Artificial intelligence helps radiologists to interpret scans and X-rays. For example, a brain haemorrhage, tumor or collapse lung can be detected faster. The technology is getting better and better, but that does not mean that we can cancel the radiology course in ten years. “I'm not afraid of my job.”
Until a few years ago, Jan-Jaap Visser, radiologist at Erasmus MC, stared for a long time at a screen looking for 'balls' on a CT scan of the lungs. Those beads can mean cancer. The detective work is therefore a precise job.
Search for globules
Nowadays Visser can be assisted in his assessment by smart software. “Artificial intelligence can detect those balls much faster. That can, among other things, save time,” says the radiologist in conversation with RTL Z.
Visser works at Erasmus MC in Rotterdam and is one of the pioneers of artificial intelligence within the Dutch Society of Radiology.
Magic word in radiology
Artificial intelligence, or artificial intelligence (AI), is now only used for specific tasks in radiology. Such as searching for the globules on a CT scan of the lungs. With other smart software, dementia can be diagnosed or mammograms can be read, x-rays of the breast.
Those options will expand in the future. After all, artificial intelligence is the magic word in radiology, it appeared at the largest radiology trade fair in the American city of Chicago earlier this month. No fewer than 120 startups at the trade show did 'something' with AI.
Tens of millions
Such as the Israeli start-up Zebra, which has developed a technique that can be used to detect a blow-away lung or hemorrhage quickly thanks to AI. Another player is Quantib, originally from Rotterdam. With AI, this startup helps diagnose dementia
The big boys also invest in artificial intelligence. Together with the Leiden Medical Center, Philips has developed an algorithm with which an MRI image can be made much faster, for example from a knee.
“Normally an MRI scan takes 15 to 30 minutes. Thanks to artificial intelligence, this can certainly be reduced by half,” says MRI specialist Thijs van Osch of the LUMC. He contributed to the algorithm.
Acute cases at the top
A competitor has equipped its mobile X-ray machine with an algorithm, Critical Care Suite , which helps radiologists to recognize a collapse and can arrange the photos in order with the most acute cases at the top.
The latter is especially important for hospitals that are very behind in defeating scans, for example in England. “Then as a patient you get a scan and you have to wait for months for the results,” says Visser. Then it helps enormously if artificial intelligence indicates which case should be treated first.
AI as an aid
At Erasmus MC, patients are not being treated better or faster thanks to AI, Visser admits. “But in developing areas there are sometimes no radiologists at all. Then artificial intelligence can already make a huge profit. For example, in detecting turberculosis,” he explains.
Will we no longer need radiologists in the West, because AI has made them superfluous? “It is sometimes said: AI is not going to replace radiologists, but radiologists who are not using AI and I agree,” says Mark van Buchem radiologist and radiology department at the LUMC. The power of human intelligence lies in the ability to improvise and that ability plays an important role in making diagnoses, according to Van Buchem.
AI replaces radiologist
“I'm not afraid of my job,” says Visser. “Radiologists look at the entire medical context. So much more than just the one abnormality.” The moment, for example, a CT scan of the lungs and abdomen is fully assessed by artificial intelligence is still far away, he expects .
More and more tasks are being done by AI. “Now a breast photo is being assessed by two radiologists. You can imagine that one of them is being replaced by smart software. And who knows both in the future, because it is a very specific task.”