Computer-Aided Detection of Non-Alcoholic Fatty Liver Disease (NAFLD), Steatosis and NASH, using Raw Signals from Point-of-Care Ultrasound and Deep Learning

Fatty Liver Disease affects approximately 20% of the Canadian population [1]. Furthermore, 4% of Canadians have developed serious inflammation and damage as a result of fat build-up in the liver [2]. This can quickly progress causing liver scarring and cancer [2]. Fatty liver disease requires early diagnosis for effective treatment to be implemented but unfortunately, there are no easily recognizable symptoms. There are high costs and complicated workflows associated with current diagnostic techniques. Alternatively, there is promising evidence which suggests that simple ultrasounds can be used for tissue characterization [3-6]. The focus of this research is to utilize machine learning algorithms and ultrasound data to diagnose liver disease in its early stages. Oncoustics has an ultrasound database that can be used for the proposed research and is capable of leveraging the software into a marketable ultrasound solution. Such technology would allow for early diagnosis though simple, affordable and accessible screening.

Faculty Supervisor:

Eran Ukwatta;Mamatha Bhat

Student:

Miriam Naim Ibrahim

Partner:

Oncoustics

Discipline:

Engineering

Sector:

Other

University:

Program: