RESEARCHERS at the Hong Kong University of Science and Technology (HKUST) have announced a significant leap forward in medical imaging, developing an artificial intelligence (AI)-powered technology that promises to drastically reduce patient radiation exposure during diagnostic procedures.
This innovation, spearheaded by Professor Li Xiaomeng of HKUST’s Department of Electronic and Computer Engineering and Associate Director of the Center for Medical Imaging and Analysis, could transform how anatomical imaging is conducted globally, offering a safer, faster, and more cost-effective alternative to conventional methods.
The system uses a novel AI algorithm capable of reconstructing highly detailed three-dimensional (3D) images of bones and other anatomical structures from a minimal number of two-dimensional (2D) X-ray radiographs. Unlike traditional computed tomography (CT) scans, which typically require 400 to 500 X-ray views to construct a 3D model, the HKUST system can achieve comparable results with as few as two to four standard X-ray snapshots.
The reduction in data input translates to a 99 percent decrease in patient radiation exposure compared to a typical CT scan.
The AI algorithm leverages advanced deep learning to infer complex volumetric data from sparse planar projections. This intelligent reconstruction process bypasses the need for extensive rotational X-ray exposure, a hallmark of CT technology. Professor Li’s team has rigorously validated the accuracy of their AI-generated 3D models, demonstrating a remarkable 97 percent similarity to those produced by conventional CT scans. Clinicians who have reviewed comparative images have often found it difficult to discern visual differences, underscoring the fidelity of the AI’s output.
The technology has a particular impact on vulnerable populations such as children, pregnant women, and elderly patients who often require repeat imaging and are therefore more susceptible to the cumulative risks of radiation exposure. Aside from enhanced patient safety, the system offers substantial operational and economic benefits to health care systems, since the imaging process can generate high-quality 3D anatomical images in under a minute, and cheaper too — an estimated six to eight times less — than a CT scan.
“Previously, designing a single implant took days of CT scans and manual modeling, not to mention the lengthy wait times for CT scan services at public hospitals. Now, we can generate 3D bone models using just two X-ray inputs in under 30 seconds. This is transformative for patients needing urgent interventions and allows us to deliver personalized solutions at scale, cutting costs by 50 percent to 70 percent,” Edmond Yau, Founder and CEO of Koln 3D, a HK-based company specializing in orthopedic and metal component 3D printing, said.
Be the first to comment