1. Skip to Content
  2. Skip to Navigation
Toggle Menu


More news

A.I. Cancer Diagnosis

15 Jan 2020

The Clinical Negligence Team at Hay & Kilner LLP are frequently instructed to represent patients who have suffered a delay in diagnosis, or a misdiagnosis of cancer. Not only can a delayed diagnosis of cancer have a huge effect on the treatment options available to the patient and their prospects of having a good outcome, these scenarios are also incredibly difficult for a patient to come to terms with on an emotional level.

Early and correct diagnosis of cancer can therefore be extremely important in giving a patient the best chance of a good outcome.

Currently, the NHS uses two radiologists to analyse a patient’s x-ray images/mammograms. In cases where the two radiologists disagree, the opinion of a third radiologist is sought.

However, a recent study published in the scientific journal ‘Nature’ suggests that Artificial Intelligence (AI) computer models can accurately diagnose breast cancer from mammogram images. In the research study, the AI model was provided anonymised mammogram images to examine and, unlike human radiologists performing the same task, the AI model was not provided any background, social or medical history for the patient.

Despite the discrepancy in information provided, the AI model was as good at reading the mammograms as the current ‘double-reading system’ of two radiologists. AI was, in fact, found to be more superior at spotting breast cancer than a single doctor examining the image alone.

Compared to one radiologist, AI saw a reduction in 1.2% of false positives, i.e. when a mammogram is incorrectly diagnosed as abnormal. There was also a 2.7% reduction in false negatives, i.e. where a cancer is missed entirely.


It is currently estimated that the UK is facing a shortage of more than 1,000 radiologists across the country and reading x-rays is described as ‘time-consuming work’. AI technology, on the other hand, could review images in a matter of seconds. Radiologists are also, by their nature, only human. They require rest, breaks and sleep to be able to perform their jobs. AI is tireless and could technically review and interpret images 24 hours a day. It is therefore thought, and possibly hoped, that AI technology could not only reduce the number of misdiagnoses made but also significantly speed up the process of diagnosis.

Whilst AI technology undoubtedly still has a long way to go before it is introduced to clinics and centres across the country, the prospect of developing this technology that could potentially make cancer screening more accurate and efficient, resulting in quicker diagnoses and overall better outcomes, would certainly be an encouraging step forward in medical intervention.

For more information on any of the above, or how we can help you or your family, please contact a member of our Clinical Negligence team, or call 0191 232 8345.