Artificial intelligence is more accurate than doctors in diagnosing breast cancer from mammograms, a study in the “Nature” journal suggests.
An international scientist team, including researchers from Google Health and Imperial College, London, has designed and trained a computer model to diagnose the cancer in X-ray images from nearly 29,000 women. The algorithm was able to read the mammograms better than six individual radiologists were. The diagnosis from the AI was as accurate as the one from two doctors working together.
In addition, the AI doesn’t get tired like a human being. Experts say that it can improve cancer detection.
How good is the AI at that?
In the research study, an AI model was given anonymized images, so that the women could not be identified. Unlike the human experts who had access to the patient’s history the AI had only the mammograms to read. The results showed that the AI model was as good as the current double-reading system of two doctors, and that it was actually more superior than a doctor at diagnosing cancer. Compared to one radiologist, the AI demonstrated a reduction of 1.2% in false positives, when a mammogram is incorrectly diagnosed as abnormal.
There was also a reduction of 2.7% in false negatives, where a cancer was missed.
Dominic King from Google Health said:
Our team is really proud of those research findings, which suggest that we are on our way to developing a tool that can help clinicians spot breast cancer with greater accuracy.
It takes over a decade of training to become a radiologist capable of interpreting mammograms.
Reading X-rays is vital, but the task can be very time-consuming, and as we know, there is an estimated shortage of more than a thousand radiologists across the UK.
Will AI replace human doctors?
No. It took humans to design and train the artificial intelligence model to read X-rays.
This was a research study, and the AI system has not been used and tested in general clinics. Even if it does get used, at least one radiologist would remain in charge of diagnosis.
The researchers say:
But the AI might largely do away with the need for dual reading of mammograms by two doctors, reducing their workload.
Professor Ara Darzi, report co-author and director of the Cancer Research UK (CRUK) Imperial Centre, told the BBC:
This went far beyond my expectations. It will have a significant impact on improving the quality of reporting, and also free up radiologists to do more important things.
Women aged between 50 and 70 should pass the NHS breast screening every three years.
The use of AI could eventually speed up the diagnosis, as images can be analyzed within seconds by the computer algorithm.
Sara Hiom, director of cancer intelligence and early diagnosis at CRUK, told the BBC:
This is promising early research which suggests that in the future it may be possible to make screening more accurate and efficient, which means less waiting and worrying for patients, as well as more accurate results.
Helen Edwards, from Surrey, was diagnosed with breast cancer at the age of 44 before she was eligible for a free screening.
She required surgery, chemotherapy, and radiotherapy, but has been cancer-free for more than a decade since then.
She was a patient representative on the CRUK panel which had to decide whether to grant Google Health permission to use the anonymized breast cancer data.
Helen told the BBC:
Initially I was a bit concerned about what Google might do with the data, but it is stripped of any identifiers. In the long term this can only benefit women. Artificial intelligence machines don’t get tired… they can work 24/7 whereas a human doctor can’t do that, so to combine the two was a great idea.