Google DeepMind is partnering with England’s National Health Service to determine whether machine learning can streamline radiotherapy planning for head and neck cancers, freeing up clinicians’ time for patient care and research.
The collaboration will see the Alphabet ($GOOG) company analyzing anonymized scans of as many as 700 former head and neck cancer patients at the University College London Hospitals NHS Foundation Trust, according to a statement. The project will evaluate the potential for machine learning in making radiotherapy planning more efficient.
Prior to giving radiotherapy, doctors must create a map of the parts of the body to be treated and parts to be avoided in a process called segmentation. A radiotherapy machine uses this information to direct treatment at tumors while leaving healthy tissue unharmed. Because there are so many vital structures so close to each other, segmentation for head and neck cancers must be “painstakingly detailed” and can take about four hours, according to the statement.
The hope is that machine learning could decrease segmentation time from 4 hours to just 1, Google DeepMind said. In addition to saving clinician hours, the project could also lead to the development of a radiotherapy segmentation algorithm with potential applications beyond head and neck cancers.
In July, DeepMind partnered with another NHS Hospital, Moorfields Eye Hospital, to use artificial intelligence to help in the early detection and treatment of preventable eye diseases. Meanwhile, a number of other players are applying artificial intelligence to a variety of situations. MedyMatch and Capital Health are collaborating on AI for the emergency room, while a Houston Methodist Hospital team is using AI to quickly and accurately analyze patient charts to predict breast cancer risk.