A research team from the Biomedical Imaging Laboratory (BIG), Faculty of Science and Technology (FST), University of Macau (UM) virtually attended the 3^rd East Asia Nuclear Medicine Association Congress, hosted in Osaka, Japan in November. Competing with various institutes of nuclear medicine in Asia, including China Medical University, National Taiwan University Hospital, Taichung Veterans General Hospital, Tokyo Medical and Dental University, Kagawa University Kansai Medical University, and Osaka University, Chen Gefei, a PhD student from BIG lab, won the gold award in the meeting for his work on deep-learning-based segmentations of the liver and tumors from contrast-enhanced computed tomography (CECT) for Yttrium-90 radioembolization dosimetry. This advancement facilitates the treatment planning for hepatocellular carcinoma and liver metastases in Yttrium-90 radioembolization.

Yttrium-90 radioembolization is an effective treatment for inoperable liver tumors, and the treatment planning requires delineation of normal liver and tumors. Deep learning-based methods are feasible for tumor segmentation on single-phase CECT, but there are limitations in differentiating tumors and cysts. In Chen’s project entitled ‘Deep Learning-based Automated Segmentation of Liver and Tumors on CT Arterial Portography and Hepatic Arteriography for Yttrium-90 Radioembolization’, the group evaluated two deep learning-based methods on single-phase and dual-phase CECTs, respectively, for normal liver and tumor segmentation. The deep learning-based method using attention U-Net trained on dual-phase CECTs can achieve superior segmentation accuracy and successfully discriminate cysts and tumors. This method can highly facilitate the Y-90 radioembolization treatment planning procedure.

This project was led by Mok Seng Peng, Greta, a professor in the Department of Electrical and Computer Engineering, Institute of Collaborative Innovation, and the Ministry of Education Frontiers Science Center for Precision Oncology. It was a collaboration between UM and the Department of Nuclear Medicine, the Taipei Veterans General Hospital, and funded by Fundo para o Desenvolvimento das Ciências e da Tecnologia (file number: FDCT 099/2021/A).