Chinese researchers have developed an image-based accurate non-invasive diagnosis method for ovarian cancer, the eighth-deadliest cancer among women worldwide.
Gao Xin with the Suzhou Institute of Biomedical Engineering and Technology affiliated to the Chinese Academy of Sciences has made the development in collaboration with eight hospitals in China.
Using the method, doctors can identify ovarian cancer by using AI-enhanced Magnetic Resonance Imaging (MRI) of patients.
Currently, clinical diagnosis depends greatly on doctors' experience.
Gao said the researchers put in MRI data of 501 ovarian cancer patients and use the AI technology of machine learning to construct analysis models for providing intelligent diagnosis results.
The team has compared the machine's work with six doctors who have been working for two to 13 years. The results show that the average diagnosis accuracy of doctors is 79.5 percent, while that of the machines is 91.7 percent.
"Clinically, the machine reading can assist doctors' work and improve the accuracy and efficiency by giving them more valuable imaging information," said Gao.
The research result has been published in the February issue of the Journal of Magnetic Resonance Imaging. (Xinhua)
52 Sanlihe Rd., Beijing,
Copyright © 2002 - Chinese Academy of Sciences