Developing an Automated Bio-Imaging Platform: The Automated Ultrasound Scoring of Joint Inflammatory and Damage Pathologies in Rheumatoid Arthritis (AUSCORA) Study


Project Outline

Ultrasound (US) allows precise disease characterisation in rheumatoid arthritis. Presently, the most widely used synotivitis scoring system is semi-quantitative and involves a manual two-step process of image acquisition and interpretation, which can be subject to intra- and inter-rater variability among operators. The team will develop a fully automated ultrasound system to detect and quantify the thickened joint capsule (synovial hypertrophy) and the degree of active inflammation, seen as Power Doppler synovium vascularity. This will make the assessment process seamless, as well as eliminate the problem of rater reliability and such a technology will serve as an index for follow up examinations and therapy management.

Project Deliverables

A fully automated system to receive ultrasound joint images and provide an objective quantitative inflammation score and report through i) automatic detection and segmentation of the synovium (measuring its size), and ii) automatic detection and segmentation of the Power Doppler signal within the synovium area, and utilizing that information to quantify joint inflammation.


Principal Investigator: Dr TAN York Kiat

Institution: Singapore General Hospital

NHIC Ref: NHIC-I2D-1506048