Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Master's Award: Canada Graduate Scholarships

University of Toronto

$17,500.00

Automated diagnosis of fibrosis in liver transplant using deep learning applied to conventional ultrasound imaging

FY 2021-22

Summary
Liver transplantation (LT) is considered the only curative treatment for patients with chronic liver disease. However, patients can develop long-term complications in their graft, particularly due to new or recurrent liver disease. These diseases can cause chronic liver injury, leading to fibrosis, which is associated with poor health outcomes and decreased graft survival. To manage these patients, liver biopsy is used to assess their level of fibrosis, but this procedure is expensive and invasive, leading to complications such as bleeding and infections. Thus, there is a need to develop non-invasive screening tools for graft fibrosis. Although several non-invasive tools exist such as laboratory tests and Transient Elastography (TE) which uses ultrasound waves to quantify liver stiffness, they have some inherent limitations. In particular, TE is expensive and requires extensive training. Fortunately, artificial intelligence (AI) is increasingly recognized as a tool that can make use of cheaper conventional ultrasound imaging. Recent studies show that AI can be used to diagnose and delineate fibrosis in non-transplant populations with great accuracy. We aim to build on these studies to develop a high-performing model that caters to LT patients, as there are limited studies on this population. We will use a deep learning model, a subtype of AI, trained on ultrasound imaging to diagnose fibrosis and subsequently combine extracted imaging features from our model with demographic, clinical, and laboratory data to predict the risk of developing fibrosis in LT recipients, facilitating early detection of fibrosis. The development and implementation of an AI model will enable better management of LT patients to prevent graft-related complications and improve patient outcomes.
Program
Master's Award: Canada Graduate Scholarships
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
CIHR Funding Decisions Database

As published

Id
466511
Name
Rabindranath Madhumitha
Region
Ontario
Country
Canada
Orgtype
Institution Paid; Research Institution
Prcname2
Special Cases - Awards Programs
Projecttitle
Automated diagnosis of fibrosis in liver transplant using deep learning applied to conventional ultrasound imaging
Programtype2
Studentship Programs
Orgnameinp2
University of Toronto
Cihramount2
$17,500
Cihrequipment2
$0
Programname2
Master's Award: Canada Graduate Scholarships
Keyworddelim
Artificial Intelligence; Liver Graft Fibrosis; Liver Transplanation; Non-Invasive Assessment
Abstract
Liver transplantation (LT) is considered the only curative treatment for patients with chronic liver disease. However, patients can develop long-term complications in their graft, particularly due to new or recurrent liver disease. These diseases can cause chronic liver injury, leading to fibrosis, which is associated with poor health outcomes and decreased graft survival. To manage these patients, liver biopsy is used to assess their level of fibrosis, but this procedure is expensive and invasive, leading to complications such as bleeding and infections. Thus, there is a need to develop non-invasive screening tools for graft fibrosis. Although several non-invasive tools exist such as laboratory tests and Transient Elastography (TE) which uses ultrasound waves to quantify liver stiffness, they have some inherent limitations. In particular, TE is expensive and requires extensive training. Fortunately, artificial intelligence (AI) is increasingly recognized as a tool that can make use of cheaper conventional ultrasound imaging. Recent studies show that AI can be used to diagnose and delineate fibrosis in non-transplant populations with great accuracy. We aim to build on these studies to develop a high-performing model that caters to LT patients, as there are limited studies on this population. We will use a deep learning model, a subtype of AI, trained on ultrasound imaging to diagnose fibrosis and subsequently combine extracted imaging features from our model with demographic, clinical, and laboratory data to predict the risk of developing fibrosis in LT recipients, facilitating early detection of fibrosis. The development and implementation of an AI model will enable better management of LT patients to prevent graft-related complications and improve patient outcomes.
Competitiondate
202112
Approvedterm2
1 yr 0 mth
Pinamesdelim
Rabindranath, Madhumitha
Approvedterm
1 an 0 mois
Cihrcontribution2
~ $17,500~ $0~