Micro-CT analysis of zebrafish models for liver cancer using deep learning methods

Liver cancer is one of the deadliest forms of cancer globally, with devastatingly low survival rates in Australia. Wasting syndrome, known as cancer-associated cachexia (CAC), is a debilitating disorder affecting liver cancer patients, leading to loss of body weight through the deterioration of muscle and body fat. Despite its impacts, little is understood about wasting syndrome's causes and no therapies exist.

Zebrafish are a recent animal model used to study human cancer. These fish develop liver cancer and an associated wasting syndrome similar to that in humans. A/Prof Andrew Cox's team pioneers using zebrafish to study wasting syndrome using various scientific methods including imaging techniques to characterise its extent.

In 2023 A/Prof Cox collaborated with Dr Jay Black to collect micro-CT data on zebrafish affected by liver cancer-associated wasting syndrome to measure changes in liver, muscle and fat tissue. Micro-CT is an x-ray imaging technique using the same principle as CT scans. It can produce high-resolution 3D imaging of organs, muscle and fat in zebrafish measuring 4-5mm long.

The team aims to develop AI-based models to automatically detect indicators of wasting syndrome in muscle and fat tissue observed in micro-CT images. Leveraging AI will accelerate screening compared to current manual methods. Results of these studies will develop our understanding of CAC. It will allow us to test new genetic regulators to improve our understanding of biology and development of anticancer therapies.

MDAP's expertise will be crucial for developing the deep learning workflows and AI pipeline, ensuring user-friendly deployment for researchers across disciplines. Their computer vision methods will optimise model performance.

Anticipated outcomes include a published deep learning model capable of rapidly quantifying cachexia indicators from micro-CT data. This would be an invaluable tool for researchers using zebrafish as an animal model for studying CAC, enabling more efficient analysis and data collection methods. Findings could ultimately translate to clinical applications enabling earlier detection and monitoring of treatment responses.