An electrocardiogram-based AI algorithm for early detection of pulmonary hypertension.

BACKGROUND: Early diagnosis of pulmonary hypertension (PH) is critical for effective treatment and management. We aimed to develop and externally validate an artificial intelligence algorithm that could serve as a PH screening tool, based on analysis of a standard 12-lead electrocardiogram (ECG). METHODS: The PH Early Detection Algorithm (PH-EDA) is a convolutional neural network developed using retrospective ECG voltage-time data, with patients classified as "PH-likely" or "PH-unlikely" (controls) based on right heart catheterisation or echocardiography. In total, 39 823 PH-likely patients and 219 404 control patients from Mayo Clinic were randomly split into training (48%), validation (12%), and test (40%) sets. ECGs taken within 1 month of PH diagnosis (diagnostic dataset) were used to train the PH-EDA at Mayo Clinic. Performance was tested on diagnostic ECGs within the test sets from Mayo Clinic (n=16 175/87 998 PH-likely/controls) and Vanderbilt University Medical Center (VUMC; n=6045/24 256 PH-likely/controls). Performance was also tested on ECGs taken 6-18 months (pre-emptive dataset), and up to 5 years prior to a PH diagnosis at both sites. RESULTS: Performance testing yielded an area under the receiver operating characteristic curve (AUC) of 0.92 and 0.88 in the diagnostic test set at Mayo Clinic and VUMC, respectively, and 0.86 and 0.81, respectively, in the pre-emptive test set. The AUC remained a minimum of 0.79 at Mayo Clinic and 0.73 at VUMC up to 5 years before diagnosis. CONCLUSION: The PH-EDA can detect PH at diagnosis and 6-18 months prior, demonstrating the potential to accelerate diagnosis and management of this debilitating disease.

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.

Mapping each pre-existing condition's association to short-term and long-term COVID-19 complications.

Understanding the relationships between pre-existing conditions and complications of COVID-19 infection is critical to identifying which patients will develop severe disease. Here, we leverage ~1.1 million clinical notes from 1803 hospitalized COVID-19 patients and deep neural network models to characterize associations between 21 pre-existing conditions and the development of 20 complications (e.g. respiratory, cardiovascular, renal, and hematologic) of COVID-19 infection throughout the course of infection (i.e. 0-30 days, 31-60 days, and 61-90 days). Pleural effusion was the most frequent complication of early COVID-19 infection (89/1803 patients, 4.9%) followed by cardiac arrhythmia (45/1803 patients, 2.5%). Notably, hypertension was the most significant risk factor associated with 10 different complications including acute respiratory distress syndrome, cardiac arrhythmia, and anemia. The onset of new complications after 30 days is rare and most commonly involves pleural effusion (31-60 days: 11 patients, 61-90 days: 9 patients). Lastly, comparing the rates of complications with a propensity-matched COVID-negative hospitalized population confirmed the importance of hypertension as a risk factor for early-onset complications. Overall, the associations between pre-COVID conditions and COVID-associated complications presented here may form the basis for the development of risk assessment scores to guide clinical care pathways.

A Literature-Derived Knowledge Graph Augments the Interpretation of Single Cell RNA-seq Datasets.

Technology to generate single cell RNA-sequencing (scRNA-seq) datasets and tools to annotate them have advanced rapidly in the past several years. Such tools generally rely on existing transcriptomic datasets or curated databases of cell type defining genes, while the application of scalable natural language processing (NLP) methods to enhance analysis workflows has not been adequately explored. Here we deployed an NLP framework to objectively quantify associations between a comprehensive set of over 20,000 human protein-coding genes and over 500 cell type terms across over 26 million biomedical documents. The resultant gene-cell type associations (GCAs) are significantly stronger between a curated set of matched cell type-marker pairs than the complementary set of mismatched pairs (Mann Whitney p = 6.15 × 10-76, r = 0.24; cohen's D = 2.6). Building on this, we developed an augmented annotation algorithm (single cell Annotation via Literature Encoding, or scALE) that leverages GCAs to categorize cell clusters identified in scRNA-seq datasets, and we tested its ability to predict the cellular identity of 133 clusters from nine datasets of human breast, colon, heart, joint, ovary, prostate, skin, and small intestine tissues. With the optimized settings, the true cellular identity matched the top prediction in 59% of tested clusters and was present among the top five predictions for 91% of clusters. scALE slightly outperformed an existing method for reference data driven automated cluster annotation, and we demonstrate that integration of scALE can meaningfully improve the annotations derived from such methods. Further, contextualization of differential expression analyses with these GCAs highlights poorly characterized markers of well-studied cell types, such as CLIC6 and DNASE1L3 in retinal pigment epithelial cells and endothelial cells, respectively. Taken together, this study illustrates for the first time how the systematic application of a literature-derived knowledge graph can expedite and enhance the annotation and interpretation of scRNA-seq data.

Pre-existing conditions are associated with COVID-19 patients' hospitalization, despite confirmed clearance of SARS-CoV-2 virus.

BACKGROUND: Consecutive negative SARS-CoV-2 PCR test results are being considered to estimate viral clearance in COVID-19 patients. However, there are anecdotal reports of hospitalization from protracted COVID-19 complications despite such confirmed viral clearance, presenting a clinical conundrum. METHODS: We conducted a retrospective analysis of 222 hospitalized COVID-19 patients to compare those that were readmitted post-viral clearance (hospitalized post-clearance cohort, n = 49) with those that were not re-admitted post-viral clearance (non-hospitalized post-clearance cohort, n = 173) between February and October 2020. In order to differentiate these two cohorts, we used neural network models for the 'augmented curation' of comorbidities and complications with positive sentiment in the Electronic Hosptial Records physician notes. FINDINGS: In the year preceding COVID-19 onset, anemia (n = 13 [26.5%], p-value: 0.007), cardiac arrhythmias (n = 14 [28.6%], p-value: 0.015), and acute kidney injury (n = 7 [14.3%], p-value: 0.030) were significantly enriched in the physician notes of the hospitalized post-clearance cohort. INTERPRETATION: Overall, this retrospective study highlights specific pre-existing conditions that are associated with higher hospitalization rates in COVID-19 patients despite viral clearance and motivates follow-up prospective research into the associated risk factors. FUNDING: This work was supported by Nference, inc.

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.

Knowledge synthesis of 100 million biomedical documents augments the deep expression profiling of coronavirus receptors.

The COVID-19 pandemic demands assimilation of all biomedical knowledge to decode mechanisms of pathogenesis. Despite the recent renaissance in neural networks, a platform for the real-time synthesis of the exponentially growing biomedical literature and deep omics insights is unavailable. Here, we present the nferX platform for dynamic inference from over 45 quadrillion possible conceptual associations from unstructured text, and triangulation with insights from single-cell RNA-sequencing, bulk RNA-seq and proteomics from diverse tissue types. A hypothesis-free profiling of ACE2 suggests tongue keratinocytes, olfactory epithelial cells, airway club cells and respiratory ciliated cells as potential reservoirs of the SARS-CoV-2 receptor. We find the gut as the putative hotspot of COVID-19, where a maturation correlated transcriptional signature is shared in small intestine enterocytes among coronavirus receptors (ACE2, DPP4, ANPEP). A holistic data science platform triangulating insights from structured and unstructured data holds potential for accelerating the generation of impactful biological insights and hypotheses.