Advancing detection and treatment of abdominal aortic aneurysm | Deakin Applied Artificial Intelligence Initiative

Challenge

Abdominal aortic aneurysm (AAA) affects over 100,000 Australians and 20 million people globally. With few if any symptoms and limited treatment options, AAA has been dubbed a ‘silent killer’.

Impact

The ultimate goal is to help clinicians predict aneurysm growth and rupture risk, and be able to offer personalised treatment options, including the development of non-surgical treatment options and better informed surgical strategies.

Supported by a $5 million 5-year NHMRC Synergy Grant, the project team, led by Distinguished Professor Jonathan Golledge, Head of the Queensland Research Centre for Peripheral Vascular Disease at James Cook University, is aiming to develop an AI-supported system capable of efficiently detecting and modelling abdominal aortic aneurysms using advanced imaging techniques.

The artificial intelligence (AI) project team, led by Professor Truyen from Deakin Applied Artificial Intelligence Initiative, is aiming to develop an AI-supported system capable of efficiently detecting and modelling abdominal aortic aneurysms using advanced imaging techniques.

Detection of abdominal aortic aneurysms is difficult and even if detected before a rupture, current treatment options are limited to monitoring or surgery, if possible, which may not reduce the critical risk.

Grants:

National Health and Medical Research Council (NHMRC) Synergy Grant

$5 million / 5 years

Solution

The Deakin AI team is working towards generating full diagnostic reports of the lower abdomen, creating AI solutions that are trained with expert radiologist input, and developing expert-informed decision support tools for clinicians.

The solution involves a three-step pipeline:

Detection – The team have already developed a model for detecting AAA using 2D imaging to identify aneurysms and is now investigating how computer vision methods could be used to build precise 3D models of the aneurysms, for improved accuracy in detection and to progress the development of new treatments.

Reporting – The efficient generation of comprehensive diagnostic reports, assisting radiologists to manage very high demand for diagnostic imaging.

Clinical Integration – Enabling treating doctors to quickly assess information about each patient using AI-generated insights, aimed at improved decision-making and patient outcomes.

These are early steps towards the ultimate goal of precision medicine. In an ideal world, that would incorporate as complete an understanding of the individual patient as possible, leading to integrated prevention and treatment options.

Research partners:

James Cook University

Macquarie University

University of Queensland

St Vincent’s Institute

Web links: