With a PhD in computer science and a focus on biomedical image analysis, Dr Priyanka Rana is bringing a data-driven approach to melanoma research at Macquarie University. Her work, in part funded by the Australian Melanoma Research Foundation, uses AI to identify patterns in melanoma tissue samples that can predict a patient’s response to immunotherapy; an advance that could significantly improve clinical decision-making. As this technology progresses, it has the potential to reshape how melanoma is treated, moving toward more personalised and effective care.
Can you tell us about your background and what led you to work at the intersection of artificial intelligence and melanoma research?
I’m a computer scientist by training, I completed my PhD in Computer Science from UNSW in 2023. My doctoral research focused on developing AI methods to analyse microscopic pathological images for a range of biomedical tasks, including cancer cell classification and protein localisation in subcellular environments, among others.
When I began my postdoctoral fellowship at the Centre for Health Informatics, Australian Institute of Health Innovation, Macquarie University, I was presented with an opportunity to work with pathological images sourced from the Melanoma Institute Australia. The project centred on predicting immunotherapy response in melanoma patients, a pursuit, I find both scientifically compelling and clinically meaningful, and it has remained the focus of my work for the past three years.
“The ultimate vision is a future where no patient undergoes a therapy that was never going to work for them, and treatment becomes truly personalised, sparing patients and their families unnecessary trauma, emotional burden, and financial cost.”
For those unfamiliar, how would you explain your AI-powered imaging research in simple terms?
In simple terms, we use AI to predict whether a patient will respond to immunotherapy before they undergo treatment. When tissue samples are stained with specific proteins, they reveal distinct visual patterns that differ between patients who respond well to immunotherapy and those who don’t. We train an AI model to recognise these patterns from pathological images, so that when presented with a new patient’s tissue sample, it can predict whether immunotherapy is likely to be effective.
This matters because up to ~60% of patients fail to respond to treatment. Early identification of non-responders is therefore critical to redirect them toward more effective alternative therapies, sparing them from debilitating side effects and unnecessary personal and financial burden.
How could this technology change the experience of someone diagnosed with melanoma?
Today, treatment decisions often involve trial and error, a patient may begin immunotherapy, endure weeks or months of debilitating side effects, only to find the treatment isn’t working. That experience can be physically exhausting, emotionally distressing, and financially costly.
This technology offers the ability to analyse a patient’s tissue sample at diagnosis and predict, using AI, whether immunotherapy is likely to benefit them. If identified early as a likely non-responder, their oncologist can immediately explore alternative therapies better suited to their biology, sparing them unnecessary suffering and delays in receiving effective treatment. For likely responders, it provides clinicians with greater confidence in their treatment plan and gives patients reassurance that the therapy they are undergoing is the right one for them.
“The path of research is rarely easy, long hours, setbacks, and moments of self-doubt, but when you recognise that your work could spare a patient and their family from unnecessary suffering, financial burden, and emotional trauma, every difficulty becomes worthwhile.”
Ultimately, this technology has the potential to make the treatment journey more personalised, more efficient, and far less burdensome, moving melanoma care toward precision medicine that puts the individual patient at the centre.
What does it mean to you personally to know your research could benefit someone’s treatment or outcome?
Knowing our research could genuinely benefit someone is the ultimate driving force. The path of research is rarely easy, long hours, setbacks, and moments of self-doubt, but when you recognise that your work could spare a patient and their family from unnecessary suffering, financial burden, and emotional trauma, every difficulty becomes worthwhile. There is no greater fulfilment. It is what gives this work its deepest purpose.
How close are we to seeing AI tools like yours used routinely in hospitals or clinics? Could this approach be applied to other cancers beyond melanoma?
We are making significant progress, but important steps remain before tools like this become routine in clinical settings, large-scale clinical validation, regulatory approval, and integration into existing workflows. The field is moving rapidly, however, and I am genuinely optimistic that AI-assisted pathology will become a meaningful part of clinical decision-making within the next five years.
What we are developing is built around the unique biology of melanoma and the detailed visual information captured by multiplex immunofluorescence imaging. While the approach may offer broader lessons for cancer care, our immediate focus is on delivering a reliable, clinician-ready tool where the need is greatest and our evidence base is strongest.
What has been the most rewarding part of your work so far? What keeps you going on the hard days?
Seeing our first framework published in the IEEE Journal of Biomedical and Health Informatics, a leading peer-reviewed journal, was genuinely rewarding. After rigorous review cycles, seeing that work validated and made openly accessible felt like a considerable milestone. It was the first tangible step toward our broader goal, and confirmation that we are on the right path.
On the hard days, I step back and remind myself of the purpose behind this work, what we are trying to achieve, and how meaningful it could be for patients and their families. That perspective never fails to re-centre me.
Looking ahead, what impact do you hope your research will have in the next 5-10 years?
In the next 5 to 10 years, I hope to see this research move from the lab into real clinical settings, where AI-assisted pathology becomes a routine part of treatment decision-making for melanoma patients. The immediate goal is to have a validated, clinically deployable tool that oncologists can trust and use with confidence.
The ultimate vision is a future where no patient undergoes a therapy that was never going to work for them, and treatment becomes truly personalised, sparing patients and their families unnecessary trauma, emotional burden, and financial cost.
“On the hard days, I step back and remind myself of the purpose behind this work, what we are trying to achieve, and how meaningful it could be for patients and their families. That perspective never fails to re-centre me.”
Read other inspiring stories
