Proteomic Profile of the ICAM1 p.K56M HFpEF Risk Variant.

A common missense variant in ICAM1 among African American individuals (rs5491; pK56M) has been associated with risk of heart failure with preserved ejection fraction (HFpEF), but the pathways that lead to HFpEF among those with this variant are not clear. In this analysis of 92 circulating proteins and their associated networks, we identified 7 circulating inflammatory proteins associated with rs5491 among >600 African American individuals. Using weighted coexpression network analysis, 3 protein networks were identified, one of which was associated with rs5491. This protein network was most highly represented by members of the tumor necrosis receptor superfamily. The rs5491 variant demonstrated an inflammatory proteomic profile in a separate cohort of African American individuals. This analysis identifies inflammatory pathways that may drive HFpEF among African American individuals with the ICAM1 pK56M (rs5491) variant.

Clinical Correlates and Prognostic Impact of Cognitive Dysfunction in Patients With Heart Failure and Preserved Ejection Fraction: Insights From PARAGON-HF.

BACKGROUND: Cognitive impairment is common in patients with heart failure and preserved ejection fraction but its clinical correlates and prognostic associations are poorly understood. METHODS: We analyzed cognitive function, using the Mini-Mental State Examination (MMSE), in patients with heart failure and preserved ejection fraction enrolled in a prespecified substudy of the PARAGON-HF trial (Prospective Comparison of Angiotensin Receptor Neprilysin Inhibitor With Angiotensin Receptor Blocker Global Outcomes in Heart Failure With Preserved Ejection Fraction). Logistic regression analyses were performed to determine the variables associated with lower MMSE scores at baseline and postbaseline decline in MMSE scores at 48 weeks. Cox proportional hazards regression and semiparametric proportional rates models were used to examine the risk of clinical outcomes related to baseline MMSE scores, and decline in MMSE scores during follow-up, adjusted for prognostic variables including NT-proBNP (N-terminal pro-B-type natriuretic peptide). RESULTS: At baseline, cognitive function was normal (MMSE score 28-30) in 1809 of 2895 patients (62.5%), borderline (score 24-27) in 794 (27.4%), and impaired (score <24) in 292 (10.1%). Variables associated with both a lower MMSE score at baseline and a decline in score from baseline included older age, a history of stroke or transient ischemia attack, and lower serum albumin. Compared with those with baseline MMSE scores of 28 to 30, patients in the lower MMSE score categories had a stepwise increase in the risk of the composite of time to first HF hospitalization or cardiovascular death, with an adjusted hazard ratio of 1.27 (95% CI, 1.06-1.53) for those with scores of 24 to 27 and 1.58 (95% CI, 1.21-2.06) for those with scores <24, respectively. These associations were also found for the individual components of the composite and all-cause death. Likewise, cognitive impairment was associated with a 50% higher risk of total (first and repeat) heart failure hospitalizations and cardiovascular deaths. Examining the change in MMSE score from baseline, a decrease in MMSE score during follow-up was associated with a higher risk of death. CONCLUSIONS: In patients with heart failure and preserved ejection fraction, even modest baseline impairment of cognitive function was associated with worse outcomes, including death. A decline in MMSE score during follow-up was a strong predictor of mortality, independent of other prognostic variables.

Aptamer Proteomics for Biomarker Discovery in Heart Failure With Preserved Ejection Fraction: The PARAGON-HF Proteomic Substudy.

BACKGROUND: Prognostic markers and biological pathways linked to detrimental clinical outcomes in heart failure with preserved ejection fraction (HFpEF) remain incompletely defined. METHODS AND RESULTS: We measured serum levels of 4123 unique proteins in 1117 patients with HFpEF enrolled in the PARAGON-HF (Efficacy and Safety of LCZ696 Compared to Valsartan, on Morbidity and Mortality in Heart Failure Patients With Preserved Ejection Fraction) trial using a modified aptamer proteomic assay. Baseline circulating protein concentrations significantly associated with the primary end point and the timing and occurrence of total heart failure hospitalization and cardiovascular death were identified by recurrent events regression, accounting for multiple testing, adjusted for age, sex, treatment, and anticoagulant use, and compared with published analyses in 2515 patients with heart failure with reduced ejection fraction from the PARADIGM-HF (Prospective Comparison of ARNI With ACEI to Determine Impact on Global Mortality and Morbidity in Heart Failure) and ATMOSPHERE (Efficacy and Safety of Aliskiren and Aliskiren/Enalapril Combination on Morbidity-Mortality in Patients With Chronic Heart Failure) clinical trials. We identified 288 proteins that were robustly associated with the risk of heart failure hospitalization and cardiovascular death in patients with HFpEF. The baseline proteins most strongly related to outcomes included B2M (ß-2 microglobulin), TIMP1 (tissue inhibitor of matrix metalloproteinase 1), SERPINA4 (serpin family A member 4), and SVEP1 (sushi, von Willebrand factor type A, EGF, and pentraxin domain containing 1). Overall, the protein-outcome associations in patients with HFpEF did not markedly differ as compared with patients with heart failure with reduced ejection fraction. A proteomic risk score derived in patients with HFpEF was not superior to a previous proteomic score derived in heart failure with reduced ejection fraction nor to clinical risk factors, NT-proBNP (N-terminal pro-B-type natriuretic peptide), or high-sensitivity cardiac troponin. CONCLUSIONS: Numerous serum proteins linked to metabolic, coagulation, and extracellular matrix regulatory pathways were associated with worse HFpEF prognosis in the PARAGON-HF proteomic substudy. Our results demonstrate substantial similarities among serum proteomic risk markers for heart failure hospitalization and cardiovascular death when comparing clinical trial participants with heart failure across the ejection fraction spectrum. REGISTRATION: URL: https://www.clinicaltrials.gov; Unique Identifiers: NCT01920711, NCT01035255, NCT00853658.

Natural Language Processing for Adjudication of Heart Failure in a Multicenter Clinical Trial: A Secondary Analysis of a Randomized Clinical Trial.

Importance: The gold standard for outcome adjudication in clinical trials is medical record review by a physician clinical events committee (CEC), which requires substantial time and expertise. Automated adjudication of medical records by natural language processing (NLP) may offer a more resource-efficient alternative but this approach has not been validated in a multicenter setting. Objective: To externally validate the Community Care Cohort Project (C3PO) NLP model for heart failure (HF) hospitalization adjudication, which was previously developed and tested within one health care system, compared to gold-standard CEC adjudication in a multicenter clinical trial. Design, Setting, and Participants: This was a retrospective analysis of the Influenza Vaccine to Effectively Stop Cardio Thoracic Events and Decompensated Heart Failure (INVESTED) trial, which compared 2 influenza vaccines in 5260 participants with cardiovascular disease at 157 sites in the US and Canada between September 2016 and January 2019. Analysis was performed from November 2022 to October 2023. Exposures: Individual sites submitted medical records for each hospitalization. The central INVESTED CEC and the C3PO NLP model independently adjudicated whether the cause of hospitalization was HF using the prepared hospitalization dossier. The C3PO NLP model was fine-tuned (C3PO + INVESTED) and a de novo NLP model was trained using half the INVESTED hospitalizations. Main Outcomes and Measures: Concordance between the C3PO NLP model HF adjudication and the gold-standard INVESTED CEC adjudication was measured by raw agreement, κ, sensitivity, and specificity. The fine-tuned and de novo INVESTED NLP models were evaluated in an internal validation cohort not used for training. Results: Among 4060 hospitalizations in 1973 patients (mean [SD] age, 66.4 [13.2] years; 514 [27.4%] female and 1432 [72.6%] male]), 1074 hospitalizations (26%) were adjudicated as HF by the CEC. There was good agreement between the C3PO NLP and CEC HF adjudications (raw agreement, 87% [95% CI, 86-88]; κ, 0.69 [95% CI, 0.66-0.72]). C3PO NLP model sensitivity was 94% (95% CI, 92-95) and specificity was 84% (95% CI, 83-85). The fine-tuned C3PO and de novo NLP models demonstrated agreement of 93% (95% CI, 92-94) and κ of 0.82 (95% CI, 0.77-0.86) and 0.83 (95% CI, 0.79-0.87), respectively, vs the CEC. CEC reviewer interrater reproducibility was 94% (95% CI, 93-95; κ, 0.85 [95% CI, 0.80-0.89]). Conclusions and Relevance: The C3PO NLP model developed within 1 health care system identified HF events with good agreement relative to the gold-standard CEC in an external multicenter clinical trial. Fine-tuning the model improved agreement and approximated human reproducibility. Further study is needed to determine whether NLP will improve the efficiency of future multicenter clinical trials by identifying clinical events at scale.

Natural Language Processing for Adjudication of Heart Failure Hospitalizations in a Multi-Center Clinical Trial.

Background: The gold standard for outcome adjudication in clinical trials is chart review by a physician clinical events committee (CEC), which requires substantial time and expertise. Automated adjudication by natural language processing (NLP) may offer a more resource-efficient alternative. We previously showed that the Community Care Cohort Project (C3PO) NLP model adjudicates heart failure (HF) hospitalizations accurately within one healthcare system. Methods: This study externally validated the C3PO NLP model against CEC adjudication in the INVESTED trial. INVESTED compared influenza vaccination formulations in 5260 patients with cardiovascular disease at 157 North American sites. A central CEC adjudicated the cause of hospitalizations from medical records. We applied the C3PO NLP model to medical records from 4060 INVESTED hospitalizations and evaluated agreement between the NLP and final consensus CEC HF adjudications. We then fine-tuned the C3PO NLP model (C3PO+INVESTED) and trained a de novo model using half the INVESTED hospitalizations, and evaluated these models in the other half. NLP performance was benchmarked to CEC reviewer inter-rater reproducibility. Results: 1074 hospitalizations (26%) were adjudicated as HF by the CEC. There was high agreement between the C3PO NLP and CEC HF adjudications (agreement 87%, kappa statistic 0.69). C3PO NLP model sensitivity was 94% and specificity was 84%. The fine-tuned C3PO and de novo NLP models demonstrated agreement of 93% and kappa of 0.82 and 0.83, respectively. CEC reviewer inter-rater reproducibility was 94% (kappa 0.85). Conclusion: Our NLP model developed within a single healthcare system accurately identified HF events relative to the gold-standard CEC in an external multi-center clinical trial. Fine-tuning the model improved agreement and approximated human reproducibility. NLP may improve the efficiency of future multi-center clinical trials by accurately identifying clinical events at scale.

Outpatient Worsening Among Patients With Mildly Reduced and Preserved Ejection Fraction Heart Failure in the DELIVER Trial.

BACKGROUND: Hospitalization is recognized as a sentinel event in the disease trajectory of patients with heart failure (HF), but not all patients experiencing clinical decompensation are ultimately hospitalized. Outpatient intensification of diuretics is common in response to symptoms of worsening HF, yet its prognostic and clinical relevance, specifically for patients with HF with mildly reduced or preserved ejection fraction, is uncertain. METHODS: In this prespecified analysis of the DELIVER trial (Dapagliflozin Evaluation to Improve the Lives of Patients With Preserved Ejection Fraction Heart Failure), we assessed the association between various nonfatal worsening HF events (those requiring hospitalization, urgent outpatient visits requiring intravenous HF therapies, and outpatient oral diuretic intensification) and rates of subsequent mortality. We further examined the treatment effect of dapagliflozin on an expanded composite end point of cardiovascular death, HF hospitalization, urgent HF visit, or outpatient oral diuretic intensification. RESULTS: In DELIVER, 4532 (72%) patients experienced no worsening HF event, whereas 789 (13%) had outpatient oral diuretic intensification, 86 (1%) required an urgent HF visit, 585 (9%) had an HF hospitalization, and 271 (4%) died of cardiovascular causes as a first presentation. Patients with a first presentation manifesting as outpatient oral diuretic intensification experienced rates of subsequent mortality that were higher (10 [8-12] per 100 patient-years) than those without a worsening HF event (4 [3-4] per 100 patient-years) but similar to rates of subsequent death after an urgent HF visit (10 [6-18] per 100 patient-years). Patients with an HF hospitalization as a first presentation of worsening HF had the highest rates of subsequent death (35 [31-40] per 100 patient-years). The addition of outpatient diuretic intensification to the adjudicated DELIVER primary end point (cardiovascular death, HF hospitalization, or urgent HF visit) increased the overall number of patients experiencing an event from 1122 to 1731 (a 54% increase). Dapagliflozin reduced the need for outpatient diuretic intensification alone (hazard ratio, 0.72 [95% CI, 0.64-0.82]) and when analyzed as a part of an expanded composite end point of worsening HF or cardiovascular death (hazard ratio, 0.76 [95% CI, 0.69-0.84]). CONCLUSIONS: In patients with HF with mildly reduced or preserved ejection fraction, worsening HF requiring oral diuretic intensification in ambulatory care was frequent, adversely prognostic, and significantly reduced by dapagliflozin. REGISTRATION: URL: https://www.clinicaltrials.gov; Unique identifier: NCT03619213.

Incremental prognostic value of biomarkers in PARADIGM-HF.

AIMS: It is uncertain how much candidate biomarkers improve risk prediction when added to comprehensive models including routinely collected clinical and laboratory variables in heart failure. METHODS AND RESULTS: Aldosterone, cystatin C, high-sensitivity troponin T (hs-TnT), galectin-3, growth differentiation factor-15 (GDF-15), kidney injury molecule-1, matrix metalloproteinase-2 and -9, soluble suppression of tumourigenicity-2, tissue inhibitor of metalloproteinase-1 (TIMP-1) and urinary albumin to creatinine ratio were measured in 1559 of PARADIGM-HF participants. We tested whether these biomarkers, individually or collectively, improved the performance of the PREDICT-HF prognostic model, which includes clinical, routine laboratory, and natriuretic peptide data, for the primary endpoint and cardiovascular and all-cause mortality. The mean age of participants was 67.3 ± 9.9 years, 1254 (80.4%) were men and 1103 (71%) were in New York Heart Association class II. During a mean follow-up of 30.7 months, 300 patients experienced the primary outcome and 197 died. Added individually, only four biomarkers were independently associated with all outcomes: hs-TnT, GDF-15, cystatin C and TIMP-1. When all biomarkers were added simultaneously to the PREDICT-HF models, only hs-TnT remained an independent predictor of all three endpoints. GDF-15 also remained predictive of the primary endpoint; TIMP-1 was the only other predictor of both cardiovascular and all-cause mortality. Individually or in combination, these biomarkers did not lead to significant improvements in discrimination or reclassification. CONCLUSIONS: None of the biomarkers studied individually or collectively led to a meaningful improvement in the prediction of outcomes over what is provided by clinical, routine laboratory, and natriuretic peptide variables.

Renal and blood pressure effects of dapagliflozin in recently hospitalized patients with heart failure with mildly reduced or preserved ejection fraction: Insights from the DELIVER trial.

AIMS: Patients recently hospitalized for heart failure (HF) often have unstable haemodynamics and experience worsening renal failure, and are at elevated risk for recurrent HF events. In DELIVER, dapagliflozin reduced HF events or cardiovascular death including among patients who were hospitalized or recently hospitalized. METHODS AND RESULTS: We examined the effects of dapagliflozin versus placebo on estimated glomerular filtration rate (eGFR) slope (acute and chronic), 1-month change in systolic blood pressure, and the occurrence of serious hypovolaemic or renal adverse events in patients with and without HF hospitalization within 30 days of randomization. The 654 (90 randomized during hospitalization, 147 1-7 days post-discharge and 417 8-30 days post-discharge) recently hospitalized patients had lower baseline eGFR compared with those without recent HF hospitalization (median [interquartile range] 55 [43, 71] vs. 60 [47, 75] ml/min/1.73 m2 ). Dapagliflozin consistently reduced the risk of all-cause (pinteraction = 0.20), cardiac-related (pinteraction = 0.75), and HF-specific (pinteraction = 0.90) hospitalizations, irrespective of recent HF hospitalization. In those recently hospitalized, acute placebo-corrected eGFR reductions with dapagliflozin were modest and similar to patients without recent hospitalization (-2.0 [-4.1, +0.1] vs. -3.4 [-3.9, -2.9] ml/min/1.73 m2 , pinteraction = 0.12). Dapagliflozin's effect to slow chronic eGFR decline was similar regardless of recent hospitalization (pinteraction = 0.57). Dapagliflozin had a minimal effect on 1-month systolic blood pressure and to a similar degree in patients with and without recent hospitalization (-1.3 vs.-1.8 mmHg, pinteraction = 0.64). There was no treatment-related excess in renal or hypovolaemic serious adverse events, irrespective of recent HF hospitalization. CONCLUSION: In patients recently hospitalized with HF, initiation of dapagliflozin had minimal effects on blood pressure and did not increase renal or hypovolaemic serious adverse events, yet afforded long-term cardiovascular and kidney protective effects. These data suggest that the benefit to risk ratio favours initiation of dapagliflozin among stabilized patients hospitalized or recently hospitalized for HF. CLINICAL TRIAL REGISTRATION: ClinicalTrials.gov NCT03619213.

Genetics of myocardial interstitial fibrosis in the human heart and association with disease.

Myocardial interstitial fibrosis is associated with cardiovascular disease and adverse prognosis. Here, to investigate the biological pathways that underlie fibrosis in the human heart, we developed a machine learning model to measure native myocardial T1 time, a marker of myocardial fibrosis, in 41,505 UK Biobank participants who underwent cardiac magnetic resonance imaging. Greater T1 time was associated with diabetes mellitus, renal disease, aortic stenosis, cardiomyopathy, heart failure, atrial fibrillation, conduction disease and rheumatoid arthritis. Genome-wide association analysis identified 11 independent loci associated with T1 time. The identified loci implicated genes involved in glucose transport (SLC2A12), iron homeostasis (HFE, TMPRSS6), tissue repair (ADAMTSL1, VEGFC), oxidative stress (SOD2), cardiac hypertrophy (MYH7B) and calcium signaling (CAMK2D). Using a transforming growth factor ß1-mediated cardiac fibroblast activation assay, we found that 9 of the 11 loci consisted of genes that exhibited temporal changes in expression or open chromatin conformation supporting their biological relevance to myofibroblast cell state acquisition. By harnessing machine learning to perform large-scale quantification of myocardial interstitial fibrosis using cardiac imaging, we validate associations between cardiac fibrosis and disease, and identify new biologically relevant pathways underlying fibrosis.

Virtual Care Team Guided Management of Patients With Heart Failure During Hospitalization.

BACKGROUND: Scalable and safe approaches for heart failure guideline-directed medical therapy (GDMT) optimization are needed. OBJECTIVES: The authors assessed the safety and effectiveness of a virtual care team guided strategy on GDMT optimization in hospitalized patients with heart failure with reduced ejection fraction (HFrEF). METHODS: In a multicenter implementation trial, we allocated 252 hospital encounters in patients with left ventricular ejection fraction ≤40% to a virtual care team guided strategy (107 encounters among 83 patients) or usual care (145 encounters among 115 patients) across 3 centers in an integrated health system. In the virtual care team group, clinicians received up to 1 daily GDMT optimization suggestion from a physician-pharmacist team. The primary effectiveness outcome was in-hospital change in GDMT optimization score (+2 initiations, +1 dose up-titrations, -1 dose down-titrations, -2 discontinuations summed across classes). In-hospital safety outcomes were adjudicated by an independent clinical events committee. RESULTS: Among 252 encounters, the mean age was 69 ± 14 years, 85 (34%) were women, 35 (14%) were Black, and 43 (17%) were Hispanic. The virtual care team strategy significantly improved GDMT optimization scores vs usual care (adjusted difference: +1.2; 95% CI: 0.7-1.8; P < 0.001). New initiations (44% vs 23%; absolute difference: +21%; P = 0.001) and net intensifications (44% vs 24%; absolute difference: +20%; P = 0.002) during hospitalization were higher in the virtual care team group, translating to a number needed to intervene of 5 encounters. Overall, 23 (21%) in the virtual care team group and 40 (28%) in usual care experienced 1 or more adverse events (P = 0.30). Acute kidney injury, bradycardia, hypotension, hyperkalemia, and hospital length of stay were similar between groups. CONCLUSIONS: Among patients hospitalized with HFrEF, a virtual care team guided strategy for GDMT optimization was safe and improved GDMT across multiple hospitals in an integrated health system. Virtual teams represent a centralized and scalable approach to optimize GDMT.

Missense Genetic Variation of ICAM1 and Incident Heart Failure.

BACKGROUND: Intercellular adhesion molecule-1 (ICAM-1) is a cell surface protein that participates in endothelial activation and is hypothesized to play a central role in heart failure (HF). We evaluated associations of ICAM1 missense genetic variants with circulating ICAM-1 levels and with incident HF. METHODS AND RESULTS: We identified 3 missense variants within ICAM1 (rs5491, rs5498 and rs1799969) and evaluated their associations with ICAM-1 levels in the Coronary Artery Risk Development in Young Adults Study and the Multi-Ethnic Study of Atherosclerosis (MESA). We determined the association among these 3 variants and incident HF in MESA. We separately evaluated significant associations in the Atherosclerosis Risk in Communities (ARIC) study. Of the 3 missense variants, rs5491 was common in Black participants (minor allele frequency [MAF] > 20%) and rare in other race/ethnic groups (MAF < 5%). In Black participants, the presence of rs5491 was associated with higher levels of circulating ICAM-1 at 2 timepoints separated by 8 years. Among Black participants in MESA (n = 1600), the presence of rs5491 was associated with an increased risk of incident HF with preserved ejection fraction (HFpEF; HR = 2.30; [95% CI 1.25-4.21; P = 0.007]). The other ICAM1 missense variants (rs5498 and rs1799969) were associated with ICAM-1 levels, but there were no associations with HF. In ARIC, rs5491 was significantly associated with incident HF (HR = 1.24 [95% CI 1.02 - 1.51]; P = 0.03), with a similar direction of effect for HFpEF that was not statistically significant. CONCLUSIONS: A common ICAM1 missense variant among Black individuals may be associated with increased risk of HF, which may be HFpEF-specific.

Machine Learning to Understand Genetic and Clinical Factors Associated With the Pulse Waveform Dicrotic Notch.

BACKGROUND: Absence of a dicrotic notch on finger photoplethysmography is an easily ascertainable and inexpensive trait that has been associated with age and prevalent cardiovascular disease. However, the trait exists along a continuum, and little is known about its genetic underpinnings or prognostic value for incident cardiovascular disease. METHODS: In 169 787 participants in the UK Biobank, we identified absent dicrotic notch on photoplethysmography and created a novel continuous trait reflecting notch smoothness using machine learning. Next, we determined the heritability, genetic basis, polygenic risk, and clinical relations for the binary absent notch trait and the newly derived continuous notch smoothness trait. RESULTS: Heritability of the continuous notch smoothness trait was 7.5%, compared with 5.6% for the binary absent notch trait. A genome-wide association study of notch smoothness identified 15 significant loci, implicating genes including NT5C2 (P=1.2×10-26), IGFBP3 (P=4.8×10-18), and PHACTR1 (P=1.4×10-13), compared with 6 loci for the binary absent notch trait. Notch smoothness stratified risk of incident myocardial infarction or coronary artery disease, stroke, heart failure, and aortic stenosis. A polygenic risk score for notch smoothness was associated with incident cardiovascular disease and all-cause death in UK Biobank participants without available photoplethysmography data. CONCLUSIONS: We found that a machine learning derived continuous trait reflecting dicrotic notch smoothness on photoplethysmography was heritable and associated with genes involved in vascular stiffness. Greater notch smoothness was associated with greater risk of incident cardiovascular disease. Raw digital phenotyping may identify individuals at risk for disease via specific genetic pathways.

Cardiac Xenotransplantation: Challenges, Evolution, and Advances.

The increased need for heart transplantation in patients with advanced heart failure has introduced demand for a greater supply of donor hearts. Progress in cross-species experimental models has led to promise for ushering in the clinical use of xenotransplantation (XTx) as a potential solution to the organ shortage worldwide. In this review, the authors first highlight the historical advances that led to the first pig-to-human heart transplantation, a landmark moment in the field of advanced heart failure. The authors discuss immunologic, infectious, and physiological challenges for implementation of XTx, as well as innovations in the science of genetic manipulation that allowed clinical translation of this therapy. The authors consider ongoing barriers that affect ongoing translation of this technology into clinical care in the current era. Finally, the authors propose a framework for advancing clinical application of XTx.

Coronary Artery Disease and Heart Failure With Preserved Ejection Fraction: The ARIC Study.

Background Whether coronary artery disease (CAD) is a significant risk factor for heart failure (HF) with preserved ejection fraction (HFpEF) is unclear. Methods and Results Among 9902 participants in the ARIC (Atherosclerosis Risk in Communities) study, we assessed the association of incident CAD with subsequent incident HFpEF (left ventricular ejection fraction [≥50%]) and HF with reduced ejection fraction (HFrEF; left ventricular ejection fraction <50%) using survival models with time-updated variables. We also assessed the extent to which echocardiographic correlates of prevalent CAD account for the relationship between CAD and incident HFpEF. Over 13-year follow-up, incident CAD developed in 892 participants and 178 subsequently developed HF (86 HFrEF, 71 HFpEF). Incident HFrEF and HFpEF risk were both greatest early after the CAD event. At >1 year post-CAD event, adjusted incidence of HFrEF and HFpEF were similar (7.2 [95% CI, 5.2-10.0] and 6.7 [4.8-9.2] per 1000 person-years, respectively) and CAD remained predictive of both (HFrEF: hazard ratio, 2.76 [95% CI, 1.99-3.84]; HFpEF: 1.85 [1.35-2.54]) after adjusting for demographics and common comorbidities. Among 4779 HF-free participants at Visit 5 (2011-2013), the 490 with prevalent CAD had lower left ventricular ejection fraction and higher left ventricular mass index, E/e', and left atrial volume index (all P<0.01). The association of prevalent CAD with incident HFpEF post-Visit 5 was not significant after adjusting for echocardiographic measures, with the greatest attenuation observed for left ventricular diastolic function. Conclusions CAD is a significant risk factor for incident HFpEF after adjustment for demographics and common comorbidities. This relationship is partially accounted for by echocardiographic alterations, particularly left ventricular diastolic function.

One Clinician Is All You Need-Cardiac Magnetic Resonance Imaging Measurement Extraction: Deep Learning Algorithm Development.

BACKGROUND: Cardiac magnetic resonance imaging (CMR) is a powerful diagnostic modality that provides detailed quantitative assessment of cardiac anatomy and function. Automated extraction of CMR measurements from clinical reports that are typically stored as unstructured text in electronic health record systems would facilitate their use in research. Existing machine learning approaches either rely on large quantities of expert annotation or require the development of engineered rules that are time-consuming and are specific to the setting in which they were developed. OBJECTIVE: We hypothesize that the use of pretrained transformer-based language models may enable label-efficient numerical extraction from clinical text without the need for heuristics or large quantities of expert annotations. Here, we fine-tuned pretrained transformer-based language models on a small quantity of CMR annotations to extract 21 CMR measurements. We assessed the effect of clinical pretraining to reduce labeling needs and explored alternative representations of numerical inputs to improve performance. METHODS: Our study sample comprised 99,252 patients that received longitudinal cardiology care in a multi-institutional health care system. There were 12,720 available CMR reports from 9280 patients. We adapted PRAnCER (Platform Enabling Rapid Annotation for Clinical Entity Recognition), an annotation tool for clinical text, to collect annotations from a study clinician on 370 reports. We experimented with 5 different representations of numerical quantities and several model weight initializations. We evaluated extraction performance using macroaveraged F1-scores across the measurements of interest. We applied the best-performing model to extract measurements from the remaining CMR reports in the study sample and evaluated established associations between selected extracted measures with clinical outcomes to demonstrate validity. RESULTS: All combinations of weight initializations and numerical representations obtained excellent performance on the gold-standard test set, suggesting that transformer models fine-tuned on a small set of annotations can effectively extract numerical quantities. Our results further indicate that custom numerical representations did not appear to have a significant impact on extraction performance. The best-performing model achieved a macroaveraged F1-score of 0.957 across the evaluated CMR measurements (range 0.92 for the lowest-performing measure of left atrial anterior-posterior dimension to 1.0 for the highest-performing measures of left ventricular end systolic volume index and left ventricular end systolic diameter). Application of the best-performing model to the study cohort yielded 136,407 measurements from all available reports in the study sample. We observed expected associations between extracted left ventricular mass index, left ventricular ejection fraction, and right ventricular ejection fraction with clinical outcomes like atrial fibrillation, heart failure, and mortality. CONCLUSIONS: This study demonstrated that a domain-agnostic pretrained transformer model is able to effectively extract quantitative clinical measurements from diagnostic reports with a relatively small number of gold-standard annotations. The proposed workflow may serve as a roadmap for other quantitative entity extraction.