2023-12-13 12:25:38
A new study led by the University of Warwick reveals that AI can analyze x-rays and diagnose medical problems as accurately, or more accurately, than doctors.
The AI, which has been trained on 2.8 million historical chest X-rays from more than 1.5 million patients, analyzes the X-rays for 37 possible conditions.
It was as accurate, or even more accurate, than the doctor’s analysis at the time the x-ray was taken for 35 of the 37 conditions (94%).
Artificial intelligence software can analyze x-rays as soon as they are taken to detect possible conditions and report any abnormalities. It then gives a percentage probability of the presence of each of the anomalies. The AI also understands the severity of different conditions and reports the most urgent ones to doctors.
To verify the accuracy of the AI, a sample of more than 1,400 analyzed x-rays was cross-reviewed by a group of experienced radiologists, who compared the diagnoses made by the AI with the historical diagnoses of the radiologists of the time.
The AI is the result of a collaboration between Warwick, King’s College London and several NHS sites, funded by a Wellcome Trust Innovator Award. The program also uses an extensive language model to understand historical reports written by clinicians – the same underlying technology used by other AI programs, such as ChatGPT.
Giovanni Montana, Warwick professor of data science and lead author, suggested that the AI tool might be used as a screening tool for radiologists, or to offer “the ultimate second opinion”, avoiding human bias.
Dr. Montana commented:
This program has been trained on millions of x-rays and is very precise. It eliminates the elements of human error, which are inevitable, and bias. If a patient is sent for an x-ray because of a heart problem, doctors will inevitably focus on the heart rather than the lungs.
[…]
This is completely understandable, but it risks spreading undetected problems to other regions. AI eliminates this human bias – it’s the ultimate second guess. Co-author, Professor Vicky Goh of King’s College London and outgoing chair of the academic committee of the Royal Society of Radiologists, commented on the study:
AI programs currently available within the NHS have only one gate limit. Comprehensive AI programs like this will be the future of medicine, with AI acting as a co-pilot for overwhelmed doctors.
[…]
Given the severe shortage of radiologists in the UK, programs like this will aid interpretation and reduce diagnostic and treatment times. It is also possible that the AI will review the X-rays showing no abnormalities, around half of them, and report this to doctors in order to improve the efficiency of the NHS. By allowing AI to eliminate X-rays with no abnormalities, radiologists will have more time to focus on more difficult and critical exams.
A recent survey by the Royal College of Radiologists found that a shortage of radiologists was leading to longer waiting times and treatment delays at 97% of UK cancer centers.
This AI software – titled X-Raydar – is designed to help reduce doctors’ workload and delays. Remarkably, the research group opened the entire software to non-commercial uses in order to accelerate the pace of research development in this area.
Source : “The X-Raydar Project” (tude de l’Universit de Warwick, Angleterre)
And you ?
What do you think ?
In your opinion, are the conclusions of this study from the University of Warwick credible or relevant?
See as well
DoctorGPT: A great language model that can pass the medical exam in the United States
ChatGPT narrowly passed an important medical exam, a historic feat for AI, but it would still be bad at math
1702473133
#Artificial #Intelligence #Analyze #Xrays #Diagnose #Medical #Problems #Doctors #Study #University #Warwick #England
