Using artificial intelligence (AI) to help read mammograms resulted in 20 percent more cancer detections and cut radiologists’ workload in half, a Swedish study published this week has found.
In what is believed to be the first randomized controlled study comparing AI to human screening, researchers had 40,003 mammograms evaluated with AI assistance and 40,030 that underwent double readings by radiologists without AI support.
“We found that using AI resulted in the detection of 20 percent (41) more cancers compared with standard screening, without affecting false positives. A false positive in screening occurs when a woman is recalled but cleared of suspicion of cancer after workup,” says Kristina Lång, researcher and associate professor in diagnostic radiology at Lund University and consultant at Skåne University Hospital, who led the study, published in The Lancet.
Test Your Knowledge: New FDA Guidance on Mammography and How it Impacts You
Can Routine Mammograms Hold Clues to Cardiovascular Disease?
Not only did the study suggest the safety of using AI screening for breast cancer detection, but it reduced the workload of the radiologists by 44 percent, or the equivalent of about five months of time saved assuming a typical radiologist workload of 50 screens an hour. This is important since there is a global shortage of radiologists, according to the Radiological Society of North America.
“We need to see whether these promising results hold up under other conditions, for example with other radiologists or other AI algorithms. There may be other ways to use AI in mammography screening, but these should preferably also need to be investigated in a prospective setting,” Lång said. “Screening is complex. The balance between benefit and harm must always be taken into account. Just because a screening method finds more cancers does not necessarily mean it’s a better method.
“What’s important is to find a method that can identify clinically significant cancers at an early stage. However, this has to be balanced with the harm of false positives and the overdiagnosis of indolent cancers.”
The Swedish study will continue by tracking the interval-cancer rate for cancers detected between screenings and to identify what cancer types were identified with or without AI.
The contents of this feature are not provided or reviewed by NPWH.