Summary: A study published in Nature Medicine shows that Swedish researchers developed AI models that outperformed human experts in detecting ovarian cancer from ultrasound images, offering improved diagnostic accuracy and reduced need for expert referrals.

Key Takeaways

  • AI Outperforms Experts in Accuracy: AI models demonstrated higher accuracy (86.3%) than both expert (82.6%) and non-expert (77.7%) examiners in detecting ovarian cancer from ultrasound images, suggesting AI’s potential to complement human expertise.
  • Reduced Referrals and Misdiagnoses: AI tools reduced the need for expert referrals by 63% and decreased the misdiagnosis rate by 18%, offering faster and more cost-effective diagnostic solutions.
  • Future Research and Development Needed: While results are promising, further studies, including prospective clinical trials, are required to evaluate the safety, adaptability, and real-world impact of AI models in different clinical settings.

——————————————————————————————————————————————————————————

A study published in Nature Medicine by Swedish researchers shows artificial intelligence (AI) models surpass human experts in detecting ovarian cancer from ultrasound images.

“Ovarian tumors are common and are often detected by chance,” says Elisabeth Epstein, MD, PhD, senior consultant in obstetrics and gynecology at Karolinska Institutet and Södersjukhuset (Stockholm South General Hospital. “There is a serious shortage of ultrasound experts in many parts of the world, which has raised concerns of unnecessary interventions and delayed cancer diagnoses. We, therefore, wanted to find out if AI can complement human experts.”

AI Surpasses Human Experts in Accuracy

The researchers have developed and validated neural network models able to differentiate between benign and malignant ovarian lesions, having trained and tested the AI on over 17,000 ultrasound images from 3,652 patients across 20 hospitals in eight countries. They then compared the models’ diagnostic capacity with a large group of experts and less experienced ultrasound examiners.

The results showed that the AI models outperformed both expert and non-expert examiners at identifying ovarian cancer, achieving an accuracy rate of 86.3%, compared to 82.6% and 77.7% for the expert and non-expert examiners respectively.

“This suggests that neural network models can offer valuable support in the diagnosis of ovarian cancer, especially in difficult-to-diagnose cases and in settings where there’s a shortage of ultrasound experts,” says Epstein.

Reducing the Need for Expert Referrals

The AI models can also reduce the need for expert referrals. In a simulated triage situation, the AI support cut the number of referrals by 63% and the misdiagnosis rate by 18%. This can lead to faster and more cost-effective care for patients with ovarian lesions.

Despite the promising results, the researchers stress that further studies are needed before the full potential of the neural network models and their clinical limitations are fully understood.

“With continued research and development, AI-based tools can be an integral part of tomorrow’s healthcare, relieving experts and optimizing hospital resources, but we need to make sure that they can be adapted to different clinical environments and patient groups,” says Filip Christiansen, doctoral student in Epstein’s research group at Karolinska Institutet and joint first author with Emir Konuk at the KTH Royal Institute of Technology in Sweden.

Evaluating the Safety of the AI Support

The Swedish researchers are now conducting prospective clinical studies to evaluate the everyday clinical safety and usefulness of the AI tool. Future research will also include a randomized multicenter study to examine its effect on patient management and healthcare costs.