Oracle opens Sydney customer excellence centre to boost AI adoption

Victor Chang Cardiac Research Institute

The Victor Chang Cardiac Research Institute’s AI-powered HeartSight app is designed to support the assessment of aortic stenosis, one of the most common and serious forms of heart valve disease and a key focus of the institute’s research.

While AI has been successfully applied to the task of analysing static medical images (for example, identifying breast cancer from mammograms), echocardiograms (ultrasound cardiac imaging) involve moving imagery. Consequently, the institute’s Valve Disease and AI Laboratory had to develop new techniques, although they still rely on human labelling and interpretation of scans to train the model.

By building the [HeartSight] application on Oracle AI Database 26ai and OCI Data Science, Victor Chang Cardiac Research Institute is taking an exciting first step toward transforming clinical practice from labour-intensive manual analysis to AI-augmented insight

“Research is core to our business,” said associate professor Mayooran Namasivayam, head of the Heart Valve Disease and Artificial Intelligence Laboratory. He explained that HeartSight automates a process that previously required significant human time and expertise.

“We have software that can replicate what I do quicker and more efficiently for less effort,” he observed.

The process examines components of the echocardiogram – for example, looking at the interactions between blood flow through the valve and valve movement to determine how restrictive the valve is – and then classifies it according to severity.

An important point is that HeartSight works with much less data than was previously needed. Rather than requiring between 80 and 120 scans, the application can make a diagnosis from a single scan, which, if necessary, can be performed by a non-expert using handheld equipment. This not only yields efficiency gains, but also makes diagnoses available to more people, especially in rural and remote regions.

While that helps patients, clinicians also benefit from better workflows, and the broader health system becomes more efficient: earlier and quicker diagnosis means more timely treatment, leading to fewer admissions. A further economic benefit is that patients lose fewer working days, meaning national productivity improves.

“AI has been transformative for our field,” Namasivayam said.

HeartSight was built using Oracle’s Apex low-code application development platform and takes advantage of Oracle AI Database 26ai. Data is securely stored in Oracle Cloud Infrastructure (OCI) Storage and processed using OCI Data Science.

“Our partnership with Oracle has been extremely beneficial, foundational and very warm,” he said.

The algorithm was developed in-house, while Oracle helped with the user interface (which was of particular importance to support plans to make HeartSight available internationally) and provided a secure, regulatory-compliant and scalable platform.

“We have been able to put [our technology] in a format that can be distributed to systems that are at the point of scanning, to handheld devices, to reading software, to cloud-based analytics… and then do real-time analytics to transform patient care,” Namasivayam explained.

HeartSight has been successfully piloted and externally validated, which was “really exciting and reassuring for our team,” he added. The next steps will involve trials in non-metropolitan areas and with patients from different ethnic backgrounds to check for scalability.

The scalability provided by OCI means the technique can be taken from the Heart Valve Disease and Artificial Intelligence Laboratory to the world. “Don’t just fix the problem for yourself… fix it for everyone,” he said.

Looking further ahead, it presents an opportunity to deliver better, more efficient and more accessible care.

The laboratory is also working to develop artificial heart valves that will last a lifetime, rather than the current 10 to 15 years. This project applies some of the Oracle technology used by the Oracle Red Bull Racing F1 team, Namasivayam observed.

“By building the [HeartSight] application on Oracle AI Database 26ai and OCI Data Science, Victor Chang Cardiac Research Institute is taking an exciting first step toward transforming clinical practice from labour-intensive manual analysis to AI-augmented insight,” said Bovis.

“We look forward to supporting the Institute through the rigorous validation stages ahead as it works towards delivering more accurate diagnoses and allowing clinicians to focus on complex decisions and delivering high-quality patient care.”

Aspen Medical

Sanja Marais, chief technology and security officer at Aspen Medical, was unable to attend the event in person, but spoke to Computer Weekly via video call.

The Australia-based company provides clinical services in various parts of the world, including managing the Lautoka Hospital in Fiji, providing medical services to the Multinational Security Support Mission in Haiti, and delivering the Australian Government’s Rural Locum Assistance Program.

“For us, Oracle is a full ecosystem,” she said. Aspen Medical uses various Oracle Fusion Cloud applications (such as for human resources management and finance), and is also “deeply invested in the Oracle Cloud Infrastructure and the Oracle Integration Cloud.”

I'm very cautious to not get on the hype cycle and try to AI everything where we already have really solid processes in place with automation

“But a lot of what we do is in and around the Apex low-code app builder and the Autonomous AI Lakehouse for all of our data and consolidation, and the Integration Cloud as middleware,” Marais explained.

Apex allows Aspen Medical to quickly build health applications to support service delivery. In a situation where the company was chosen to send in first responders, “we were able to spin up an app that was live, resilient, secure in the cloud, and portable in less than a week on Apex” rather than going through a procurement process that would have taken months.

Another example is a pharmaceutical stock management system developed with Apex that saves the business thousands of dollars a year. It is used wherever Aspen clinicians are working and is integrated with other parts of the Oracle environment to create sales and purchase orders.

This is particularly significant in the current geopolitical environment where stock management is going to become a huge issue, she noted.

Oracle Autonomous AI Lakehouse has allowed Aspen Medical to centralise its data management, improving operational efficiency and significantly reducing IT overhead and infrastructure costs.

In addition to helping consolidate and standardise data sets, it provides “data equity” for users, allowing them to use natural language to generate a report instead of writing SQL statements. This empowers individuals to ask questions such as, “were there any outliers in the quarterly report?” because all the data from the Fusion applications is available in the lakehouse.

“We’re implementing that now, which means our operational teams will be using that data capability with natural language and AI to do their reporting,” said Marais, suggesting that improving operational staff activities ultimately improves patient outcomes. “Oracle Autonomous AI Lakehouse provides a secure, central platform that sits at the heart of our operations.”

Other areas where AI is in use at Aspen Medical include data interpretation, tender writing, clinical governance, and the company’s health and wellbeing programme.

“We are very conservative when it comes to AI in direct delivery to our patients. We are embedding AI in our operational processing, in our data lakes, and trying to slowly get users used to talking to AI and prompting AI,” Marais said. However, the company is looking to make use of AI in telehealth, subject to the regulations and guidelines that apply to the health sector.

While AI is being employed to enhance user interfaces and, in some cases, to help with integrations in the absence of an appropriate application programming interface (API), Marais does not see it as a cure-all. Robotic process automation (RPA) still has a place at Aspen Medical. While AI is good for extracting fields from unstructured data, such as a medical referral in PDF form, RPA is more suitable for moving that data to another system.

Marais added: “I'm very cautious to not get on the hype cycle and try to AI everything where we already have really solid processes in place with automation.” While AI will find its way into decision support products, she noted that some aspects of such systems need to remain algorithmic.

“RPA has been around for so long, we use it heavily. We use it between our systems, we use it to move things, we use it in lieu of solid APIs. It is a workhorse. It doesn’t think for itself and hallucinate... It just does what it needs to do. So, I think a combination of AI interpreting unstructured data, combined with a solid automation process, is going to serve us better than a full AI solution.”

Other plans include adding more data sets to the lakehouse and extending self-service capabilities across the organisation while controlling access rights appropriately. Marais’s team is also using Apex to develop dashboards that combine clinical and financial data so clients can monitor the delivery of services under their contracts.

Oracle’s Bovis added: “Managing vast amounts of clinical and operational data across complex global healthcare environments is a significant challenge for providers.

“With the high performance, scalability, reliability, and automation of Oracle Autonomous AI Lakehouse, Aspen Medical is turning data into real-time insights that drive operational efficiency and enable smarter decision-making. This data-driven transformation demonstrates how healthcare organisations can take advantage of the latest technologies to improve patient outcomes while building more sustainable, responsive systems.”