AI for accuracy

It’s early days but artificial intelligence (AI) applications are dramatically improving the personalisation of medical treatment, improving the accuracy of diagnostics and robotic surgery, and driving administrative efficiencies.

South Australia is positioned to become a leader in the application of AI in health – but more needs to be done to provide the regulatory environment and to foster an innovation ecosystem, says Professor Seyedali Mirjalili, director of the Centre for Artificial Intelligence Research and Optimisation at Torrens University Australia.

‘AI can be very helpful, especially when you are dealing with high volume of data that humans are not capable of understanding and analysing. It is being used a lot in imaging – CT, Xray and MRIs – it can look into the history of millions of other images and learn how to recognise complex patterns such as tumors and provide quantitative assessments of radiographic characteristics,’ explains Professor Mirjalili.

‘The other main use I am excited about is the use of AI in robotic and telesurgery. The AI algorithm helps the surgeon perform the procedure, and the robotic system can be much less error-prone than humans, especially in minimally invasive surgery.

‘AI systems can efficiently evolve over time and transfer the learning to other AI systems accurately and quickly – unlike humans, who can take a lot longer to learn new techniques and teach it to others.’

Professor Mirjalili says that in an age of increasing medical complexity and cost, AI has the potential for administrative efficiencies, providing 24/7 advice for patients through the use of smart chatbot technology such as ChatGPT (Chat Generative Pre-trained Transformer), and for scheduling appointments.

Australian researchers have used artificial intelligence to develop a world-first benchmark for measuring brain atrophy in neurodegenerative diseases, including Alzheimer’s disease.

Assessing the onset and progression of Alzheimer’s using MRI has traditionally been challenging as changes in the thickness of the brain's cortex are often in the sub-millimetre range.

Advanced machine learning techniques are routinely used in brain research to assess changes in cortical thickness, but until now, a lack of a clinically accurate ‘ground truth’ dataset meant we could not evaluate their sensitivity to the detection of small atrophy levels.

Scientists from CSIRO and the Queensland University of Technology have used machine learning to produce a set of artificial MRI images of brains with predefined signs of neurodegeneration in the cortex region, the outer layer of the brain most affected by Alzheimer’s.

Prior to this breakthrough, published in Medical Image Analysis, the only way to get a ground truth measure of cortical thickness was by studying the brain post-mortem.

This new technique allows researchers to set the amount and location of brain degeneration they want to compare against so they can get a clear picture of what method of cortical thickness quantification performs the best.

The technique can test the sensitivity of methods to a miniscule level. It can determine whether a method can detect changes in thickness of just 0.01 millimetres.

The technique can be applied to research in any brain disease that involves neurodegeneration, representing a significant step forward to better understanding dementia and other debilitating brain diseases. It can also potentially be used to predict the level of cortical degeneration a person can expect over time.

The synthetic dataset images have been made publicly available so clinicians and scientists can use the synthetic images to conduct their own assessments of cortical thickness quantification methods.

AI is already helping automate doctors’ workflow and improve efficiencies through converting text to voice and automated data mining from different sources, he says.

One of his research projects at the private Torrens University, based in Adelaide, involves building an AI system to facilitate the early detection of mental health conditions in children, so they may be referred to psychiatrists.

The research team is building a system that uses large data sets to identify the probability of a child having, for example, an 80% chance of attention-deficit/hyperactivity disorder (ADHD), based on the parents’ responses to a standardised questionnaire.

The next step will be to build a database of text and video to help with another layer of diagnostics, with the final objective of commercialising the system.

Professor Mirjalali predicts more use of AI in precision medicine, providing treatments based on a person’s unique genes and proteins, by using AI to analyse large data sets.

“Typically, treatments are based on cohorts and groups,’ he says. ‘But we are unique as humans and diseases develop uniquely in our bodies, and there is always a unique solution. AI can hyperpersonalise medical treatment.’

AI is also being used to analyse large amounts of data to identify drug targets and optimise the end design of clinical trials.

South Australia, with its growing international reputation as a site for clinical trials, is well positioned to be a leader in AI applications for health, says Professor Mirjalili. But, he adds, there is more to be done to foster collaboration between technology companies, researchers, regulators and healthcare providers. He says the state government has shown a commitment to invest in AI and digital health, with initiatives such as the Digital Health Cooperative Research Centre and the creation of a dedicated AI hub. Major entities and organisations such as the Australian Institute for Machine Learning, MIT big data living lab, Google cloud, and Amazon AWS are part of the conversation.

‘Adelaide is attractive because of the city’s high-tech capabilities and world leading data ecosystems, supported by universities that are making significant progress in this space, leading in international research,’ he says.

‘We need to have all the players working together to accelerate the development of AI and its adoption.

‘With recent concerns about cyber security, we need better policies in place to ensure that data is used ethically and responsibly – and that includes regulation around data sharing, data storage, and data access.

‘We also need guidelines and policies to ensure AI is used ethically, adhering to patient autonomy and non-discrimination.

The fuel of AI systems is data, and if data is biased towards a gender or race, the outcome of the AI will be biased too.’

Ensuring that the application of AI is transparent – rather than a black box with data in and recommendations out – is an important piece of the adoption puzzle, says Professor Mirjalili.

Standardisation and interoperability of systems to enable the integration of new and existing health systems, and a culture of innovation, will also be vital.

‘Medical professionals should not be afraid that they will lose their jobs,’ he says. ‘AI will make their jobs easier, so we will see gradually machines do the leg work and humans do the tasks that require special skills such as strategic decision making, persuasion and caring.’

Hazelwood Clinic is a well established GP owned and operated practice.

We are seeking a VR GP to join our busy Eastern Suburbs practice – 2-4 sessions per week to full-time.

We are a private billing practice with well equipped consulting rooms and a friendly and supportive medical, nursing and admin team and value a respectful team environment.

We provide a wide variety of medical services including onsite Spirometry, 24 Hour blood pressure monitoring, Holter monitor and ECG

Enquiries to Jenny Lambert, 0419 409059 or email jennylambert@hazclin.com.au