AI-Assisted Breast Cancer Screening: A Breakthrough in Early Detection
Breast cancer is one of the most common cancers affecting women worldwide. Early detection is crucial for successful treatment and improved survival rates. However, traditional mammogram screenings can be challenging for radiologists, who may miss subtle signs of cancer due to factors such as fatigue, overwork, and the visual complexity of the scans. Fortunately, a new study published in The Lancet Digital Health reveals that AI-assisted breast cancer screening can significantly enhance the accuracy of diagnosis.
The study, led by the German startup Vara, tested an AI system that assists radiologists in analyzing mammograms. The AI system was trained on over 367,000 mammograms, including radiologists' notes, original assessments, and information on whether the patient ultimately had cancer. The AI system was tested in two approaches: solo performance and assisting a human expert.
The results were promising. When the AI system worked alongside a radiologist, it was 2.6% better at detecting breast cancer than a doctor working alone. Moreover, the AI system was able to automatically classify 63% of all mammograms as "confident normal," reducing the workload for radiologists. This is a significant breakthrough, as radiologists examining mammograms currently miss one in eight cancers.
The software being tested comes from Vara, a startup based in Germany that also led the study. The company’s AI is already used in over a fourth of Germany’s breast cancer screening centers and was introduced earlier this year to a hospital in Mexico and another in Greece.
The Vara team, with help from radiologists at the Essen University Hospital in Germany and the Memorial Sloan Kettering Cancer Center in New York, tested two approaches. In the first, the AI works alone to analyze mammograms. In the other, the AI automatically distinguishes between scans it thinks look normal and those that raise a concern. It refers the latter to a radiologist, who reviews them before seeing the AI’s assessment. Then the AI issues a warning if it detected cancer when the doctor did not. In the study, the AI examined old scans and compared its assessments with those of the radiologist who reviewed them originally.
"In the proposed AI-driven process nearly three-quarters of the screening studies didn’t need to be reviewed by a radiologist, while improving accuracy overall,” says Charles Langlotz.
The potential benefits of AI-assisted breast cancer screening are immense. By improving the accuracy of diagnosis, AI systems can save lives by detecting cancers that doctors may miss. Additionally, by reducing the workload for radiologists, AI systems can free up healthcare professionals to see more patients and ease the burden in places where there is a dire lack of specialists.
Thilo Töllner, a radiologist who heads a German breast cancer screening center, has used the program for two years. He’s sometimes disagreed when the AI classified scans as confident normal and manually filled out reports to reflect a different conclusion, but he says “normals are almost always normal.” Mostly, “you just have to press enter.”
Curtis Langlotz, director of Stanford's Center for Artificial Intelligence in Medicine and Imaging, is impressed, but he says the next step would be to confirm how well the AI performs over a long period of time in actual clinics with real patients.
So far, attempts to fully replace radiologists with AI have failed. A 2021 review found that in 34 of 36 studies, the AI did worse than a single radiologist at screening for breast cancer from mammograms. All 36 were less accurate than the consensus of two radiologists, which some countries require.
“We often say that AI will not replace radiologists,” Langlotz says. “This study doesn’t change that, but in the proposed AI-driven process nearly three-quarters of the screening studies didn’t need to be reviewed by a radiologist, while improving accuracy overall.” That, he says, is “groundbreaking.”
Langlotz adds that this approach could ease the shortage of radiologists, especially in countries such as Malawi, where there is one radiologist per 8.8 million people, or India, a country of 1.4 billion served by one radiologist for every 100,000 people. Even the US, which proportionally has 10 times as many radiologists as India, is projected to be short 17,000 radiologists by 2033.
Töllner is optimistic that more radiologists using AI will mean earlier breast cancer detection, which could improve survival rates. He also hopes Vara will help quash the high number of false positives—patients recalled for further testing who are actually fine.
The study conducted by Vara is a significant step forward in the field of AI-assisted breast cancer diagnosis and holds great promise for the future of early cancer detection.
Author: Nardeep Singh
Source: MIT Review
