Fight Tumour Team Wins CDI and Vector Institute ML Challenge for Cancer Image Segmentation
The Fight Tumour Team has been announced as the winner of the inaugural Cancer Digital Intelligence (CDI) and Vector Institute Machine Learning Challenge for Cancer Image Segmentation. This collaborative competition aimed to address computational limitations in medical image object detection and contouring. Participants, comprising researchers from the University Health Network (UHN) and the Vector Institute, developed machine learning models using deidentified 3D radiological images from the RADCURE dataset, which focuses on head and neck cancer patients. The goal was to automate the segmentation of tumors for radiation treatment planning, a task that traditionally requires hours of manual work by radiologists. The winning team, consisting of Jun Ma, Rex Ma, and Ronald Xie, achieved the highest score for model reliability. The McIntosh Lab Team secured second place, distinguished by their model's low complexity and fast inference time. As part of their prize, the winners will present their findings at the Toronto Machine Learning Summit. This initiative highlights the potential of AI to reduce clinical workload and improve patient care, aligning with broader strategies for responsible health data sharing and innovation in healthcare delivery.
Wire timeline
Fight Tumour Team Wins CDI and Vector Institute ML Challenge for Cancer Image Segmentation
The Fight Tumour Team has been announced as the winner of the inaugural Cancer Digital Intelligence (CDI) and Vector Institute Machine Learning Challenge for Cancer Image Segmentation. This collaborative competition aimed to address computational limitations in medical image object detection and contouring. Participants, comprising researchers from the University Health Network (UHN) and the Vector Institute, developed machine learning models using deidentified 3D radiological images from the RADCURE dataset, which focuses on head and neck cancer patients. The goal was to automate the segmentation of tumors for radiation treatment planning, a task that traditionally requires hours of manual work by radiologists. The winning team, consisting of Jun Ma, Rex Ma, and Ronald Xie, achieved the highest score for model reliability. The McIntosh Lab Team secured second place, distinguished by their model's low complexity and fast inference time. As part of their prize, the winners will present their findings at the Toronto Machine Learning Summit. This initiative highlights the potential of AI to reduce clinical workload and improve patient care, aligning with broader strategies for responsible health data sharing and innovation in healthcare delivery.
Vector Institute for Artificial Intelligence