A recent study conducted by researchers at the Karolinska Institute in Sweden has demonstrated that artificial intelligence (AI) can outperform human experts in detecting ovarian cancer from ultrasound images. The AI models achieved an accuracy rate of 86% in distinguishing between benign and malignant ovarian lesions, surpassing the 82% accuracy of human experts and the 77% accuracy of less experienced clinicians. This advancement holds significant promise for improving early detection and treatment outcomes for ovarian cancer patients.
The study involved training an AI program using over 17,000 ultrasound images from nearly 3,700 patients across 20 hospitals in eight countries. In a retrospective triage simulation, the AI-driven diagnostic support system reduced referrals to experts by 63% while significantly enhancing diagnostic performance compared to current practices. Senior physician Elisabeth Epstein from Stockholm South General Hospital highlighted the potential impact of this technology, especially given the shortage of ultrasound experts in many regions. She noted that AI could complement human expertise, potentially reducing unnecessary interventions and delays in cancer diagnosis. Click for More Details
