Training AI To Spot Lung Cancer

An artificial neural network trained to recognize cancerous nodules could help detect lung cancers earlier and improve patient survival.

AsianScientist (July 7, 2017) – Researchers at Beihang University in China have developed an artificial neural network that can aid lung cancer detection. Their results have been published in Science China Information Sciences.

Lung cancer is one of the most deadly cancers, with a five-year survival rate of 16 percent. Early detection is important as the survival rate rises to 52 percent if lung cancer is diagnosed in the early stages. In clinical practice, computer tomography (CT) can capture fine-grained details for both lung nodules and surround structures, acting as the golden standard for diagnosis. However, the high sensitivity of CT imaging also leads to huge data and complex ambiguities, which makes it hard for radiologists for distinguishing pathological structures from healthy.

In recent years, computer aided detection (CADe) systems have shown great potential in assisting doctors in making diagnoses. However, it is hard to detect lung nodules as they come in a wide range of shapes and are influenced by surrounding vessels and tissues. Furthermore, there are only minor differences between diseased and healthy structures.

In the present study, the researchers used an artificial neural network-based approach to the extraction of lung nodules from chest CTs. Unlike classical methods, this approach focused on the inner structures of nodule voxels and used artificial neural networks to generalize these characteristics.

The authors say that their method can be easily integrated into existing CADe systems and rapidly accommodate and process new data streams with few human interactions. They also proposed a novel voting method based on geometrical and statistical features to better extract initial candidate regions while suppressing ambiguous structures.


The article can be found at: Liu et al. (2017) A CADe System for Nodule Detection in Thoracic CT Images Based on Artificial Neural Network.

———

Source: Science China Press.
Disclaimer: This article does not necessarily reflect the views of AsianScientist or its staff.

Asian Scientist Magazine is an award-winning science and technology magazine that highlights R&D news stories from Asia to a global audience. The magazine is published by Singapore-headquartered Wildtype Media Group.

Related Stories from Asian Scientist