy (IMT) with Bacillus Calmette-Guérin (BCG) and allogeneic leukemic lymphoblasts pretreated with formaldehyde or irradiated in vitro. Leukemic recurrences were observed 19, 11, and 9 years after cessation of CT and appeared morphologically similar to the original blasts. A second complete remission was easily achieved in all three patients, but two went on to repeated relapse (one has died following the fourth recurrence). We speculate that some residual leukemic cells, remaining after the inadequate, short-term CT, were responsible for these unusual evolutions, and we question a possible delaying role of IMT in prolonging remission. Other possible etiologies are discussed.

We report a series of 34 adult patients with lymphoblastic lymphoma treated at our Institute between 1977 and 1986. At presentation, mediastinal involvement was seen in 76%, extranodal involvement (not including bone marrow) in 20%, stage IV in 53% and circulating blasts in 15%. In the 15 cases where immunological studies were performed, 12 proved to be of T type and the other three lacked both T and B markers. Chemotherapy with acute lymphocytic leukemia (ALL) protocols was given to 58% and lymphoma protocols to the other patients, with CNS prophylaxis given to a total of 28 patients. Although complete remission (CR) was observed overall in 74% of patients, 5 year survival was only 22% (42% in stages I-II and 8% stages III-IV) with improved results seen with recent aggressive anti-leukemic protocols. Five relapsed patients entered second CR and two were still in CR at this time. CNS relapse occurred in a total of eight patients (three without prophylaxis), and was isolated in two of six patients who achieved CR (despite CNS prophylaxis). In an analysis of prognostic factors, only the attainment of a CR was statistically significant (P less than 0.001). Thus we are unable to confirm other studies which demonstrated prognostic variations in this disease and believe that all such patients should receive aggressive chemotherapy with substantial CNS prophylaxis similar to those currently used for ALL.

Two cases of ruptured duodenal ulcer occurring in patients previously treated with metrazol are described. The theory that there may be a causal relationship between the preceding metrazol therapy and the later gastrointestinal pathology is advanced.

Previous reports in the literature have crystallized quite well the results effected by frontal lobotomy in the treatment of the major psychoses. Our present investigation, which deals particularly with dementia praecox, gives further support to previous conclusions. Apparently one can anticipate that some 50% to 60% of appropriately selected patients will exhibit significant improvement following frontal lobotomy, with approximately equal division between levels of moderate improvement and of marked improvement. There appear to be relatively few relapses with the passage of time—at least within a 2-year period—and, statistically, these relapses are compensated by other cases that exhibit their optimal improvement after a year or more of postoperative interlude. In general, results are somewhat better in affective disorders and other nonschizophrenic conditions than in dementia praecox. Among the latter, best results are observed in the mixed and paranoid subgroups. Symptoms related to tension and depressiv...