Chinese researchers have developed a computer deep learning neural network for grading one type of liver cancer with high accuracy. The model shows the potential to be an asset to liver cancer treatment.

Hepatocellular carcinoma (HCC) is the most common type of primary liver cancer and often occurs in people with chronic liver diseases, such as cirrhosis caused by hepatitis B or hepatitis C infection. It is the third leading cause of cancer-related deaths worldwide.

Histological grading of HCC is of great significance in clinical diagnoses, treatments and prognoses. However, evaluate HCC grading from radiology images is challenging for medical professionals, so they must rely heavily on their previous experience.

As machine learning has shown great promise in medical image analysis in recent years, researchers have been proposing to build computer neural networks to classify HCC subtypes.

Researchers from Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Sciences, the Second Hospital of Suzhou University and other research institutions developed the neural network for HCC grading by combining two classic neural networks and training the network with enhanced nuclear magnetic resonance images of 75 patients.

The researchers reported in the journal Computers in Biology and Medicine that their proposed model achieved an accuracy of 83 percent in classifying HCC.

The researchers said that the model performed on par with experienced doctors, demonstrating that machine learning can achieve high performance when working on challenging image classification tasks.

In future studies, the researchers plan to integrate the model into liver cancer diagnosis and treatment system, which is expected to help doctors make better surgery plans for liver cancer patients. (Xinhua)