The use of artificial intelligence in breast cancer screening reduces cancer diagnosis rates by 12% and increases early detection rates in subsequent years, according to a first-of-its-kind trial.
researchers said Study This was the largest trial to date looking at the use of AI in cancer screening. It involved 100,000 women in Sweden who were part of mammography screening and were randomly assigned between April 2021 and December 2022 to either AI-assisted screening or standard readings by two radiologists.
The AI system worked by analyzing mammograms and assigning low-risk cases for a single reading by the radiologist and high-risk cases for a double reading, as well as highlighting suspicious findings to support the radiologist.
According to research published in The Lancet, AI-assisted mammography screening reduced cancer diagnoses by 12% in the years following a breast screening appointment. There were 1.55 cancers per 1,000 women in the AI-assisted group, while there were 1.76 cancers per 1,000 women in the control group.
More than four out of five cancer cases (81%) in the AI-assisted mammography group were detected at the screening stage, compared to only three quarters (74%) in the control group, and there were also about one third (27%) less aggressive subtypes of cancer in the AI group compared to the control.
Dr. Christina Long of Lund University in Sweden and lead author of the study said AI-assisted mammography can help detect cancer at an early stage, but it has some drawbacks.
“Widely introducing AI-assisted mammography into breast cancer screening programs could help reduce workload pressures among radiologists, as well as detect more cancers at an early stage, including cancers with aggressive subtypes,” Long said.
“However, introducing AI into healthcare should be done carefully, using tested AI tools and with ongoing monitoring to ensure we have good data on how AI impacts different regional and national screening programs and how this may vary over time.”
Breast cancer is the leading cause of death among women aged 35 to 50, with more than 2 million people diagnosed with the disease each year globally.
Although studies have shown that AI can bring clear benefits to mammography screening, researchers do not support replacing health care professionals with AI, because screening still requires at least one human radiologist to perform readings with AI support.
Dr Soumya Murthy, senior strategic evidence manager at Cancer Research UK, said the findings were promising but urged caution.
“Using AI to help read mammograms may be more efficient, but there are concerns it may miss some cancers. This study helps address concerns, but the results are from a single center, so more research will definitely be needed to know whether this will help save lives,” Murthy said.
He said: “With the number of people affected by cancer expected to increase in the coming years, innovations like this will be vital to improving the NHS, but it is vital that they are properly evaluated to ensure that they help rather than harm people affected by cancer.”
Simon Vincent, Chief Scientific Officer at Breast Cancer Now, said: “This first trial underlines the huge potential of AI to support radiologists in breast cancer screening. Screening is a vital tool for early detection, and the earlier the disease is detected, the greater the chance of successful treatment.
“This study shows real promise for early diagnosis that can improve and save lives, which is why the trial launched last year in the UK to explore the use of AI within the NHS breast screening system will be important in determining the safest and most effective way of using these tools to detect more cancers early.”
