June 18, 2024

ANU Researchers Unveil AI Solution to Enhance Brain Tumor Classification with Unprecedented Accuracy: DEPLOY

A groundbreaking AI tool, DEPLOY, has been developed by researchers at The Australian National University (ANU) in collaboration with experts from the National Cancer Institute in the US, to expedite and improve the accuracy of brain tumor classification.

Dr. Danh-Tai Hoang, a researcher involved in the project, emphasized the importance of precise tumor diagnosis and categorization for effective patient treatment. He stated, “The current standard for identifying various types of brain tumors is DNA methylation-based profiling. However, the time it takes to perform this test can be a significant limitation, often taking several weeks or more, and it is not widely available in hospitals globally.”

DNA methylation plays a crucial role in controlling gene activity by determining which genes are active or inactive. However, the time-consuming nature and limited availability of DNA methylation-based profiling have led researchers to explore alternative solutions.

To address these challenges, the ANU researchers and their US collaborators developed DEPLOY, an AI model that predicts DNA methylation and subsequently classifies brain tumors into ten major subtypes. DEPLOY utilizes microscopic images of a patient’s tissue, known as histopathology images, for analysis.

The model was trained and validated on extensive datasets comprising approximately 4,000 patients from the US and Europe. The study, published in the journal Nature Medicine, reported that DEPLOY achieved an unprecedented accuracy of 95%.

Moreover, when tested on a subset of 309 particularly challenging cases, DEPLOY provided diagnoses that were more clinically relevant than the initial diagnoses provided by pathologists. This demonstrates the potential future role of DEPLOY as a complementary tool, enhancing a pathologist’s initial diagnosis or even prompting re-evaluation in cases of disparities.

The researchers are optimistic about the future applications of DEPLOY, as it could potentially be extended to help classify other types of cancer as well.

1. Source: Coherent Market Insights, Public sources, Desk research
2. We have leveraged AI tools to mine information and compile it