HKUST researchers develop AI-based microscopic imaging system

A research team from the Hong Kong University of Science and Technology (HKUST) has developed an AI-powered microscopic imaging system called Computational High-throughput Autofluorescence Microscopy by Pattern Illumination (CHAMP). CHAMP aims to improve tumor tissue evaluation during cancer surgeries by providing rapid and high-quality visualization of cancer cells, potentially reducing the need for repeat surgeries. It identifies cancer cells in three minutes with over 90% accuracy, matching traditional test turnaround times.

Using ultraviolet (UV) light, CHAMP excites tissue samples to produce a greyscale image, which a deep learning algorithm then converts into a histological view for immediate interpretation. The system does not require tissue processing and is applicable to various organs. The research has led to six US provisional patents, with ongoing studies for other cancers. Supported by the RAISe+ scheme, HKUST plans a large-scale clinical trial across five hospitals. Additionally, Professor Terence Wong founded PhoMedics Limited to commercialize CHAMP for lung and breast cancer applications.