Identification and risk stratification of coronary disease by artificial intelligence-enabled ECG.

Background: Atherosclerotic cardiovascular disease (ASCVD) is the leading cause of death worldwide, driven primarily by coronary artery disease (CAD). ASCVD risk estimators such as the pooled cohort equations (PCE) facilitate risk stratification and primary prevention of ASCVD but their accuracy is still suboptimal. Methods: Using deep electronic health record data from 7,116,209 patients seen at 70+ hospitals and clinics across 5 states in the USA, we developed an artificial intelligence-based electrocardiogram analysis tool (ECG-AI) to detect CAD and assessed the additive value of ECG-AI-based ASCVD risk stratification to the PCE. We created independent ECG-AI models using separate neural networks including subjects without known history of ASCVD, to identify coronary artery calcium (CAC) score ≥300 Agatston units by computed tomography, obstructive CAD by angiography or procedural intervention, and regional left ventricular akinesis in ≥1 segment by echocardiogram, as a reflection of possible prior myocardial infarction (MI). These were used to assess the utility of ECG-AI-based ASCVD risk stratification in a retrospective observational study consisting of patients with PCE scores and no prior ASCVD. The study period covered all available digitized EHR data, with the first available ECG in 1987 and the last in February 2023. Findings: ECG-AI for identifying CAC ≥300, obstructive CAD, and regional akinesis achieved area under the receiver operating characteristic (AUROC) values of 0.88, 0.85, and 0.94, respectively. An ensembled ECG-AI identified 3, 5, and 10-year risk for acute coronary events and mortality independently and additively to PCE. Hazard ratios for acute coronary events over 3-years in patients without ASCVD that tested positive on 1, 2, or 3 versus 0 disease-specific ECG-AI models at cohort entry were 2.41 (2.14-2.71), 4.23 (3.74-4.78), and 11.75 (10.2-13.52), respectively. Similar stratification was observed in cohorts stratified by PCE or age. Interpretation: ECG-AI has potential to address unmet need for accessible risk stratification in patients in whom PCE under, over, or insufficiently estimates ASCVD risk, and in whom risk assessment over time periods shorter than 10 years is desired. Funding: Anumana.

Development of Electronic Health Record-Based Machine Learning Models to Predict Barrett's Esophagus and Esophageal Adenocarcinoma Risk.

INTRODUCTION: Screening for Barrett's esophagus (BE) is suggested in those with risk factors, but remains underutilized. BE/esophageal adenocarcinoma (EAC) risk prediction tools integrating multiple risk factors have been described. However, accuracy remains modest (area under the receiver-operating curve [AUROC] ≤0.7), and clinical implementation has been challenging. We aimed to develop machine learning (ML) BE/EAC risk prediction models from an electronic health record (EHR) database. METHODS: The Clinical Data Analytics Platform, a deidentified EHR database of 6 million Mayo Clinic patients, was used to predict BE and EAC risk. BE and EAC cases and controls were identified using International Classification of Diseases codes and augmented curation (natural language processing) techniques applied to clinical, endoscopy, laboratory, and pathology notes. Cases were propensity score matched to 5 independent randomly selected control groups. An ensemble transformer-based ML model architecture was used to develop predictive models. RESULTS: We identified 8,476 BE cases, 1,539 EAC cases, and 252,276 controls. The BE ML transformer model had an overall sensitivity, specificity, and AUROC of 76%, 76%, and 0.84, respectively. The EAC ML transformer model had an overall sensitivity, specificity, and AUROC of 84%, 70%, and 0.84, respectively. Predictors of BE and EAC included conventional risk factors and additional novel factors, such as coronary artery disease, serum triglycerides, and electrolytes. DISCUSSION: ML models developed on an EHR database can predict incident BE and EAC risk with improved accuracy compared with conventional risk factor-based risk scores. Such a model may enable effective implementation of a minimally invasive screening technology.

Clinical Administration Characteristics of Subcutaneous and Intravenous Administration of Daratumumab in Patients With Multiple Myeloma at Mayo Clinic Infusion Centers.

PURPOSE: Median duration of daratumumab (DARA) administration for treatment of multiple myeloma is 3-7 hours for the intravenous formulation (DARA IV) and 3-5 minutes for the subcutaneous formulation (DARA SC). Here, we describe clinical administration characteristics of DARA using a novel method for data extraction from electronic health records. METHODS: Time-based measurements were extracted using a scheduling/pharmacy software program that tracked patient movement through appointments for patients initiating DARA in Mayo Clinic infusion centers from April 5, 2017, to October 14, 2021. Cohorts included patients who received DARA IV or DARA SC, or converted from DARA IV to DARA SC. The DARA SC cohort was further analyzed before (DARA SC initial) and after (DARA SC shortened) a reduction in the postadministration observation time mandated by the treatment plan. Events associated with administration-related reactions (ARRs) were also identified. RESULTS: Median total clinic times were 2.7-3.0 hours shorter for DARA SC versus DARA IV. Median clinic times were highest at dose 1 and decreased with subsequent doses. Median total chair times were 2.7-2.8 hours shorter for DARA SC versus DARA IV. Incidences of ARR-related events with DARA SC were low across doses. CONCLUSION: Reduced clinic times were observed with DARA SC, indicating that use of DARA SC as a treatment option results in time savings that may free clinic resources. Furthermore, novel methods of electronic health record data extraction can provide insights that may help inform clinic resource optimization.

Cerebral Venous Sinus Thrombosis is not Significantly Linked to COVID-19 Vaccines or Non-COVID Vaccines in a Large Multi-State Health System.

OBJECTIVE: To assess the association of COVID-19 vaccines and non-COVID-19 vaccines with cerebral venous sinus thrombosis (CVST). MATERIALS AND METHOD: We retrospectively analyzed a cohort of 771,805 vaccination events across 266,094 patients in the Mayo Clinic Health System between 01/01/2017 and 03/15/2021. The primary outcome was a positive diagnosis of CVST, identified either by the presence of a corresponding ICD code or by an NLP algorithm which detected positive diagnosis of CVST within free-text clinical notes. For each vaccine we calculated the relative risk by dividing the incidence of CVST in the 30 days following vaccination to that in the 30 days preceding vaccination. RESULTS: We identified vaccination events for all FDA-approved COVID-19 vaccines including Pfizer-BioNTech (n = 94,818 doses), Moderna (n = 36,350 doses) and Johnson & Johnson - J&J (n = 1,745 doses). We also identified vaccinations events for 10 common FDA-approved non-COVID-19 vaccines (n = 771,805 doses). There was no statistically significant difference in the incidence rate of CVST in 30-days before and after vaccination for any vaccine in this population. We further found the baseline CVST incidence in the study population between 2017 and 2021 to be 45 to 98 per million patient years. CONCLUSIONS: This real-world evidence-based study finds that CVST is rare and is not significantly associated with COVID-19 vaccination in our patient cohort. Limitations include the rarity of CVST in our dataset, a relatively small number of J&J COVID-19 vaccination events, and the use of a population drawn from recipients of a SARS-CoV-2 PCR test in a single health system.

Real-time analysis of a mass vaccination effort confirms the safety of FDA-authorized mRNA COVID-19 vaccines.

BACKGROUND: As the coronavirus disease 2019 (COVID-19) vaccination campaign unfolds, it is important to continuously assess the real-world safety of Food and Drug Administration (FDA)-authorized vaccines. Curation of large-scale electronic health records (EHRs) enables near-real-time safety evaluations that were not previously possible. METHODS: In this retrospective study, we deployed deep neural networks over a large EHR system to automatically curate the adverse effects mentioned by physicians in over 1.2 million clinical notes between December 1, 2020 and April 20, 2021. We compared notes from 68,266 individuals who received at least one dose of BNT162b2 (n = 51,795) or mRNA-1273 (n = 16,471) to notes from 68,266 unvaccinated individuals who were matched by demographic, geographic, and clinical features. FINDINGS: Individuals vaccinated with BNT162b2 or mRNA-1273 had a higher rate of return to the clinic, but not the emergency department, after both doses compared to unvaccinated controls. The most frequently documented adverse effects within 7 days of each vaccine dose included myalgia, headache, and fatigue, but the rates of EHR documentation for each side effect were remarkably low compared to those derived from active solicitation during clinical trials. Severe events, including anaphylaxis, facial paralysis, and cerebral venous sinus thrombosis, were rare and occurred at similar frequencies in vaccinated and unvaccinated individuals. CONCLUSIONS: This analysis of vaccine-related adverse effects from over 1.2 million EHR notes of more than 130,000 individuals reaffirms the safety and tolerability of the FDA-authorized mRNA COVID-19 vaccines in practice. FUNDING: This study was funded by nference.

Plasma IL-6 levels following corticosteroid therapy as an indicator of ICU length of stay in critically ill COVID-19 patients.

Intensive care unit (ICU) admissions and mortality in severe COVID-19 patients are driven by "cytokine storms" and acute respiratory distress syndrome (ARDS). Interim clinical trial results suggest that the corticosteroid dexamethasone displays better 28-day survival in severe COVID-19 patients requiring ventilation or oxygen. In this study, 10 out of 16 patients (62.5%) that had an average plasma IL-6 value over 10 pg/mL post administration of corticosteroids also had worse outcomes (i.e., ICU stay >15 days or death), compared to 8 out of 41 patients (19.5%) who did not receive corticosteroids (p-value = 0.0024). Given this potential association between post-corticosteroid IL-6 levels and COVID-19 severity, we hypothesized that the glucocorticoid receptor (GR or NR3C1) may be coupled to IL-6 expression in specific cell types that govern cytokine release syndrome (CRS). Examining single-cell RNA-seq data from BALF of severe COVID-19 patients and nearly 2 million cells from a pan-tissue scan shows that alveolar macrophages, smooth muscle cells, and endothelial cells co-express NR3C1 and IL-6, motivating future studies on the links between the regulation of NR3C1 function and IL-6 levels.

The bed nucleus of the stria terminalis and functionally linked neurocircuitry modulate emotion processing and HPA axis dysfunction in posttraumatic stress disorder.

BACKGROUND: The bed nucleus of the stria terminalis (BNST) plays an important role in rodent posttraumatic stress disorder (PTSD), but evidence to support its relevance to human PTSD is limited. We sought to understand the role of the BNST in human PTSD via fMRI, behavioral, and physiological measurements. METHODS: 29 patients with PTSD (childhood sexual abuse) and 23 healthy controls (HC) underwent BOLD imaging with an emotional word paradigm. Symptom severity was assessed using the Clinician-Administered PTSD Scale and HPA-axis dysfunction was assessed by measuring the diurnal cortisol amplitude index (DCAI). A data-driven multivariate analysis was used to determine BNST task-based functional co-occurrence (tbFC) across individuals. RESULTS: In the trauma-versus-neutral word contrast, patients showed increased activation compared to HC in the BNST, medial prefrontal cortex (mPFC), posterior cingulate gyrus (PCG), caudate heads, and midbrain, and decreased activation in dorsolateral prefrontal cortex (DLPFC). Symptom severity positively correlated with activity in the BNST, caudate head, amygdala, hippocampus, dorsal anterior cingulate gyrus (dACG), and PCG, and negatively with activity in the medial orbiotofrontal cortex (mOFC) and DLPFC. Patients and HC showed marked differences in the relationship between the DCAI and BOLD activity in the BNST, septal nuclei, dACG, and PCG. Patients showed stronger tbFC between the BNST and closely linked limbic and subcortical regions, and a loss of negative tbFC between the BNST and DLPFC. CONCLUSIONS: Based upon novel data, we present a new model of dysexecutive emotion processing and HPA-axis dysfunction in human PTSD that incorporates the role of the BNST and functionally linked neurocircuitry.

Augmented curation of clinical notes from a massive EHR system reveals symptoms of impending COVID-19 diagnosis.

Understanding temporal dynamics of COVID-19 symptoms could provide fine-grained resolution to guide clinical decision-making. Here, we use deep neural networks over an institution-wide platform for the augmented curation of clinical notes from 77,167 patients subjected to COVID-19 PCR testing. By contrasting Electronic Health Record (EHR)-derived symptoms of COVID-19-positive (COVIDpos; n = 2,317) versus COVID-19-negative (COVIDneg; n = 74,850) patients for the week preceding the PCR testing date, we identify anosmia/dysgeusia (27.1-fold), fever/chills (2.6-fold), respiratory difficulty (2.2-fold), cough (2.2-fold), myalgia/arthralgia (2-fold), and diarrhea (1.4-fold) as significantly amplified in COVIDpos over COVIDneg patients. The combination of cough and fever/chills has 4.2-fold amplification in COVIDpos patients during the week prior to PCR testing, in addition to anosmia/dysgeusia, constitutes the earliest EHR-derived signature of COVID-19. This study introduces an Augmented Intelligence platform for the real-time synthesis of institutional biomedical knowledge. The platform holds tremendous potential for scaling up curation throughput, thus enabling EHR-powered early disease diagnosis.

Multimodal SHG-2PF Imaging of Microdomain Ca2+-Contraction Coupling in Live Cardiac Myocytes.

RATIONALE: Cardiac myocyte contraction is caused by Ca(2+) binding to troponin C, which triggers the cross-bridge power stroke and myofilament sliding in sarcomeres. Synchronized Ca(2+) release causes whole cell contraction and is readily observable with current microscopy techniques. However, it is unknown whether localized Ca(2+) release, such as Ca(2+) sparks and waves, can cause local sarcomere contraction. Contemporary imaging methods fall short of measuring microdomain Ca(2+)-contraction coupling in live cardiac myocytes. OBJECTIVE: To develop a method for imaging sarcomere level Ca(2+)-contraction coupling in healthy and disease model cardiac myocytes. METHODS AND RESULTS: Freshly isolated cardiac myocytes were loaded with the Ca(2+)-indicator fluo-4. A confocal microscope equipped with a femtosecond-pulsed near-infrared laser was used to simultaneously excite second harmonic generation from A-bands of myofibrils and 2-photon fluorescence from fluo-4. Ca(2+) signals and sarcomere strain correlated in space and time with short delays. Furthermore, Ca(2+) sparks and waves caused contractions in subcellular microdomains, revealing a previously underappreciated role for these events in generating subcellular strain during diastole. Ca(2+) activity and sarcomere strain were also imaged in paced cardiac myocytes under mechanical load, revealing spontaneous Ca(2+) waves and correlated local contraction in pressure-overload-induced cardiomyopathy. CONCLUSIONS: Multimodal second harmonic generation 2-photon fluorescence microscopy enables the simultaneous observation of Ca(2+) release and mechanical strain at the subsarcomere level in living cardiac myocytes. The method benefits from the label-free nature of second harmonic generation, which allows A-bands to be imaged independently of T-tubule morphology and simultaneously with Ca(2+) indicators. Second harmonic generation 2-photon fluorescence imaging is widely applicable to the study of Ca(2+)-contraction coupling and mechanochemotransduction in both health and disease.

Non-linear optical flow cytometry using a scanned, Bessel beam light-sheet.

Modern flow cytometry instruments have become vital tools for high-throughput analysis of single cells. However, as issues with the cellular labeling techniques often used in flow cytometry have become more of a concern, the development of label-free modalities for cellular analysis is increasingly desired. Non-linear optical phenomena (NLO) are of growing interest for label-free analysis because of the ability to measure the intrinsic optical response of biomolecules found in cells. We demonstrate that a light-sheet consisting of a scanned Bessel beam is an optimal excitation geometry for efficiently generating NLO signals in a microfluidic environment. The balance of photon density and cross-sectional area provided by the light-sheet allowed significantly larger two-photon fluorescence intensities to be measured in a model polystyrene microparticle system compared to measurements made using other excitation focal geometries, including a relaxed Gaussian excitation beam often used in conventional flow cytometers.

Label-free identification and characterization of human pluripotent stem cell-derived cardiomyocytes using second harmonic generation (SHG) microscopy.

Pluripotent stem cell-derived cardiomyocytes (PSC-CMs) are a potentially unlimited source of cardiomyocytes (CMs) for cardiac transplantation therapies. The establishment of pure PSC-CM populations is important for this application, but is hampered by a lack of CM-specific surface markers suitable for their identification and sorting. Contemporary purification techniques are either non-specific or require genetic modification. We report a second harmonic generation (SHG) signal detectable in PSC-CMs that is attributable to sarcomeric myosin, dependent on PSC-CM maturity, and retained while PSC-CMs are in suspension. Our study demonstrates the feasibility of developing a SHG-activated flow cytometer for the non-invasive purification of PSC-CMs.

Detection of morphological markers of vulnerable atherosclerotic plaque using multimodal spectroscopy.

Vulnerable plaques, which are responsible for most acute ischemic events, are presently invisible to x-ray angiography. Their primary morphological features include a thin or ulcerated fibrous cap, a large necrotic core, superficial foam cells, and intraplaque hemorrhage. We present evidence that multimodal spectroscopy (MMS), a novel method that combines diffuse reflectance spectroscopy (DRS), intrinsic fluorescence spectroscopy (IFS), and Raman spectroscopy (RS), can detect these markers of plaque vulnerability. To test this concept, we perform an MMS feasibility study on 17 human carotid artery specimens. Following the acquisition of spectra, each specimen is histologically evaluated. Two parameters from DRS, hemoglobin concentration and a scattering parameter, are used to detect intraplaque hemorrhage and foam cells; an IFS parameter that relates to the amount of collagen in the topmost layers of the tissue is used to detect the presence of a thin fibrous cap; and an RS parameter related to the amount of cholesterol and necrotic material is used to detect necrotic core. Taken together, these spectral parameters can generally identify the vulnerable plaques. The results indicate that MMS provides depth-sensitive and complementary morphological information about plaque composition. A prospective in vivo study will be conducted to validate these findings.