A Non-radiation Artificial Intelligence Spine Deformity Diagnosis System

Professor Cheung, Jason Pui Yin   (Principal Investigator (PI))

16

2024-02-23

755405

A Non-radiation Artificial Intelligence Spine Deformity Diagnosis System

A Non-radiation, Artificial Intelligence, Spine Deformity Diagnosis System

Orthopaedics/Traumatology

PiH/065/24

Research Talent Hub for ITF Projects (RTH-ITF)

2024

On-going

Spinal deformity is prevalent in adolescents (3%) and elderly (60%). Back pain is a leading global cause of disability and a large percentage is contributed to deformity. Without proper intervention, the disease can progress rapidly, thus close follow-up of the patients is critical. However, this is associated with an enormous economic burden on the public healthcare system. The first line of medical imaging is X-ray for diagnosing and assessment of spine deformities. Clinical decisions are based mainly on the alignment parameters like Cobb angles and lumbar lordosis measured by specialists on X-rays. Repetitive X-rays are required at nearly every routine follow-up to guide timely interventions. Repeated X-rays are associated with high radiation exposure which increases the risk of cancer. Thus, the availability of an easily accessible and non-radiation system that also objectively assesses the deformity of the patients is crucial. In this project, we will develop a portable and low-cost radiation-free AI-powered system for spine deformity diagnosis using depth-sensing technology. Specifically, we will (1) establish a large dataset using depth camera with paired X-ray images labelled by specialists, (2) design and implement different AI-integrated pipelines for spinal deformity severity grading, and (3) clinically validate the developed system.