Semantron 23 Summer 2023

21 st -century technology, specifically AI, in modern medicine

Sam Morrissey

The advancement of modern technology has touched all our lives and transformed the way we live and work, and therefore it is no surprise that it has had a major impact on the field of medicine. The use of artificial intelligence within medicine has greatly expanded over the past decade, with the NHS

launching a long-term plan to be a world leader in the use of AI screening, and medical or surgical assistance to deliver better and more efficient care. 1 Trials have already begun, yielding promising results. One example is an AI system that has been trained using deep learning, a system where artificial neural networks designed to mimic the human brain are developed via algorithms. In this scenario, an AI developed to detect eye diseases will be showed positive and negative slides hundreds or thousands of times, with the algorithm learning to detect the discrepancies between an infection or clean test. It can then be expanded in complexity by introducing different

types of diseases, where the AI will be able to determine whether an infection is present, then whether the infection is EG bacterial or viral conjunctivitis. A system trialled at Moorfields Eye Hospital in London found that it made the correct diagnosis and referrals for 50 eye diseases with an astonishing 94% accuracy. 2 The NHS’s aim is to render over 30 million outpatient referrals unnecessary saving the NHS £1 billion which can be redirected to other areas of care. Another area the NHS is exploring is radiological screenings using AI assistance. The NHS is performing more diagnostic tests that ever before, with a 20% hike in MRI screenings between March 2016 and March 2019. Each scan needs to be looked at by 2 clinical professionals, which creates an immense burden within radiology departments. AI can relieve this by being able to detect a positive or

negative result with a high degree of certainty, acting as a first or second opinion and thereby freeing up staff, as negative tests do not need to be looked at. An example of how this is being used is in mammography, with an investigation determining that radiologist workload could be reduced by up to 62.5% and with a marginally higher accuracy of 98.6% than an attending (US

1 See learning-within-5-years/#:~:text=As%20set%20out%20in%20the,part%20of%20the%20NHS%20routine. 2 See Chadwick, J.


Made with FlippingBook - Online catalogs