An Artificial Intelligence System For The Detection And Characterization Of Early Liver Cancer
Grant Data
Project Title
An Artificial Intelligence System For The Detection And Characterization Of Early Liver Cancer
Principal Investigator
Professor Seto, Wai Kay Walter
(Project Coordinator (PC))
Co-Investigator(s)
Dr Chiu Wan Hang Keith
(Co-Investigator)
Emeritus Professor Li Wai Keung
(Co-Investigator)
Professor Yu Philip Leung Ho
(Co-Investigator)
Professor Yuen Richard Man Fung
(Co-Investigator)
Duration
24
Start Date
2019-05-01
Completion Date
2021-04-30
Amount
4712540
Conference Title
An Artificial Intelligence System For The Detection And Characterization Of Early Liver Cancer
Keywords
An Artificial Intelligence System, Detection And Characterization, Early Liver Cancer
Discipline
Others - Computing Science and Information Technology
HKU Project Code
ITS/122/18FP
Grant Type
Innovation and Technology Support Programme (Tier 2)
Funding Year
2018
Status
Completed
Objectives
Liver cancer, the third most common cause of cancer deaths in Hong Kong, is currently a disease of unmet needs with poor prognosis. Radiological diagnosis of early liver cancer is highly challenging, and often leads to delay in treatment. Our objective is to improve detection of liver cancer by developing an artificial intelligence algorithm that can automatically detect liver cancers on computed tomography (CT) scans. The algorithm, as an integrated computer-aided detection software, can be used to improve diagnostic accuracy of early liver cancer. There is currently no similar product in the market. Artificial intelligence algorithm will be developed by University of Hong Kong researchers with a track record in deep learning, based on anonymised images retrieved from 6 different institutes, and placed on a dedicated secured cloud server via a virtual private network. Processed scans will be fed back to the Radiology Information System as a fully end-to-end automatic output interface, with priority for reporting upgraded if liver cancer is suspected. The software has a high potential for commercialization and patent application. Applying in large healthcare systems e.g. Hospital Authority may lead to significant efficiency saving, reduction in diagnostic errors, improving patient care, and ultimately reducing liver cancer deaths.