Automatic cine-based detection of patients at high risk of heart failure with reduced ejection fraction in echocardiograms

ABSTRACT

Heart failure with reduced ejection fraction (HFrEF) is associated with high mortality rates. The identification of patients with HFrEF in urgent and emergent situations can be delayed in the absence of clinicians skilled in acquiring and interpreting echocardiography (echo) images. Although the standard-of-care for measuring ejection fraction (EF) involves segmentation of the left ventricle in echo, experienced interpreters often opt for visual estimation of EF in clinical settings. In this paper, we mimic this process using dual-channel deep neural networks for segmentation-free classification of echo cine loops: low-risk ($40\mathrm{\%}<\mathrm{E}\mathrm{F}\le 75\mathrm{\%}$) vs. high-risk (reduced EF, i.e. $\mathrm{E}\mathrm{F}\le 40\mathrm{\%}$). The proposed architecture utilises densely connected networks for extraction of unsupervised spatial features. Temporal embedding is then achieved by aggregating these frame-level feature vectors and feeding them through recurrent blocks. We use 949 and 237 clinical echo exams to train and validate the proposed method, respectively. We demonstrate that direct cine-based detection of patients at high-risk of HFrEF is feasible using densely connected convolutional and recurrent-based architectures.

KEYWORDS:

• Echocardiography
• recurrent
• convolutional
• ejection fraction

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the Canadian Institutes of Health Research [x].

Notes on contributors

Delaram Behnami

Delaram Behnami is a PhD candidate in Electrical and Computer Engineering at UBC.

Christina Luong

Dr. Christina Luong is a cardiologist and clinical assistant professor in the Cardiology Division at UBC.

Hooman Vaseli

Hooman Vaseli is an undergraduate student in the Department of Electrical and Computer Engineering (ECE) at UBC.

Hany Girgis

Hany Girgis is a cardiology researcher at  Vancouver General Hospital (VGH), and an Associate Professor of Cardiology Fayoum University, Egypt.

Amir Abdi

Amir Abdi is a PhD candidate and machine learning scientist at ECE, UBC.

Dale Hawley

Dale Hawley is a Senior Systems Analyst /Designer, Information Management/Information Technology Services (IMITS) for Provincial Health Services Authority (PHSA), Providence Health Care (PHC), and Vancouver Coastal Health (VCH).

Ken Gin

Ken Gin is a Clinical Professor, Department of Medicine at UBC, Head of the Division of Cardiology VGH Medical Manager of Cardiology Programs Vancouver Acute, Past Director of Atrial Fibrillation Clinic and CCU VGH, Associate Director Coronary Care Unit and Echo at VGH.

Robert Rohling

Robert Rohling is a Professor with a joint appointment in Electrical and Computer Engineering & Mechanical Engineering at UBC. Dr. Rohling's research is in the field of biomedical engineering with a specialization in medical ultrasound. Dr. Rohling is also the Director of the Institute for Computing, Information and Cognitive Systems (ICICS).

Purang Abolmaesumi

Dr. Purang Abolmaesumi is a Professor of Medical Imaging, Image Guided Therapy and Applied Machine Learning at the University of British Columbia.

Teresa Tsang

Dr. Teresa Tsang is a cardiologist, Director of Echocardiography at Vancouver General Hospital and UBC Hospital, and Professor of Medicine,  Associate Head of Research for the Department of Medicine.