BEIJING -- Chinese scientists have designed a new method to predict lymph node metastasis for early lung cancer patients, which could help improve therapeutic plans.

The evaluation of lymph node metastasis mainly depends on radiologists' judgment of computed tomography (CT), which lacks high precision and might lead to inappropriate therapeutic plans which increase the risk of complications.

Researchers from the Suzhou Institute of Biomedical Engineering and Technology under the Chinese Academy of Sciences and Shanghai Changzheng Hospital developed a 3D deep learning system based on CT images and prior clinical knowledge for accurate prediction of lymph node metastasis in early lung cancer patients.

They trained and validated the system with CT images and clinical records of 501 early lung cancer patients. Results showed that the system's diagnostic accuracy rate reached 92.8 percent, higher than that of the radiologists' judgment and previous diagnostic systems.

The research helps enhance the diagnostic precision of lung cancer development and facilitate the application of precision medicine. It also provides a reference for the development of an intelligent image auxiliary diagnosis system.

The research was published in the journal Lung Cancer.