Artificial intelligence improves ovarian cancer diagnosis with lab tests in China

In this study, researchers in China collected data from 98 lab tests and clinical information of women with and without ovarian cancer from three hospitals between January 2012 and April 2021.

The researchers used a special method called multicriteria decision making-based classification fusion (MCF) to create an artificial intelligence model, which combined information from 20 different AI models to make predictions about ovarian cancer.

Results showed that the MCF model, using 52 features including lab tests and age, had high accuracy in predicting ovarian cancer. It performed better than traditional markers like carbohydrate antigen 125 (CA125) and human epididymal protein (HE4), especially in early-stage cases.

Researchers concluded that this model not only accurately detected ovarian cancer but also surpassed traditional markers (CA125 and HE4) in identifying early-stage cases. Going forward researchers believe the MCF model could prove beneficial in identifying cases particularly in primary care settings and routine health check ups where there could be limited experience in gynaecological oncology.