In the dimly lit radiology room of Bács-Kiskun County Hospital outside Budapest, Dr. Éva Ambrózay, a radiologist with over 20 years of experience, scrutinized a mammogram on her computer monitor. Although two radiologists had previously deemed the X-ray free of breast cancer signs, Dr. Ambrózay paid close attention to several red-circled areas flagged by AI software as potentially cancerous. “This is something,” she remarked, promptly scheduling a biopsy for the patient.

Hungary, with its robust breast cancer screening program, has emerged as a leading testing ground for this technology. Since 2021, AI systems have been deployed in five hospitals and clinics performing over 35,000 screenings annually, assisting radiologists in identifying potential cancers they might have missed. Similar efforts are underway in the United States, Britain, and the European Union, where hospitals and clinics are testing or contributing data to develop these AI systems.

As AI continues to dominate Silicon Valley, with innovations like ChatGPT demonstrating remarkable human-like communication abilities, its application in breast cancer screening exemplifies how AI is increasingly integrating into daily life.

Dr. Éva Ambrózay relies on AI software to assist in detecting signs of cancer that doctors might overlook. However, the widespread adoption of this technology still faces several hurdles. Additional clinical trials are necessary before AI systems can be widely used as automated second or third readers of breast cancer scans. The technology must also prove its accuracy across women of various ages, ethnicities, and body types, and demonstrate its ability to identify complex breast cancer forms while reducing false positives.

Despite concerns about AI replacing human radiologists, many experts believe that AI will be most effective and trusted when used alongside trained doctors. AI’s potential to save lives is immense, as noted by Dr. László Tabár, a leading mammography educator in Europe, who was impressed by AI’s performance in screening trials.

Hundreds of images a day in 2016, Geoff Hinton, a leading AI researcher, predicted that AI would surpass radiologists’ skills within five years. His work in developing image recognition systems demonstrated the technology’s potential, but the reality proved more complex. Peter Kecskemethy, co-founder of Kheiron Medical Technologies, recognized the challenges of applying AI to detect illnesses like breast cancer. Growing up in Hungary, where his mother was a radiologist, he witnessed the difficulty of identifying small malignancies in images.

Kecskemethy’s company, Kheiron, has developed AI tools to assist radiologists by training their systems on over five million historical mammograms from clinics in Hungary, Argentina, and academic institutions like Emory University. The AI learns to spot cancerous growths by analyzing shape, density, location, and other factors. Last year, Kheiron’s AI matched human radiologists’ performance in a test involving more than 275,000 breast cancer cases, reducing radiologists’ workload by 30% and increasing cancer detection rates by 13%.

Dr. Tabár, an expert in mammography, tested the AI with challenging cases and was impressed by its accuracy. Kheiron’s technology has been adopted in various clinics, including MaMMa Klinika in Budapest, where it has identified several missed cancer cases since 2021.

Kheiron emphasizes that AI should complement doctors, not replace them. The technology is being integrated into screening programs in Scotland, England, Finland, and Oman. “An AI-plus-doctor should replace doctor alone, but an AI should not replace the doctor,” Kecskemethy stated.

The National Cancer Institute estimates that about 20% of breast cancers are missed during screening mammograms. Dr. Constance Lehman, a professor of radiology at Harvard Medical School, urges doctors to keep an open mind about AI’s role. At Bács-Kiskun County Hospital, Dr. Ambrózay was initially skeptical but soon recognized AI’s value when it identified a tiny tumor she had difficulty seeing. The AI detected something “that seemed to appear out of nowhere,” she noted.

Conclusion

AI is proving to be a game-changer in breast cancer detection, offering a powerful tool to assist radiologists and improve diagnostic accuracy. As the technology continues to evolve and integrate into healthcare systems worldwide, its potential to save lives and enhance public health becomes increasingly evident. Embracing AI in medical practice will be crucial for harnessing its full benefits while ensuring the highest standards of patient care.

‍