Our specialties
Exams & training

Member benefits

View
| 3 mins

Exploring how artificial intelligence may support human interpretation of medical images

The application of artificial intelligence (AI) to our working and personal lives has been forecast for great expansion in 2016. Taking hold of this emerging development, The Royal College of Radiologists (RCR) recently held a forum that brought together clinicians, academics and researchers from the worlds of computer science, mathematics and medicine. The topic was machine learning – that branch of artificial intelligence which focuses on developing computer programmes that can grow and change when exposed to new data without being explicitly programmed.


Hosted by RCR President, Dr Giles Maskell, the meeting was called by the RCR because its medical specialties of clinical radiology and clinical oncology are highly technical and rely heavily on digital resources. Radiologists interpret medical images to support the diagnosis of patients with disease and injury, and clinical oncologists use these images to guide the treatment of cancer. The specialties are key to the diagnosis of patients with disease and injury and in the treatment of cancer.

Forum participants explored what opportunities machine learning held for the College’s specialties and identified potential barriers to progress including:

  1. how machine learning in medical imaging can be applied to high volume, commonplace imaging techniques;
  2. how the vast amounts of data held within the NHS can be accessed to be applied to machine learning;
  3. how best to address the issues of consent to use data, notably imaging data, to make the best use of resources while assuring patients that their data will be used appropriately and securely.

Participants showed a real willingness to share ideas and to facilitate this, and the RCR has set up an online forum to explore these and other issues further. 

Speaking after the event, Dr Maskell said:

“I was delighted that Fellows of this College came together to join with colleagues from the worlds of computer science and mathematics.  The wish to continue the conversation was clear and that is something we want to enable and support at the RCR.

“It was also apparent that despite suggestions in some quarters that these emerging technologies could take the place of clinical radiologists in the future, there was a clear message from the computer science community that machine learning will support rather than replace human interpretation. As the volume of data acquired in imaging studies continues to increase at an astonishing rate, radiologists will come to rely more and more on computers to support them in what will otherwise be an impossible task.”

Antonio Criminisi of Microsoft Research in Cambridge who led the discussions said:

“Machine learning can provide radiologists and oncologists with automatic tools for the quantitative analysis of disease in medical images. These tools promise to boost the productivity of human experts without removing control from them.”

Criminisi added that

“Embracing this fascinating new technology will translate into both improved patient outcomes and potentially large cost savings for the care provider”.

Ends

Further information: Bruce Sparrow 020 7406 5941 / 07554 998197 [email protected]

Notes for Editors:

  1. In January 2016, the RCR hosted an event for about 40 leading thinkers from the worlds of clinical radiology, clinical oncology, computer science and mathematics. The meeting received a series of presentations about cutting edge research and exploration into the application of machine learning to medical imaging and planning of radiotherapy treatment. This was followed by a lively discussion on the possibilities and potential of machine learning and some of the challenges.
  2. The Royal College of Radiologists has almost 10,000 Fellows and members worldwide, representing the specialties of clinical oncology and clinical radiology. The College sets and maintains the standards for entry to and practise in the specialties in addition to leading and supporting practitioners throughout their careers.