CS & EE Seminar: Digital Biomarker Discovery for Continuous and Non-...03-01-2020

Speaker: Beren Semiz, Georgia Tech

Title:  Digital Biomarker Discovery for Continuous and Non-Invasive Health Monitoring

Date/Time: January 3, 2020  /  14.40-15.30

Place: FENS G035

Abstract: In the United States, 6.5 million adults live with heart failure and more than 70,000 patients have advanced heart failure requiring heart transplantation. While the number of patients awaiting a transplant has doubled in the last 15 years, the worldwide availability of donor hearts has decreased by a third, thus necessitating the use of mechanical assist devices, such as left ventricular assist devices (LVADs). The number of patients implanted with LVADs has thus grown rapidly, with more than 15,000 patients currently living with the support of LVADs. However, up to 15% of patients are readmitted to clinics due to development of pump thrombosis (occurrence of blood clotting within the pump). In this talk, I will describe a novel method for remote monitoring of patients with LVADs to enable early detection and tracking of pump thrombosis. This monitoring is achieved by deriving digital biomarkers, which are measured through home-based sensors and wearables. I will show that acoustical features extracted from the pump sounds have high correlation with the actual blood biomarkers, and when combined with machine learning algorithms, they can improve the detection accuracy of suspected thrombosis. Additionally, I will discuss how the scoring of post-treatment data can suggest residual pump thrombosis, which can otherwise not be identified by pump parameters or blood biomarkers.

Bio: Beren Semiz is a fourth year PhD student at Georgia Institute of Technology in Electrical and Computer Engineering Department. She received the M.Sc. degree in Electrical and Computer Engineering from Georgia Institute of Technology in 2018, and B.Sc. degree in Electrical and Electronics Engineering from Koc University in 2016 where she was a Vehbi Koc Scholar. Beren is a recipient of the Fulbright Scholarship. Her research interests include non-invasive wearable device design, biomedical signal processing, and applied machine learning to derive clinically useful digital biomarkers for non-invasive and continuous health monitoring. Her research projects received mainstream media coverage in the United States, including CNN and Fox. Her interdisciplinary research has led to several publications in major engineering venues (IEEE JBHI, IEEE Sensors, and IEEE BHI/BSN) and leading medical conferences (ACC, HFSA, ACR/ARP). Her work on Acoustical Analysis of Left Ventricular Assist Devices was selected as one of the spotlight presentations in the Heart Failure Society of America 2018 Meeting (3.8% selection rate).

Contact: Öznur Taştan