Over 100 radiologists, clinical oncologists, computer scientists, industry experts and funders came together for a pioneering stakeholder summit earlier this week to discuss the challenges of developing artificial intelligence (AI) in imaging and cancer treatment.
The event, “Grand challenges in artificial intelligence in clinical radiology and clinical oncology”, took place at The Wellcome Collection in London on 16 May. The brainchild of Dr Nicola Strickland, President of The Royal College of Radiologists, it was organised in partnership with The Alan Turing Institute, Health Data Research UK (HDR UK) and the Engineering and Physical Sciences Research Council.
The packed programme featured a keynote speech from renowned surgeon Lord Ara Darzi, who called on delegates to urgently collaborate to progress AI in clinical practice, both for the benefit of patients and to maintain the UK as a world leader in innovation after leaving the EU.
HDR UK’s Professor Andrew Morris outlined the potential for national data sharing to realise benefits for patients and improve AI training, stressing the timeliness of the meeting given Government preparations to fund AI initiatives into general diagnostics and cancer detection.
Recurring themes expressed by experts throughout the day, as well as during detailed breakout discussion sessions, revolved around how academics, funders and technology companies could and should collaborate to accelerate AI, as well as the ongoing challenge of sourcing large, robust sets of data to train algorithms and machine learning programmes.
Debate was also heard around the exiting gap in regulation and standards for healthcare AI, and whether professional bodies such as royal colleges should input into product development via standard-setting.
Delegates were introduced to an array of developing projects from universities and industry, including AI algorithms and machine learning that detects lung cancer nodules, models pulmonary hypertension and helps oncologists outline tumours ahead of radiotherapy.