Experts Highlight Artificial Intelligence Potential in Breast Cancer
While evaluating the potential of artificial intelligence in breast cancer screenings for risk prediction and identifying hidden cases, it is emphasized that more real-world data is needed before clinical integration.
Artificial intelligence systems used in breast cancer screenings have the potential to predict future risks by analyzing seemingly normal mammograms. While experts state that this technology is promising, they express the need for prospective and real-world data for widespread clinical use.
AI-Powered Risk Prediction in Mammograms
Artificial intelligence systems can not only detect obvious lesions but also analyze seemingly normal mammograms. By examining patterns that radiologists might miss, they can assign future cancer risk scores.
Early Diagnosis and Future Cancer Cases
In a study, artificial intelligence assigned the highest risk score to 23 percent of women who developed interval cancers three screening periods before diagnosis. In the mammogram immediately prior to diagnosis, this rate rose to 39 percent.
Clinical Responses and Potential Side Effects
AI's technical ability to detect high risk in negative mammograms raises new questions for clinicians. While additional imaging can provide early detection, it can increase false positives and unneeded biopsies.
Need for Prospective Evidence
Associate Professor Dr. Hannah Milch from the David Geffen School of Medicine at UCLA emphasizes that finding cancer retrospectively is different from providing early diagnosis in routine screening. Prospective evidence showing that AI changes patient outcomes is needed.
Research Challenges and Lack of Standards
Comparing research on artificial intelligence tools is quite difficult due to the variability in methods and test cohorts. Studies use different screening intervals, technology types, and algorithms.
Future Studies and Monitoring Processes
Future research is expected to prospectively evaluate whether artificial intelligence reduces interval cancer rates. Additionally, long-term patient outcomes and post-market surveillance studies are of great importance.