AI can outperform human doctors when it comes to identifying ovarian cancer from ultrasound images.

A new study published in the journal Nature Medicine shows that specially trained AI program achieved an accuracy rate of more than 86% in identifying ovarian cancer by scanning ultrasounds, compared to just under 83% for human experts and nearly 78% for non-expert examiners.

“Ovarian tumors are common and are often detected by chance,” researcher Elisabeth Epstein, a senior physician with the Stockholm South General Hospital’s Department of Clinical Science and Education, said in a news release from the Karolinska Institute in Sweden.

These results suggest that AI “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,” Epstein said.

For the study, researchers trained an AI program to be able to tell the difference between benign and malignant ovarian lesions, using more than 17,000 ultrasound images from nearly 3,700 patients across 20 hospitals in eight countries.

The AI program also cut back on the need for expert referrals, by backing up human doctors, researchers found.

In a simulated care scenario, support from the AI cut the number of referrals by 63% and the misdiagnosis rate by 18%.

Overall, these results show that AI could contribute to faster and more cost-effective care for patients with ovarian lesions, researchers concluded.

However, the research team noted that more studies are needed to validate their findings and fully explore the potential helpfulness of AI.

They are now conducting research to evaluate the everyday clinical safety and usefulness of the program, and are planning for a clinical trial to examine the AI’s effect on patient management and health care costs.

“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,” co-lead researcher Filip Christiansen, a doctoral student in Epstein’s research group at Karolinska Institute, said in a news release.