Role of Artificial Intelligence in Breast Cancer Screening

According to the World Health Organization (WHO), breast cancer caused an estimated 670,000 deaths globally in 2022 [1]. Furthermore, about one in eight women is expected to develop breast cancer over the course of her lifetime, making it the most common type of cancer in women [2]. In light of this fact, we want to draw attention to technological innovations in breast cancer on International Women’s Day. Specifically, this article covers how the latest advancements in artificial intelligence (AI) are shaping radiological breast cancer screening, and the potential AI offers in improving real-world breast cancer diagnostics beyond clinical research.

How is AI used in breast cancer screening?

Although AI has become a part of everyday vocabulary, its specific meaning and context in breast cancer screening are briefly outlined in this section.

Breast cancer screening involves a multitude of radiological investigations, such as mammography (MG), magnetic resonance imaging (MRI) and ultrasonography (USG). MG is a routine form of investigation that involves sending X-rays (ionizing radiation) across the breast, while USG and MRI are considered to be supplementary examinations [3]. In today’s modern IT systems, medical images (e.g., produced by MG) are stored in digital format (digital MG) rather than the traditional film-based imaging. Due to this digitalization, AI models can access these medical images, perform computational analyses, and extract specific features from them. The basis for the underlying algorithms which enable this are machine learning (ML) and deep learning (DL) – which are subsets of AI. These algorithms “learn” from broader datasets which then help them build predictive models, ultimately enable them to identify abnormalities within individual images.

It should be noted that the use of ML and DL in breast cancer screening is fundamentally different from Generative AI (Gen-AI), such as ChatGPT, which focuses on producing new content (e.g., text or images), often using large-language models (LLMs) [4].

In summary, AI models analyze digital medical images (e.g., produced through MG) and aim to provide information details on image abnormalities, thereby assisting in breast cancer screening. The information details obtained from the analyses from AI varies from model to model. For example, some AI models put out an abnormality score to indicate a cancerous lesion and heat map for localization that is overlayed on the MG image. While other types of AI models can perform breast density assessments and give a short-term risk score. The confirmatory diagnosis of breast cancer involves histopathological examination, and AI also has the potential to assist in analyzing histopathological images [5]. However, the potential of AI for histopathological examination is beyond the scope of this article.

Since the majority of AI-powered medical devices approved by regulatory agencies are in (cancer) radiology, and breast cancer screening in particular has benefited from AI-based devices [6] [7], we will focus on the AI models built for radiological workflows and imaging analyses for breast cancer screening.

Radiological AI tools for breast cancer: research, availability and implications

Several retrospective studies have shown that AI-based systems in MG can perform as well as or better than radiologists and, in some cases, can even detect tumors that appear between scheduled MG screenings that were previously missed by radiologists [8] [9].

Prospective and randomized analyses have replicated these results. For example, the MASAI trial randomly assigned 80,033 women to either AI-supported screening (n = 40,003) or double reading by two radiologists without AI (n= 40,030). The results revealed AI-supported MG screening resulted in a similar cancer detection rate as double reading, and led to reduced screen-reading workload by 44.3% [10]. Another nationwide prospective study, PRAIM, screened 463,094 women (260,739 with AI support) by 119 radiologists. The examinations were assigned to the AI group when at least one of the two radiologists read and submitted the report with the AI-supported viewer. All examinations for which neither radiologist submitted the report using the AI-supported viewer formed the control group. A suspicion of breast cancer after MG was confirmed either through preoperative or surgical biopsy. The results demonstrated that radiologists in the AI-supported screening group achieved higher detection rates than the control group, indicating that AI can improve MG screening results [11].

Figure 1: Potential applications of AI in breast cancer screening using mammography

Figure 1: Potential applications of AI in breast cancer screening using mammography.

Regulatory clearance for using AI-based systems in cancer screening

The use of AI in radiology is not limited to research, and numerous AI-enabled systems are now available for commercial use. As of 2025, over 1000 AI-enabled medical devices are listed on the Food and Drug Administration (FDA) website under the category of radiology [12]. Among them, relevant AI solutions for breast cancer screening include the ProFound AI Breast Health Suite, Lunit INSIGHT MMG and Breast-SlimView. All three solutions use MG images, have regulatory clearance (either CE in Europe or FDA in the USA), and are available for commercial use. ProFound AI can even identify early cancerous lesions and increase the overall cancer detection rate, while Lunit INSIGHT MMG can generate heatmaps with an abnormality score reflecting the probability of the presence of breast cancer. Some evidence suggests Lunit INSIGHT MMG can be particularly useful in detecting abnormalities in dense breasts [13] [14]. This feature is of particular importance as mammograms of dense breast can be challenging for radiologists to interpret and often require subsequent examinations.

On the other hand, Breast-SlimView masks normal physiological areas of the breasts to display only potentially suspicious areas and is designed as a clinical decision support system for radiologists.

Role of radiologists and AI support in breast cancer screening

While interpreting positive results from AI tools and reviewing commercial AI solutions, caution must be taken to not underestimate the crucial role radiologists play in breast cancer screening. Presently, it should not be assumed that AI can replace radiologists.

Instead, AI can complement the work of radiologists, especially during tedious, labor intensive tasks and offer decision-support and triage tools [8]. This is essential as many countries around the world report a shortage of sufficiently trained radiologists [5].

Challenges and final take-away

The implementation of AI models remains a challenge in the real-world, particularly because of accessibility issues, limited clinical adoption or lack of large-scale clinical validation [5]. However, there are significant efforts made to overcome these challenges. For example, a large-scale trial, EDITH, was launched by the National Health Service (NHS) on World Cancer Day, 2025. EDITH, a randomized controlled trial, aims to include nearly 660,000 women across the UK who are routinely invited for breast cancer screening. The current standard of screening requires two NHS experts to look at the obtained images. EDITH attempts to identify the difference in clinical- and cost-effectiveness by replacing the second reader with AI in one of the trial arms. The trial is expected to be completed by 2029 and it can be estimated that the implications of this trial could potentially transform breast cancer screening care. Successful results could significantly reduce the workload for radiologists [15].

There is great potential for AI-based tools to empower clinical workflows. As AI enters clinical practice in breast cancer screening settings, efforts remain in its validation for full clinical adoption.

References

  1. Breast cancer n.d. https://www.who.int/news-room/fact-sheets/detail/breast-cancer (accessed February 6, 2026).
  2. Krebs – Breast cancer n.d. https://www.krebsdaten.de/Krebs/EN/Content/Cancer_sites/Breast_cancer/breast_cancer_node.html (accessed February 9, 2026).
  3. Zhang J, Wu J, Zhou XS, et al. Recent advancements in artificial intelligence for breast cancer: Image augmentation, segmentation, diagnosis, and prognosis approaches. Semin Cancer Biol 2023;96:11–25. https://doi.org/10.1016/j.semcancer.2023.09.001.
  4. Chua BN, Thng DKH, Toh TB, et al. Artificial intelligence for breast cancer management. Commun Med 2026;6:79. https://doi.org/10.1038/s43856-025-01342-3.
  5. Ahn JS, Shin S, Yang S-A, et al. Artificial Intelligence in Breast Cancer Diagnosis and Personalized Medicine. J Breast Cancer 2023;26:405–35. https://doi.org/10.4048/jbc.2023.26.e45.
  6. Luchini C, Pea A, Scarpa A. Artificial intelligence in oncology: current applications and future perspectives. Br J Cancer 2022;126:4–9. https://doi.org/10.1038/s41416-021-01633-1.
  7. Kuno M, Osumi H, Udagawa S, et al. Artificial Intelligence in Clinical Oncology: From Productivity Enhancement to Creative Discovery. Curr Oncol 2025;32. https://doi.org/10.3390/curroncol32110588.
  8. Chua BN, Thng DKH, Toh TB, et al. Artificial intelligence for breast cancer management. Commun Med 2026;6:79. https://doi.org/10.1038/s43856-025-01342-3.
  9. Nanaa M, Gupta VO, Hickman SE, et al. Accuracy of an Artificial Intelligence System for Interval Breast Cancer Detection at Screening Mammography. Radiology 2024;312:e232303. https://doi.org/10.1148/radiol.232303.
  10. Lång K, Josefsson V, Larsson A-M, et al. Artificial intelligence-supported screen reading versus standard double reading in the Mammography Screening with Artificial Intelligence trial (MASAI): a clinical safety analysis of a randomised, controlled, non-inferiority, single-blinded, screening accuracy study. Lancet Oncol 2023;24:936–44. https://doi.org/10.1016/S1470-2045(23)00298-X.
  11. Eisemann N, Bunk S, Mukama T, et al. Nationwide real-world implementation of AI for cancer detection in population-based mammography screening. Nat Med 2025;31:917–24. https://doi.org/10.1038/s41591-024-03408-6.
  12. Health C for D and R. Artificial Intelligence-Enabled Medical Devices. FDA 2025.
  13. Mansour S, Soliman S, Kansakar A, et al. Strengths and challenges of the artificial intelligence in the assessment of dense breasts. BJR Open 2022;4:20220018. https://doi.org/10.1259/bjro.20220018.
  14. Kwon M-R, Chang Y, Ham S-Y, et al. Screening mammography performance according to breast density: a comparison between radiologists versus standalone intelligence detection. Breast Cancer Res BCR 2024;26:68. https://doi.org/10.1186/s13058-024-01821-w.
  15. World-leading AI trial to tackle breast cancer launched. GOVUK 2025. https://www.gov.uk/government/news/world-leading-ai-trial-to-tackle-breast-cancer-launched (accessed February 9, 2026).© 2026 Springer-Verlag GmbH, Impressum