Detecting QT prolongation from a single-lead ECG with deep learning.

For a number of antiarrhythmics, drug loading requires a 3-day hospitalization with continuous monitoring for QT-prolongation. Automated QT monitoring with wearable ECG monitors would enable out-of-hospital care. We therefore develop a deep learning model that infers QT intervals from ECG Lead-I-the lead that is often available in ambulatory ECG monitors-and use this model to detect clinically meaningful QT-prolongation episodes during Dofetilide drug loading. QTNet-a deep neural network that infers QT intervals from Lead-I ECG-was trained using over 3 million ECGs from 653 thousand patients at the Massachusetts General Hospital and tested on an internal-test set consisting of 633 thousand ECGs from 135 thousand patients. QTNet is further evaluated on an external-validation set containing 3.1 million ECGs from 667 thousand patients at another healthcare institution. On both evaluations, the model achieves mean absolute errors of 12.63ms (internal-test) and 12.30ms (external-validation) for estimating absolute QT intervals. The associated Pearson correlation coefficients are 0.91 (internal-test) and 0.92 (external-validation). Finally, QTNet was used to detect Dofetilide-induced QT prolongation in a publicly available database (ECGRDVQ-dataset) containing ECGs from subjects enrolled in a clinical trial evaluating the effects of antiarrhythmic drugs. QTNet detects Dofetilide-induced QTc prolongation with 87% sensitivity and 77% specificity. The negative predictive value of the model is greater than 95% when the pre-test probability of drug-induced QTc prolongation is below 25%. These results show that drug-induced QT prolongation risk can be tracked from ECG Lead-I using deep learning.

Dynamics of collagen oxidation and cross linking in regenerating and irreversibly infarcted myocardium.

In mammalian hearts myocardial infarction produces a permanent collagen-rich scar. Conversely, in zebrafish a collagen-rich scar forms but is completely resorbed as the myocardium regenerates. The formation of cross-links in collagen hinders its degradation but cross-linking has not been well characterized in zebrafish hearts. Here, a library of fluorescent probes to quantify collagen oxidation, the first step in collagen cross-link (CCL) formation, was developed. Myocardial injury in mice or zebrafish resulted in similar dynamics of collagen oxidation in the myocardium in the first month after injury. However, during this time, mature CCLs such as pyridinoline and deoxypyridinoline developed in the murine infarcts but not in the zebrafish hearts. High levels of newly oxidized collagen were still seen in murine scars with mature CCLs. These data suggest that fibrogenesis remains dynamic, even in mature scars, and that the absence of mature CCLs in zebrafish hearts may facilitate their ability to regenerate.

Multi-scale assessment of roost selection by 'ope'ape'a, the Hawaiian hoary bat (Lasiurus semotus).

The Hawaiian hoary bat (Lasiurus semotus; Chiroptera: Vespertilionidae), commonly and locally known as 'ope'ape'a, is a solitary, insectivorous, and foliage-roosting species distributed across a wide range of habitats in lowland and montane environments. The species, as with many others in the Hawaiian archipelago, are facing a suite of challenges due to habitat loss and degradation, introduced predators and pests, and climate change. An understanding of the roost requirements of foliage-roosting tree bats is critical to their conservation as these habitats provide several important benefits to survival and reproduction. Because little is known about 'ope'ape'a roost ecology and considerable effort is needed to capture and track bats to roost locations, we examined resource selection at multiple spatial scales-perch location within a roost tree, roost tree, and forest stand. We used a discrete choice modeling approach to investigate day-roost selection and describe attributes of roost trees including those used as maternity roosts. 'Ope'ape'a were found roosting in 19 tree species and in an assortment of landcover types including native and non-native habitats. Our results are largely consistent with findings of other studies of foliage-roosting, insectivorous tree bats where bats selected roost locations that may offer protection and thermoregulatory benefits.

Tissue perfusion pressure enables continuous hemodynamic evaluation and risk prediction in the intensive care unit.

Treatment of circulatory shock in critically ill patients requires management of blood pressure using invasive monitoring, but uncertainty remains as to optimal individual blood pressure targets. Critical closing pressure, which refers to the arterial pressure when blood flow stops, can provide a fundamental measure of vascular tone in response to disease and therapy, but it has not previously been possible to measure this parameter routinely in clinical care. Here we describe a method to continuously measure critical closing pressure in the systemic circulation using readily available blood pressure monitors and then show that tissue perfusion pressure (TPP), defined as the difference between mean arterial pressure and critical closing pressure, provides unique information compared to other hemodynamic parameters. Using analyses of 5,988 admissions to a modern cardiac intensive care unit, and externally validated with 864 admissions to another institution, we show that TPP can predict the risk of mortality, length of hospital stay and peak blood lactate levels. These results indicate that TPP may provide an additional target for blood pressure optimization in patients with circulatory shock.

Dynamics of Collagen Oxidation and Cross Linking in Regenerating and Irreversibly Infarcted Myocardium.

In mammalian hearts myocardial infarction produces a permanent collagen-rich scar. Conversely, in zebrafish a collagen-rich scar forms but is completely resorbed as the myocardium regenerates. The formation of cross-links in collagen hinders its degradation but cross-linking has not been well characterized in zebrafish hearts. Here, a library of fluorescent probes to quantify collagen oxidation, the first step in collagen cross-link (CCL) formation, was developed. Myocardial injury in mice or zebrafish resulted in similar dynamics of collagen oxidation in the myocardium in the first month after injury. However, during this time, mature CCLs such as pyridinoline and deoxypyridinoline developed in the murine infarcts but not in the zebrafish hearts. High levels of newly oxidized collagen were still seen in murine scars with mature CCLs. These data suggest that fibrogenesis remains dynamic, even in mature scars, and that the absence of mature CCLs in zebrafish hearts may facilitate their ability to regenerate.

Myeloperoxidase PET Imaging Tracks Intracellular and Extracellular Treatment Changes in Experimental Myocardial Infarction.

Myeloperoxidase (MPO) is a highly oxidative, pro-inflammatory enzyme involved in post-myocardial infarction (MI) injury and is a potential therapeutic target. While multiple MPO inhibitors have been developed, the lack of an imaging reporter to select appropriate patients and assess therapeutic efficacy has hampered clinical development. Thus, a translational imaging method to detect MPO activity non-invasively would help to better understand the role MPO plays in MI and facilitate novel therapy development and clinical validation. Interestingly, many MPO inhibitors affect both intracellular and extracellular MPO, but previous MPO imaging methods can only report extracellular MPO activity. In this study, we found that an MPO-specific PET imaging agent (18F-MAPP) can cross cell membranes to report intracellular MPO activity. We showed that 18F-MAPP can track the treatment effect of an MPO inhibitor (PF-2999) at different doses in experimental MI. The imaging results were corroborated by ex vivo autoradiography and gamma counting data. Furthermore, extracellular and intracellular MPO activity assays revealed that 18F-MAPP imaging can report the changes induced by PF-2999 on both intracellular and extracellular MPO activities. These findings support 18F-MAPP as a translational candidate to noninvasively report MPO activity and accelerate drug development against MPO and other related inflammatory targets.

Prediction of operative mortality for patients undergoing cardiac surgical procedures without established risk scores.

OBJECTIVE: Current cardiac surgery risk models do not address a substantial fraction of procedures. We sought to create models to predict the risk of operative mortality for an expanded set of cases. METHODS: Four supervised machine learning models were trained using preoperative variables present in the Society of Thoracic Surgeons (STS) data set of the Massachusetts General Hospital to predict and classify operative mortality in procedures without STS risk scores. A total of 424 (5.5%) mortality events occurred out of 7745 cases. Models included logistic regression with elastic net regularization (LogReg), support vector machine, random forest (RF), and extreme gradient boosted trees (XGBoost). Model discrimination was assessed via area under the receiver operating characteristic curve (AUC), and calibration was assessed via calibration slope and expected-to-observed event ratio. External validation was performed using STS data sets from Brigham and Women's Hospital (BWH) and the Johns Hopkins Hospital (JHH). RESULTS: Models performed comparably with the highest mean AUC of 0.83 (RF) and expected-to-observed event ratio of 1.00. On external validation, the AUC was 0.81 in BWH (RF) and 0.79 in JHH (LogReg/RF). Models trained and applied on the same institution's data achieved AUCs of 0.81 (BWH: LogReg/RF/XGBoost) and 0.82 (JHH: LogReg/RF/XGBoost). CONCLUSIONS: Machine learning models trained on preoperative patient data can predict operative mortality at a high level of accuracy for cardiac surgical procedures without established risk scores. Such procedures comprise 23% of all cardiac surgical procedures nationwide. This work also highlights the value of using local institutional data to train new prediction models that account for institution-specific practices.

Deep Learning to Predict Mortality After Cardiothoracic Surgery Using Preoperative Chest Radiographs.

BACKGROUND: The Society of Thoracic Surgeons Predicted Risk of Mortality (STS-PROM) estimates mortality risk only for certain common procedures (eg, coronary artery bypass or valve surgery) and is cumbersome, requiring greater than 60 inputs. We hypothesized that deep learning can estimate postoperative mortality risk based on a preoperative chest radiograph for cardiac surgeries in which STS-PROM scores were available (STS index procedures) or unavailable (non-STS index procedures). METHODS: We developed a deep learning model (CXR-CTSurgery) to predict postoperative mortality based on preoperative chest radiographs in 9283 patients at Massachusetts General Hospital (MGH) having cardiac surgery before April 8, 2014. CXR-CTSurgery was tested on 3615 different MGH patients and externally tested on 2840 patients from Brigham and Women's Hospital (BWH) having surgery after April 8, 2014. Discrimination for mortality was compared with the STS-PROM using the C-statistic. Calibration was assessed using the observed-to-expected ratio (O/E ratio). RESULTS: For STS index procedures, CXR-CTSurgery had a C-statistic similar to STS-PROM at MGH (CXR-CTSurgery: 0.83 vs STS-PROM: 0.88; P = .20) and BWH (0.74 vs 0.80; P = .14) testing cohorts. The CXR-CTSurgery C-statistic for non-STS index procedures was similar to STS index procedures in the MGH (0.87 vs 0.83) and BWH (0.73 vs 0.74) testing cohorts. For STS index procedures, CXR-CTSurgery had better calibration than the STS-PROM in the MGH (O/E ratio: 0.74 vs 0.52) and BWH (O/E ratio: 0.91 vs 0.73) testing cohorts. CONCLUSIONS: CXR-CTSurgery predicts postoperative mortality based on a preoperative CXR with similar discrimination and better calibration than the STS-PROM. This may be useful when the STS-PROM cannot be calculated or for non-STS index procedures.

QTNet: Deep Learning for Estimating QT Intervals Using a Single Lead ECG.

QT prolongation often leads to fatal arrhythmia and sudden cardiac death. Antiarrhythmic drugs can increase the risk of QT prolongation and therefore require strict post-administration monitoring and dosage control. Measurement of the QT interval from the 12-lead electrocardiogram (ECG) by a trained expert, in a clinical setting, is the accepted method for tracking QT prolongation. Recent advances in wearable ECG technology, however, raise the possibility of automated out-of-hospital QT tracking. Applications of Deep Learning (DL) - a subfield within Machine Learning - in ECG analysis holds the promise of automation for a variety of classification and regression tasks. In this work, we propose a residual neural network, QTNet, for the regression of QT intervals from a single lead (Lead-I) ECG. QTNet is trained in a supervised manner on a large ECG dataset from a U.S. hospital. We demonstrate the robustness and generalizability of QTNet on four test-sets; one from the same hospital, one from another U.S. hospital, and two public datasets. Over all four datasets, the mean absolute error (MAE) in the estimated QT interval ranges between 9ms and 15.8ms. Pearson correlation coefficients vary between 0.899 and 0.914. By contrast, QT interval estimation on these datasets with a standard method for automated ECG analysis (NeuroKit2) yields MAEs between 22.29ms and 90.79ms, and Pearson correlation coefficients 0.345 and 0.620. These results demonstrate the utility of QTNet across distinct datasets and patient populations, thereby highlighting the potential utility of DL models for ubiquitous QT tracking.Clinical Relevance- QTNet can be applied to inpatient or ambulatory Lead-I ECG signals to track QT intervals. The method facilitates ambulatory monitoring of patients at risk of QT prolongation.

Timing of diuretic administration effects on urine volume in hospitalized patients.

Importance: Some medications have effects that depend on the time of day they are given. Current knowledge of the time-of-day effects of specific medications in hospitalized patients with cardiovascular disease is very limited. In hospitalized patients, increased medication efficiency might reduce dose (and associated side effects) and/or the length of time in the Intensive Care Unit (ICU) or hospital-potentially improving patient outcomes and patient and family quality of life and reducing financial costs. We studied whether the time of day or night patients in Cardiac or Intensive Care Units receive a diuretic affects urine volume. Methods: In this observational study, data were collected from 7,685 patients (63% male, 18 to 98 years old) admitted to one hospital's Acute Care Cardiac units, Cardiac ICUs, Cardiac Surgery ICUs, and/or Non-cardiac ICUs who received intravenous furosemide (a diuretic), had measurements of urine volume, were hospitalized for ≥3 days between January 2016 to July 2021 and were older than 18 years. The outcomes of interest were urine volume normalized by the most recent (not older than 24 h) weight or body mass index (BMI), (i) in the hour after the time of diuretic administration, and (ii) when no diuretics were administered for the previous 3 h. Results: We identified diuretic medication administration time 23:00-04:59 as a predictor of higher urine volume response. For patients without recent diuretic medication, higher urine volume was predicted 11:00-16:59 and 17:00-22:59. Other factors that affected urine volume response to the diuretic were sex, age, medication dose, creatinine concentration, diagnoses, and hospital unit. Discussion: Time-of-day of medication administration may be a factor associated with increased medication efficiency. Randomized controlled trials should be conducted to quantify the relative effect of modifiable factors, such as time of medication administration, that may affect short- and longer-term outcomes.

Human acute inflammatory recovery is defined by co-regulatory dynamics of white blood cell and platelet populations.

Inflammation is the physiologic reaction to cellular and tissue damage caused by trauma, ischemia, infection, and other pathologic conditions. Elevation of white blood cell count (WBC) and altered levels of other acute phase reactants are cardinal signs of inflammation, but the dynamics of these changes and their resolution are not well established. Here we studied inflammatory recovery from trauma, ischemia, and infection by tracking longitudinal dynamics of clinical laboratory measurements in hospitalized patients. We identified a universal recovery trajectory defined by exponential WBC decay and delayed linear growth of platelet count (PLT). Co-regulation of WBC-PLT dynamics is a fundamental mechanism of acute inflammatory recovery and provides a generic approach for identifying high-risk patients: 32x relative risk (RR) of adverse outcomes for cardiac surgery, 9x RR of death from COVID-19, 9x RR of death from sepsis, and 5x RR of death from myocardial infarction.

Neutrophils incite and macrophages avert electrical storm after myocardial infarction.

Sudden cardiac death, arising from abnormal electrical conduction, occurs frequently in patients with coronary heart disease. Myocardial ischemia simultaneously induces arrhythmia and massive myocardial leukocyte changes. In this study, we optimized a mouse model in which hypokalemia combined with myocardial infarction triggered spontaneous ventricular tachycardia in ambulatory mice, and we showed that major leukocyte subsets have opposing effects on cardiac conduction. Neutrophils increased ventricular tachycardia via lipocalin-2 in mice, whereas neutrophilia associated with ventricular tachycardia in patients. In contrast, macrophages protected against arrhythmia. Depleting recruited macrophages in Ccr2 -/- mice or all macrophage subsets with Csf1 receptor inhibition increased both ventricular tachycardia and fibrillation. Higher arrhythmia burden and mortality in Cd36 -/- and Mertk -/- mice, viewed together with reduced mitochondrial integrity and accelerated cardiomyocyte death in the absence of macrophages, indicated that receptor-mediated phagocytosis protects against lethal electrical storm. Thus, modulation of leukocyte function provides a potential therapeutic pathway for reducing the risk of sudden cardiac death.

Remote monitoring in the use of extracorporeal membrane oxygenation and acute mechanical circulatory support.

PURPOSE OF REVIEW: To provide an overview of the role of remote monitoring tools in management of critically-ill patients requiring acute mechanical circulatory support (MCS). RECENT FINDINGS: Tele-critical care systems have received new interest during the COVID-19 pandemic, which has stretched the capacity of health systems everywhere. At the same time, utilization of MCS and extracorporeal membrane oxygenation (ECMO) technologies has increased during the pandemic. The opportunity for remote monitoring and clinical decision support for ECMO and acute MCS devices has been recognized by industry partners, with several major platforms implementing technology infrastructure for it in available products. Healthcare systems face challenges interfacing multiple devices from multiple manufacturers with each other and with their designated electronic health records. Furthermore, the availability of data must be combined with algorithms for alerting on clinical events and with implementation systems to act upon these alerts. Studies are not yet published validating remote monitoring platforms for ECMO and MCS in clinical care. SUMMARY: Remote monitoring for MCS devices represents a major opportunity for further investigation to improve the utilization of these devices and better serve patients.

Optical coherence tomography in coronary atherosclerosis assessment and intervention.

Since optical coherence tomography (OCT) was first performed in humans two decades ago, this imaging modality has been widely adopted in research on coronary atherosclerosis and adopted clinically for the optimization of percutaneous coronary intervention. In the past 10 years, substantial advances have been made in the understanding of in vivo vascular biology using OCT. Identification by OCT of culprit plaque pathology could potentially lead to a major shift in the management of patients with acute coronary syndromes. Detection by OCT of healed coronary plaque has been important in our understanding of the mechanisms involved in plaque destabilization and healing with the rapid progression of atherosclerosis. Accurate detection by OCT of sequelae from percutaneous coronary interventions that might be missed by angiography could improve clinical outcomes. In addition, OCT has become an essential diagnostic modality for myocardial infarction with non-obstructive coronary arteries. Insight into neoatherosclerosis from OCT could improve our understanding of the mechanisms of very late stent thrombosis. The appropriate use of OCT depends on accurate interpretation and understanding of the clinical significance of OCT findings. In this Review, we summarize the state of the art in cardiac OCT and facilitate the uniform use of this modality in coronary atherosclerosis. Contributions have been made by clinicians and investigators worldwide with extensive experience in OCT, with the aim that this document will serve as a standard reference for future research and clinical application.

Intravital Microscopy in Atherosclerosis Research.

Atherosclerosis is a lipid-driven inflammatory disorder that narrows the arterial lumen and can induce life-threatening complications from coronary artery disease, cerebrovascular disease, and peripheral artery disease. On a mechanistic level, the development of novel cellular-resolution intravital microscopy imaging approaches has recently enabled in vivo studies of underlying biological processes governing disease onset and progress. In particular, multiphoton microscopy has emerged as a promising intravital imaging tool utilizing two-photon-excited fluorescence and second-harmonic generation that provides subcellular resolution and increased imaging depths beyond confocal and epifluorescence microscopy. In this chapter, we describe the state-of-the-art multiphoton microscopy applied to the study of murine atherosclerosis.

Patient contrastive learning: A performant, expressive, and practical approach to electrocardiogram modeling.

Supervised machine learning applications in health care are often limited due to a scarcity of labeled training data. To mitigate the effect of small sample size, we introduce a pre-training approach, Patient Contrastive Learning of Representations (PCLR), which creates latent representations of electrocardiograms (ECGs) from a large number of unlabeled examples using contrastive learning. The resulting representations are expressive, performant, and practical across a wide spectrum of clinical tasks. We develop PCLR using a large health care system with over 3.2 million 12-lead ECGs and demonstrate that training linear models on PCLR representations achieves a 51% performance increase, on average, over six training set sizes and four tasks (sex classification, age regression, and the detection of left ventricular hypertrophy and atrial fibrillation), relative to training neural network models from scratch. We also compared PCLR to three other ECG pre-training approaches (supervised pre-training, unsupervised pre-training with an autoencoder, and pre-training using a contrastive multi ECG-segment approach), and show significant performance benefits in three out of four tasks. We found an average performance benefit of 47% over the other models and an average of a 9% performance benefit compared to best model for each task. We release PCLR to enable others to extract ECG representations at https://github.com/broadinstitute/ml4h/tree/master/model_zoo/PCLR.

Late onset atrial fibrillation in patients undergoing surgical aortic valve replacement.

OBJECTIVE: Aortic valve disease is a risk factor for atrial fibrillation (AF), and AF is associated with increased late mortality and morbidity after cardiac surgery. The evolution of alternative approaches to AF prophylaxis, including less invasive technologies and medical therapies, has altered the balance between risk and potential benefit for prophylactic intervention at the time of surgical aortic valve replacement (SAVR). Such interventions impose incremental risk, however, making an understanding of predictors of new onset AF that persists beyond the perioperative episode relevant. METHODS: We conducted a retrospective single-institution cohort analysis of patients undergoing SAVR with no history of preoperative AF (n = 1014). These patients were cross-referenced against an institutional electrocardiogram (ECG) database to identify those with ECGs 3-12 months after surgery. Logistic regression was used to identify predictors of late AF. RESULTS: Among the 401 patients (40%), who had ECGs in our institution 3-12 months after surgery, 16 (4%) had late AF. Patients with late AF were older than patients without late AF (73 vs. 65, p = .025), and underwent procedures that were more urgent/emergent (38% vs. 15%, p = .015), with higher predicted risk of mortality (2.2% vs. 1.3%, p = .012). Predictors associated with the development of late AF were advanced age, higher preoperative creatinine level and urgent/emergent surgery. CONCLUSIONS: The incidence of late AF 3-12 months after SAVR, is low. Prophylactic AF interventions at the time of SAVR may not be warranted.

Optical Coherence Tomography of Plaque Vulnerability and Rupture: JACC Focus Seminar Part 1/3.

Plaque rupture is the most common cause of acute coronary syndromes and sudden cardiac death. Characteristics and pathobiology of vulnerable plaques prone to plaque rupture have been studied extensively over 2 decades in humans using optical coherence tomography (OCT), an intravascular imaging technique with micron scale resolution. OCT studies have identified key features of plaque vulnerability and described the in vivo characteristics and spatial distribution of thin cap fibroatheromas as major precursors to plaque rupture. In addition, OCT data supports the evolving understanding of coronary heart disease as a panvascular process associated with inflammation. In the setting of high atherosclerotic burden, plaque ruptures often occur at multiple sites in the coronary arteries, and plaque progression and healing are dynamic processes modulated by systemic risk factors. This review details major investigations with intravascular OCT into the biology and clinical implications of plaque vulnerability and plaque rupture.

White Blood Cell and Platelet Dynamics Define Human Inflammatory Recovery.

Inflammation is the physiologic reaction to cellular and tissue damage caused by pathologic processes including trauma, infection, and ischemia 1 . Effective inflammatory responses integrate molecular and cellular functions to prevent further tissue damage, initiate repair, and restore homeostasis, while futile or dysfunctional responses allow escalating injury, delay recovery, and may hasten death 2 . Elevation of white blood cell count (WBC) and altered levels of other acute phase reactants are cardinal signs of inflammation, but the dynamics of these changes and their resolution are not established 3,4 . Patient responses appear to vary dramatically with no clearly defined signs of good prognosis, leaving physicians reliant on qualitative interpretations of laboratory trends 4,5 . We retrospectively, observationally studied the human acute inflammatory response to trauma, ischemia, and infection by tracking the longitudinal dynamics of cellular and serum markers in hospitalized patients. Unexpectedly, we identified a conserved pattern of recovery defined by co-regulation of WBC and platelet (PLT) populations. Across all inflammatory conditions studied, recovering patients followed a consistent WBC-PLT trajectory shape that is well-approximated by exponential WBC decay and delayed linear PLT growth. This recovery trajectory shape may represent a fundamental archetype of human physiologic response at the cellular population scale, and provides a generic approach for identifying high-risk patients: 32x relative risk of adverse outcomes for cardiac surgery patients, 9x relative risk of death for COVID-19, and 5x relative risk of death for myocardial infarction.