Arruda‐Olson, Adelaide M.

The analysis and interpretation of cardiac magnetic resonance (CMR) images are often time-consuming. The automated segmentation of cardiac structures can reduce the time required for image analysis. Spatial similarities between different CMR image types were leveraged to jointly segment multiple sequences using a segmentation model termed a multi-image type UNet (MI-UNet). This model was developed from 72 exams (46% female, mean age 63 ± 11 years) performed on patients with hypertrophic cardiomyopathy. The MI-UNet for steady-state free precession (SSFP) images achieved a superior Dice similarity coefficient (DSC) of 0.92 ± 0.06 compared to 0.87 ± 0.08 for a single-image type UNet (p < 0.001). The MI-UNet for late gadolinium enhancement (LGE) images also had a superior DSC of 0.86 ± 0.11 compared to 0.78 ± 0.11 for a single-image type UNet (p = 0.001). The difference across image types was most evident for the left ventricular myocardium in SSFP images and for both the left ventricular cavity and the left ventricular myocardium in LGE images. For the right ventricle, there were no differences in DCS when comparing the MI-UNet with single-image type UNets. The joint segmentation of multiple image types increases segmentation accuracy for CMR images of the left ventricle compared to single-image models. In clinical practice, the MI-UNet model may expedite the analysis and interpretation of CMR images of multiple types.

Background: Patients with peripheral artery disease (PAD) are at increased risk for major adverse limb and cardiac events including mortality. Developing screening tools capable of accurate PAD identification is a necessary first step for strategies of adverse outcome prevention. This study aimed to determine whether machine analysis of a resting Doppler waveform using deep neural networks can accurately identify patients with PAD. Methods: Consecutive patients (4/8/2015 – 12/31/2020) undergoing rest and postexercise ankle–brachial index (ABI) testing were included. Patients were randomly allocated to training, validation, and testing subsets (70%/15%/15%). Deep neural networks were trained on resting posterior tibial arterial Doppler waveforms to predict normal (> 0.9) or PAD (⩽ 0.9) using rest and postexercise ABI. A separate dataset of 151 patients who underwent testing during a period after the model had been created and validated (1/1/2021 – 3/31/2021) was used for secondary validation. Area under the receiver operating characteristic curves (AUC) were constructed to evaluate test performance. Results: Among 11,748 total patients, 3432 patients met study criteria: 1941 with PAD (mean age 69 ± 12 years) and 1491 without PAD (64 ± 14 years). The predictive model with highest performance identified PAD with an AUC 0.94 (CI = 0.92–0.96), sensitivity 0.83, specificity 0.88, accuracy 0.85, and positive predictive value (PPV) 0.90. Results were similar for the validation dataset: AUC 0.94 (CI = 0.91–0.98), sensitivity 0.91, specificity 0.85, accuracy 0.89, and PPV 0.89 (postexercise ABI comparison). Conclusion: An artificial intelligence-enabled analysis of a resting Doppler arterial waveform permits identification of PAD at a clinically relevant performance level.

Abstract According to the 2020 report of the American Heart Association’s Heart & Stroke Statistics report, nearly 1,000 people are dying daily because of sudden out-of-hospital cardiac arrests and unfortunately, their survival rate is as low as 10%. Hypertrophic Cardiomyopathy (HCM), a relatively rare genetic heart disease is one of these diseases but finding the right patient for the implantation of ICD is still a research question. Implantation of cardioverter-defibrillator (ICD) can save the life of some of these patients. Due to the complexity of the identification of HCM patients, financial burdens, and the clinical risks involved in the ICD implantation procedure, HCM patients will go into a monitoring state before reaching the implantation trigger. Our study cohort shows about 82% of HCM deaths, did not have an ICD, which highlights the need to improve the pre-screening algorithms. In the current paper, we have proposed a new deep learning-based unsupervised clustering technique to facilitate the prioritization of patients to undergo ICD device implantation. This model uses over 900 echocardiographic measurements to find patients who benefit more from the ICD implantation procedure. Our model was trained and tested over 6 years of echo reports collected at Mayo Clinic. This model can be used as a decision support assistant for cardiologists in finding the right HCM patient when decision-making is hard.

With stress echo (SE) 2020 study, a new standard of practice in stress imaging was developed and disseminated: the ABCDE protocol for functional testing within and beyond CAD. ABCDE protocol was the fruit of SE 2020, and is the seed of SE 2030, which is articulated in 12 projects: 1-SE in coronary artery disease (SECAD); 2-SE in diastolic heart failure (SEDIA); 3-SE in hypertrophic cardiomyopathy (SEHCA); 4-SE post-chest radiotherapy and chemotherapy (SERA); 5-Artificial intelligence SE evaluation (AI-SEE); 6-Environmental stress echocardiography and air pollution (ESTER); 7-SE in repaired Tetralogy of Fallot (SETOF); 8-SE in post-COVID-19 (SECOV); 9: Recovery by stress echo of conventionally unfit donor good hearts (RESURGE); 10-SE for mitral ischemic regurgitation (SEMIR); 11-SE in valvular heart disease (SEVA); 12-SE for coronary vasospasm (SESPASM). The study aims to recruit in the next 5 years (2021–2025) ≥10,000 patients followed for ≥5 years (up to 2030) from ≥20 quality-controlled laboratories from ≥10 countries. In this COVID-19 era of sustainable health care delivery, SE2030 will provide the evidence to finally recommend SE as the optimal and versatile imaging modality for functional testing anywhere, any time, and in any patient.

BACKGROUND Epidemiological cohorts can assist with post-marketing surveillance as they provide community-based outcomes data. Concern was raised about the adverse impact of sulfonylureas (SU) drugs on the outcome of diabetes after myocardial infarction (MI). We evaluated the outcome of MI according to treatment among diabetic persons compared to non-diabetic counterparts in a community-based cohort. METHODS All MIs that occurred in Olmsted County, MN between 1979 and 2002 were validated by standardized epidemiological including cardiac pain, biomarkers and ECG Minnesota coding. Diabetes mellitus (DM) was defined by the National Diabetes Data Group criteria. RESULTS Among 2732 persons with MI (mean age 70 ± 12 years, 44% men), 486 (18%) had DM. Subjects with DM were treated with second generation SU (SU2) (24%), insulin (47%) or diet alone (22%) Diabetic subjects were more likely to be overweight, hyperlipidemic and to receive reperfusion therapy during hospitalization than non-diabetic subjects. One year after the MI, 518 subjects died. Diabetic subjects taking SU2 experienced better survival than subjects without diabetes, while subjects treated with diet alone or insulin experienced worse survival. (Figure ). After adjusting for age, sex, reperfusion and duration of DM, an excess risk of death was detected among persons with insulin-treated DM, particularly those with long-standing DM. No association was found for DM patients treated with SU2 or diet. CONCLUSION These data do not support the concern for an adverse impact of SU2 on survival and underscore the potential role of community studies for population surveillance.

Introduction: Stress echocardiography (SE) has an established role in the evaluation of symptomatic patients, but its utility in asymptomatic patients is unclear. Current guidelines discourage the use of SE in most asymptomatic patients, citing lack of data to support its use. We therefore sought to describe the yield of SE in a group of asymptomatic individuals according to risk profile. Hypothesis: We hypothesized that SE would rarely be abnormal in asymptomatic individuals. Methods: We conducted a retrospective cohort study of all asymptomatic individuals who underwent exercise or pharmacologic SE between 2003 and 2013. Reasons for SE included screening for CAD, preoperative evaluation, abnormal ECG, coronary calcification, and assessment of established CAD (defined as prior myocardial infarction or coronary revascularization). We grouped subjects using both the 10-year Framingham risk equation and ATP3-suggested categories of low-risk (<10%), intermediate-risk (10-20%), and high-risk (>20%) for non-diabetic subjects without known CAD and by ACC/AHA Pooled Cohort risk estimates for subjects with and without diabetes. Subjects with established CAD were considered separately. Abnormal echocardiograms were defined as having ischemic, fixed, or mixed defects. Results: The study population comprised 19,170 subjects; 63% men, mean±SD age was 64±13 years, and 52% underwent pharmacologic SE. The likelihood of abnormal SE results was proportional to risk, with the highest yield in those with established CAD (59%). Those with diabetes had a yield similar to those with a 10 year risk of 10-20% by both ATP3/FRS and ACC/AHA. Conclusions: SE abnormalities are commonly seen in asymptomatic individuals; the likelihood of abnormal SE results was proportional to patients’ estimated risk. The frequency of SE abnormalities in diabetic patients is similar to that of intermediate-risk, but not high-risk patients.

Background: Two-dimensional volumetric exercise stress echocardiography (ESE) provides an integrated view of left ventricular (LV) preload reserve through end-diastolic volume (EDV) and LV contractile reserve (LVCR) through end-systolic volume (ESV) changes. Purpose: To assess the dependence of cardiac reserve upon LVCR, EDV, and heart rate (HR) during ESE. Methods: We prospectively performed semi-supine bicycle or treadmill ESE in 1344 patients (age 59.8 ± 11.4 years; ejection fraction = 63 ± 8%) referred for known or suspected coronary artery disease. All patients had negative ESE by wall motion criteria. EDV and ESV were measured by biplane Simpson rule with 2-dimensional echocardiography. Cardiac index reserve was identified by peak-rest value. LVCR was the stress-rest ratio of force (systolic blood pressure by cuff sphygmomanometer/ESV, abnormal values ≤2.0). Preload reserve was defined by an increase in EDV. Cardiac index was calculated as stroke volume index * HR (by EKG). HR reserve (stress/rest ratio) <1.85 identified chronotropic incompetence. Results: Of the 1344 patients, 448 were in the lowest tertile of cardiac index reserve with stress. Of them, 303 (67.6%) achieved HR reserve <1.85; 252 (56.3%) had an abnormal LVCR and 341 (76.1%) a reduction of preload reserve, with 446 patients (99.6%) showing ≥1 abnormality. At binary logistic regression analysis, reduced preload reserve (odds ratio [OR]: 5.610; 95% confidence intervals [CI]: 4.025 to 7.821), chronotropic incompetence (OR: 3.923, 95% CI: 2.915 to 5.279), and abnormal LVCR (OR: 1.579; 95% CI: 1.105 to 2.259) were independently associated with lowest tertile of cardiac index reserve at peak stress. Conclusions: Heart rate assessment and volumetric echocardiography during ESE identify the heterogeneity of hemodynamic phenotypes of impaired chronotropic, preload or LVCR underlying a reduced cardiac reserve.

Abstract Hypertrophic Cardiomyopathy (HCM) is the most common genetic heart disease in the US and is known to cause sudden death (SCD) in young adults. While significant advancements have been made in HCM diagnosis and management, there is a need to identify HCM cases from electronic health record (EHR) data to develop automated tools based on natural language processing guided machine learning (ML) models for accurate HCM case identification to improve management and reduce adverse outcomes of HCM patients. Cardiac Magnetic Resonance (CMR) Imaging, plays a significant role in HCM diagnosis and risk stratification. CMR reports, generated by clinician annotation, offer rich data in the form of cardiac measurements as well as narratives describing interpretation and phenotypic description. The purpose of this study is to develop an NLP-based interpretable model utilizing impressions extracted from CMR reports to automatically identify HCM patients. CMR reports of patients with suspected HCM diagnosis between the years 1995 to 2019 were used in this study. Patients were classified into three categories of yes HCM, no HCM and, possible HCM. A random forest (RF) model was developed to predict the performance of both CMR measurements and impression features to identify HCM patients. The RF model yielded an accuracy of 86% (608 features) and 85% (30 features). These results offer promise for accurate identification of HCM patients using CMR reports from EHR for efficient clinical management transforming health care delivery for these patients.

Abstract Hypertrophic cardiomyopathy (HCM) is a heritable, phenotypically diverse, and often asymptomatic heart muscle disease which is a major cause of sudden cardiac death (SCD) in young adults. The gold-standard for the diagnosis of HCM is echocardiography (echo), which is an ultrasound-based cardiac imaging modality. Across all sites of the Mayo Clinic enterprise, echo images and measurement data are reviewed, interpreted, and reported via the Echo Information Management System (EIMS). The objective of this paper is to develop a machine learning model for the identification of HCM from cardiac measurements obtained by the echo. We developed a novel machine learning model on patient demographic information and echo measurements that were retrieved from the EIMS digital data registry and selected by cardiologists. Random forest (RF) was utilized to investigate the predictive performance of these features on the identification of HCM patients. The HCM cohort consists of 3,548 patients with at least one HCM diagnostic billing code (ICD-9 or ICD-10), from 2014 to 2019. The class labels “HCM yes” and “HCM no” were assigned by manual review of medical records as well as the outcomes of the gold standard imaging tests for HCM diagnosis. The developed model performed well in finding HCM patients with an accuracy of 95%, recall of 99%, and precision of 97%. The F1 score was 98 %, while 4% of patients were misclassified. This model will be translated into clinical practice for a clinical decision support system in EIMS to assist providers in the accurate diagnosis of HCM from echo data automatically while ensuring high-quality echo interpretation.