Artificial intelligence (AI) is proving to be a potential life-saving tool for the nearly 20,000 women in the U.S. who are diagnosed with ovarian cancer each year. A recent study led by researchers at Karolinska Institutet in Sweden revealed that AI models outperformed human doctors in detecting ovarian cancer.
The study, published in Nature Medicine, tested an AI model's ability to differentiate between benign and malignant lesions on the ovaries. The AI model was trained on over 17,000 ultrasound images from 3,652 patients across 20 hospitals in eight countries.
The AI model achieved an impressive accuracy rate of 86% in detecting ovarian cancer, surpassing the 82% accuracy rate of human experts and the 77% accuracy rate of less experienced individuals.
According to study author Elisabeth Epstein, a professor at Karolinska Institutet, ovarian tumors are often detected incidentally, highlighting the importance of accurate diagnostics. AI-driven diagnostic support can significantly enhance triage efficiency, reduce diagnostic errors, and address the shortage of expert examiners, ultimately leading to improved patient outcomes.
Dr. Brian Slomovitz, a gynecologic oncology expert, emphasized the importance of early detection in reducing ovarian cancer-related deaths. While AI shows promise in improving cancer diagnostics, limitations such as data bias and regulatory concerns need to be addressed.
Dr. Harvey Castro, an emergency medicine physician, highlighted the need for additional research to assess AI's real-world adaptability, long-term impact on healthcare costs and outcomes, and its effectiveness across diverse populations and clinical settings.
The researchers acknowledged the study's limitations and emphasized that AI should be used as a diagnostic support tool, not a replacement for human physicians. Clinical studies are planned to further evaluate AI's performance in real clinical settings.
Collaborating with the KTH Royal Institute of Technology, the research was funded by various organizations, including the Swedish Research Council and the Swedish Cancer Society.
