Study Suggests AI Can Screen For Breast Cancer Similar To A Human Radiologist

Artificial intelligence is revolutionizing radiology, according to some experts.

A new study in The Lancet Oncology journal has found that AI can detect breast cancer as successfully as an experienced radiologist. The study shows that AI will increase the availability of radiologists to perform more advanced diagnostic work. According to The Washington Post, this study is not the first to explore using machines for medical diagnostics. The authors said it is the first randomized controlled trial to examine the use of artificial intelligence in mammography screening.

The Lancet Oncology published results from a trial with 80,000 female participants in Sweden. The journal revealed that AI detected 20% more cases of breast cancer than usual readings performed by radiologists. At least one radiologist verified the results based on risk factors for the patient.

Based on the study, researchers concluded that AI is a “safe” method for mammography screening. Furthermore, it relieves radiologists of pressure and reduces wait times for patients.

Experts advise that AI models for screening are trained and tested before being used in healthcare facilities. Thus, it will take time before AI usage is seen in these settings. The study’s lead author, Kristina Lang, says the study’s results on safety are “promising;” however, they are “not enough on their own to confirm that AI is ready to be implemented in mammography screening.” Other experts such as Nereo Segnan and Antonio Ponti, affiliated with the Center for Epidemiology and Cancer Prevention in Turin, Italy, regard the study as “breathtaking.”

An article published with the study indicates that using AI screening techniques can reduce breast cancer mortality by identifying it in its earliest stages when it is more treatable.

In a post on X, formerly Twitter, Robert O’Connor, director of Ireland’s National Clinical Trials Office, referred to the analysis as “game-changing.”

Game changing #Breast #Cancer screening paper showing the potential of #AI to improve detection and efficiency out in @TheLancetOncol https://t.co/Xctxk1tabr
1/n

— Dr Robert O'Connor (@DrRobertOConnor) August 2, 2023

World Health Organization states breast cancer resulted in 685,000 deaths globally in 2020.

James O’Connor, a professor of radiology at the Institute of Cancer Research in London, said using AI to screen patients will “have the potential to save a lot of time, and this could help with shortages in workflow.” However, he is concerned about patients’ willingness to use AI in their care. According to O’Connor, the notion that AI will replace radiologists is “nonsense” as there is a shortage of experienced radiologists.

Experts are concerned that AI may over-diagnose cancers or lead to over-treating patients because it detects cancers that have not spread or may be low-grade. There is also discussion about whether AI is more or less successful at diagnosing breast cancer in certain age groups and ethnicities.