Research Projects

1) 3D Segmentation of Epicardial Adipose Tissue from Cardiac MRI

Obesity is a wide-ranging health problem and fat depots around the heart have been linked to the risk of cardiovascular disease.  Magnetic Resonance Imaging (MRI) has emerged as the standard for quantifying epicardial fat for clinical studies but requires the analysis of a large number of images in a short time.  We work on 3D segmentation of epicardial adipose tissue (EAT) and the development of tools to aid in studies of EAT and its relationship to cardiovascular disease and risk. 

2) Identification of Epicardial Adipose Tissue Using Echocardiography

Magnetic resonance imaging (MRI) can provide three-dimensional (3D) assessment of EAT, but it is expensive, time-consuming, and is only available at large institutions.
Echocardiography is safe, real-time, inexpensive, and can also be used to quantify cardiac structure and function. The goal of this project is to utilize 3D volumetric information from MRI data to develop a shape-based model to be used in conjunction with real-time echocardiography and advanced processing of the radio-frequency (RF) ultrasound signals for volumetric assessment of EAT. Machine learning algorithms are used to differentiate tissue types based on features from the ultrasound spectra. Leveraging the specific individual strengths of MRI and echocardiography has the potential to yield a more powerful, yet less expensive analysis tool suited for large studies of intervention and their effect on EAT and cardiovascular health.

3) Spectral Analysis of Ultrasound Radiofrequency Signals for the Identification of Intercostal Nerves

Ultrasound has long been used to assess and image soft-tissue structures in vivo.  However, it has also been used to assess mechanical properties of tissue via elasticity imaging techniques and algorithms.  Acoustic Radiation Force Impulse (ARFI) imaging can be used to both mechanically excite the tissue and observe the response of the tissue.  Images can be created to highlight the mechanical properties of different tissues, thus providing contrast between structures difficult to discern via traditional b-mode imaging.  More specifically, the project is investigating the feasibility of imaging intercostal nerves during image-guided procedures performed by anesthesiologists.  It is a collaboration with the Departments of Biomedical Engineering and General Anesthesiology at the Cleveland Clinic Foundation.