Knee Osteoarthritis (KOA)

is a degenerative disorder with the loss of knee articular cartilage.  

Early prediction of KOA progression is essential to stop disease progression, but society still lacks a reliable prediction method.  Hence, KOA patients only receive non-personalised treatments, which often comes with poor treatment outcome.

Screen Shot 2019-11-08 at 3.35.19 AM.png

10%

Hong Kong population
suffer from KOA

100,000+

New cases in Hong Kong in 2019

1 YEAR

Average waiting time of KOA treatments in Hong Kong public hospitals

The Problem

Current hospital flow

1

2

Problem 1: Lack of reliable KOA progression prediction method

Problem 2: Non-personalised treatments

Solution: An AI-based Progression Risk Assessment System for Knee Osteoarthritis

Suggested hospital flow

2

1

3

Benefit 1: Preliminary screening and triage
Benefit 2: Assist medical personnel in constructing a personalized protocol
Benefit 3: May improve treatment outcome and disease management

Supported by PolyU’s Large Equipment Fund & PolyU’s University Research Facility in Big Data Analytics

polyu.jpg
image.png

Our team is developing an artificial intelligence (AI) -based KOA triage system by predicting the KOA progression risk at different time points in the future. This AI differentiates the KOA progression likelihood by studying the patient’s clinical and imaging data. It also generates an assessment report, highlighting any abnormal physiological conditions of the patient. Therefore, the AI assists healthcare professionals to construct a personalised treatment plan to improve treatment outcome.  

 

Besides, we value point-of-care medical services, as a free mobile application is developing for public hospital patients to record relevant data from home for healthcare practitioners to monitor. With a continuous input of medical data, the KOA progression risk can be assessed progressively to facilitate a dynamic treatment protocol.

Currently, our AI is developed using the Osteoarthritis Initiative in the US with over 20,000 subjects. In the future, we are planning to recruit an additional dataset from the US for better generalization ability to capture any abnormal conditions. Moreover, we are applying the access medical data to over 200,000 Hong Kong patients via the Hospital Authority Data Collaboration Lab to enhance the AI adaptation to the local population.
 

title-90002aBG3.gif
logo_dsp.png

Who are the users?

Screen Shot 2019-11-08 at 3.35.19 AM.png

Public clinical practitioners

Private clinical practitioners

Public hospitals and clinics

Private hospitals and clinics

KOA patients