The integration of artificial intelligence (AI) in breast cancer detection opens a new era of enhanced diagnosis capabilities. Danish and Dutch scientists examined mammograms of over 100,000 women to compare diagnostic accuracy before and after the implementation of AI. The results were published in the journal Radiology.

Global Challenge of Breast Cancer Detection

Breast cancer (BC) is the most common oncological disease among women worldwide. According to the World Health Organization (WHO), 2.3 million new cases were diagnosed in 2022 alone. This disease is also the leading cause of cancer-related deaths among women, with 670,000 succumbing to it in 2022. Early detection through regular mammography-based screening is crucial for improving survival rates. Research shows that annual mammograms for women over 40 reduce the risk of death from BC within the first 10 years of follow-up.

However, these screening programs place a significant burden on radiologists, who must review numerous mammograms, often involving double readings to ensure accuracy. This workload is further complicated by a shortage of specialized breast cancer radiologists in many countries.

How Artificial Intelligence Improves Breast Cancer Screening

A Danish study offers compelling evidence of AI’s potential in enhancing breast cancer screening. Conducted in the Capital Region of Denmark, the research compared two groups of women. The first group (60,751 participants) underwent screening before the implementation of an artificial intelligence system, while the second group (58,246 participants) was screened after AI integration. The average age of participants was 58. The AI system, “Transpara,” was trained to flag suspicious lesions and provide risk assessments, helping radiologists prioritize cases. After the AI system was introduced, the number of unnecessary recalls decreased by 20.5%. This means that fewer women were called back for unnecessary diagnostic tests. At the same time, the cancer detection rate increased from 0.70% to 0.82%, a statistically significant improvement.

Reducing Workload and Increasing Accuracy

Another important finding of the study was the AI system’s ability to reduce the number of false-positive results—cases where women are incorrectly identified as potentially having cancer. The false-positive rate decreased from 2.39% to 1.63%, while the positive predictive value, or the percentage of recalls that led to an actual cancer diagnosis, improved from 22.6% to 33.6%. These findings suggest that AI can help radiologists focus their efforts on patients with the highest risk, potentially improving outcomes for breast cancer patients.

A Global Perspective on AI in Breast Cancer Diagnosis

The success of AI-assisted mammography in Denmark aligns with global efforts to improve breast cancer diagnosis. Studies conducted in other countries have also shown that AI can maintain or even enhance detection rates while reducing the number of unnecessary follow-up screenings. This approach increases screening efficiency and alleviates the psychological burden on patients caused by false positives and additional testing.

AI is also capable of predicting the effectiveness of chemotherapy for breast cancer, already being used in clinical practice.

As AI technologies continue to evolve, their integration into breast cancer diagnosis is likely to become more widespread, providing a valuable tool that complements human expertise. While artificial intelligence cannot replace radiologists, it offers tools that make screening more accurate and efficient.