Across the world, more than 500 million people live with diabetes, and for roughly one in three of them, a foot ulcer becomes a quiet crisis — one that can end in amputation or early death when specialist care arrives too late or not at all. Researchers at the University of Melbourne and RMIT have developed THERMUL, an AI-powered thermal imaging tool that reads the heat signatures of wounds through a smartphone camera to predict, within the first weeks of treatment, whether a foot ulcer will heal or spiral toward serious complication. The tool's promise lies not only in its clinical precision
AI thermal imaging tool THERMUL shows promise in predicting diabetes foot ulcer healing
Diabetes-related foot ulcers affect 500M+ people globally, with 34% developing ulcers that can lead to infection, hospitalization, amputation, and early death without proper management.
Heat patterns in a wound's first image can signal whether it's on track or at risk
Mark
Why does it matter that this tool is portable and smartphone-based? Couldn't clinicians in remote areas just send photos to specialists?
Mimi
They could, but that introduces delay. A patient in a remote town might wait days for a specialist to review an image and respond. THERMUL gives the local clinician—maybe a nurse or GP—an immediate prediction right there. It's the difference between waiting and acting.
Mark
The study involved 110 patients. Is that enough to trust the algorithm?
Mimi
It's a solid foundation, but it's why they're expanding now. They need to test it across different populations, different healthcare settings, different climates even. The algorithm learned from 110 people; it needs to prove it works for thousands.
Mark
You mentioned heat patterns. What exactly is the AI seeing that a human eye can't?
Mimi
A thermal camera shows temperature variations across the wound surface. The AI is trained to spot subtle patterns—maybe a particular distribution of warmth, or lack of it—that correlate with healing. A clinician might see a wound and think it looks okay. The thermal data might reveal inflammation or poor blood flow that isn't visually obvious.
Mark
If this works, what happens to the specialist foot doctors in regional areas?
Mimi
They don't disappear. They become consultants for the complex cases. THERMUL handles the triage—it flags which ulcers need urgent escalation. That frees specialists to focus on the patients who truly need their expertise, and it means more people get appropriate care sooner.
Mark
The cost is $875 million a year. How much of that is preventable?
Mimi
That's the question the funding will help answer. If THERMUL catches even a fraction of ulcers before they become infected or require amputation, the savings compound quickly. But you have to prove it works first, and you have to get it into people's hands.
El Pulso
- Diabetes-related foot ulcers are silently catastrophic — affecting hundreds of millions globally, they can progress to infection, amputation, and death when early warning signs go unrecognized.
- The current standard of care depends on regular specialist visits that people in regional and remote Australia often cannot access, creating a dangerous gap between need and treatment.
- THERMUL captures thermal images of wounds at the first clinic visit and uses AI to identify heat patterns that signal whether healing is on track or heading toward serious trouble within 12 weeks.
- A clinical study across Austin Health and St Vincent's Hospital Melbourne enrolled 110 patients, generating the training data that taught the algorithm to distinguish healing wounds from those at risk.
- With $540,000 in funding now secured, researchers are expanding validation across metropolitan, regional, and remote settings — and refining the model to work reliably across diverse communities.
- THERMUL is not yet in clinical use, but its trajectory points toward a future where a smartphone and a thermal camera could bring specialist-level wound assessment to the most underserved corners of the country.
Across the world, more than 500 million people live with diabetes, and for roughly one in three of them, a foot ulcer becomes a quiet crisis — one that can end in amputation or early death when specialist care arrives too late or not at all. Researchers at the University of Melbourne and RMIT have developed THERMUL, an AI-powered thermal imaging tool that reads the heat signatures of wounds through a smartphone camera to predict, within the first weeks of treatment, whether a foot ulcer will heal or spiral toward serious complication. The tool's promise lies not only in its clinical precision but in its portability — the possibility that a person in a remote community might receive the same quality of prognostic insight as someone in a major city hospital. With $540,000 in new funding secured, the team is now working to validate THERMUL across diverse populations and settings before it enters clinical practice.
In clinics across Australia, a patient with diabetes presents with a foot ulcer — and what happens next depends heavily on where they live. For those near specialist services, careful monitoring can catch complications early. For those in regional or remote communities, that monitoring is often out of reach. It is this gap that THERMUL, a smartphone-based AI tool developed through a partnership between the University of Melbourne and RMIT, is designed to close.
Diabetes affects more than 500 million people worldwide, and roughly one in three will develop a foot ulcer at some point. Without proper management, these wounds can become infected, require hospitalization, lead to amputation, or contribute to early death. Australia's healthcare system spends an estimated $875 million annually on diabetes-related foot disease — a burden that falls unevenly on a system already stretched thin in its outer reaches.
The research team, working across Austin Health and St Vincent's Hospital Melbourne, enrolled 110 patients in a clinical study. Thermal cameras photographed each wound at the initial visit, then at two and four weeks. At twelve weeks, clinicians assessed whether each ulcer had healed. That data trained an AI algorithm to recognize which early heat patterns corresponded to wounds that healed — and which ones were headed for trouble.
Professor Elif Ekinci, who leads the Australian Centre for Accelerating Diabetes Innovations, explained that certain thermal signatures in a wound's early presentation can signal delayed healing before it becomes visible through conventional assessment. The tool's portability amplifies its significance: a clinician conducting a home visit in a remote community could use THERMUL to access the same quality of prognostic insight as a specialist in a major hospital.
The project began with seed funding from ACADI and has since secured $490,539 through Australia's Economic Accelerator program and $50,000 from the University of Melbourne's Proof of Concept Fund. That investment will support expanded validation across metropolitan, regional, and remote settings, and allow the AI model to be refined using data from more diverse populations — a necessary step before clinical translation. Partner organizations including the Royal Flying Doctor Service Victoria are already part of the network working toward that goal.
THERMUL is not yet in clinical practice, but the direction is clear. A tool that reads heat to predict healing, carried in a clinician's pocket, could meaningfully reshape who receives timely, expert-informed care — and where.
Somewhere in a clinic in Melbourne, a patient with diabetes presents with a foot ulcer. A clinician pulls out a lightweight thermal camera, captures an image, and within moments receives a prediction: will this wound heal on its own, or is it headed for trouble? This is the promise of THERMUL, a smartphone-based tool developed through a partnership between the University of Melbourne and RMIT that uses artificial intelligence to read the heat signatures in wound photographs and forecast healing outcomes.
The problem THERMUL addresses is both vast and intimate. Diabetes affects more than 500 million people worldwide, and roughly one in three of them will develop a foot ulcer at some point in their lives. These are not minor complications. Left unmanaged, they can become infected, land patients in hospitals, lead to amputation of the lower limb, or contribute to early death. Yet the standard clinical approach to monitoring these wounds—checking progress over the first four weeks of treatment—relies on regular access to specialist care. For people living in regional, rural, or remote parts of Australia, that access is often a luxury they don't have.
The research team, working across Austin Health and St Vincent's Hospital Melbourne, enrolled 110 patients with diabetes-related foot ulcers in a clinical study. Using thermal cameras, they photographed each patient's wound at the initial clinic visit, then again at two weeks and four weeks. At the twelve-week mark, clinicians assessed whether each ulcer had actually healed. The researchers then fed this data into an AI algorithm, training it to recognize which heat patterns in those early thermal images corresponded to wounds that went on to heal and which ones stalled or worsened.
Professor Elif Ekinci, who heads the University of Melbourne's Department of Medicine and directs the Australian Centre for Accelerating Diabetes Innovations, explained the insight: certain thermal signatures in a wound's initial presentation can signal whether it's on track or at risk of delayed healing. This matters because it means clinicians could potentially spot trouble early and escalate care before complications set in—before infection deepens, before amputation becomes necessary.
The economic case is substantial. Diabetes-related foot disease costs Australia's healthcare system an estimated $875 million annually. Professor Dinesh Kumar at RMIT noted that THERMUL could reduce that burden while also addressing a deeper inequity: the tool's portability means it could bring specialist-level decision support to people in areas where specialist diabetes foot services are scarce or nonexistent. A person receiving home care in a remote community could have their wound assessed with the same precision as someone in a major hospital.
The project began with seed funding from ACADI, which got the team to prototype stage. Recently, they secured $490,539 through Australia's Economic Accelerator program and $50,000 from the University of Melbourne's Proof of Concept Fund. That money will support expanded validation across metropolitan, regional, and remote healthcare settings, and will allow the researchers to refine the AI model using data from more diverse population groups—a crucial step toward ensuring the tool works reliably across different communities.
THERMUL is not yet in clinical practice. It is still moving through validation and development phases, with a network of partner organizations including the Royal Flying Doctor Service Victoria and Software Medical Devices Pty Ltd working toward eventual translation into real-world care. But the trajectory is clear: a tool that reads heat to predict healing, deployed through a smartphone, could reshape how and where diabetes foot ulcers are managed in Australia and potentially beyond.
Citas Notables
These patterns may provide an early indication of a wound's likely healing trajectory, potentially allowing clinicians to identify people who might benefit from earlier intervention or escalation of care.— Professor Elif Ekinci, University of Melbourne
A portable, non-contact tool like THERMUL has the potential to bring specialist-level decision support closer to where people live and receive care.— Dr Rajna Ogrin, Bolton Clarke
Quieres la nota completa?
Lee el original en Inside State Government ↗
Technology