JERUSALEM, June 22 (Xinhua) -- Researchers have developed an innovative AI-powered method to make cancer Magnetic Resonance Imaging (MRI) scans faster and more accurate, the Israel Institute of Technology said in a statement on Monday.
Although the new method was developed for breast cancer imaging, the researchers said it could also be applied to scans of the brain, head and neck, and potentially other medical imaging technologies in the future.
The new technology, called ELITE and described in Nature Communications, combines artificial intelligence (AI) with advanced mathematical models to improve dynamic MRI scans, which are widely used to detect breast cancer in people at high risk of developing the disease.
MRI scans can provide highly detailed images, but producing those images usually takes time. Traditional dynamic MRI exams generate one image every one to two minutes, making it difficult for doctors to track how contrast material moves through tissue in real time.
The researchers said they addressed this challenge by combining mathematical modelling with a deep-learning system trained to remove image noise and reconstruct missing information. The result is a scan capable of generating one image per second while maintaining high image quality, they said.
In experiments involving 54 patients, the technology improved tumor visibility and diagnostic sensitivity compared with existing methods.
The ability to follow contrast agents almost continuously could help doctors detect small tumors earlier and better distinguish between benign and malignant growths, the researchers concluded. ■
