Professor H. Peter Soyer is a world leader in preventative dermato-oncology and pioneered the use of dermoscopy in pigmented skin lesions for early skin cancer detection. He is the inaugural Chair in Dermatology and Director of the Dermatology Research Centre at the Frazer Institute within the University of Queensland. Here he leads a multinational team of dedicated researchers including Dr Katie Lee, who has been researching melanoma herself for over a decade. Under NHMRC Ideas Grant funding, Professor Soyer and Dr Lee, among other chief investigators, are tackling melanoma at the pixel level through deep image analysis. For this innovative research, Professor Soyer is the recipient of the NHMRC Marshall and Warren Ideas Grant Award.
Melanoma is Australia’s third most common cancer and can usually be cured with surgery if it is detected early and before metastases set in.
For Dr Lee, her intrigue into skin cancer research began after she had finished her undergraduate degree and when her uncle was diagnosed with melanoma at age 53. The cancer had already metastasised and treatment was unsuccessful.
Current melanoma screening practices focus on moles. However, 70% of melanomas appear on normal skin. Herein lies the danger as normal appearing skin can carry melanoma promoting changes independent of visible risk factors like chronic sun damage.
Dr Lee questioned why skin cancers form at one site, and not on another equally sun damaged patch of skin. Their hypothesis was that the community of skin cells work together to produce a microenvironment that makes it either melanoma prone or melanoma resistant.
Her research has since focused on using ultra sensitive molecular techniques to find markers that distinguish normal skin near melanomas, from skin away from any melanomas or moles. Knowing this information is critical to understanding how and why melanoma forms.
Funding provided through NHMRC Ideas Grants scheme has enabled Professor H. Peter Soyer to assemble a team of research experts to combine their knowledge and skills with other recent advances in melanoma detection. Techniques such as rapid 3D total body photography and machine learning techniques can help distinguish melanomas from moles and measure risk factors like chronic sun damage.
Working with the teams of Dr Zhen Yu (Monash University), Professor Brock Christensen (Dartmouth College, US) and Professor Harald Kittler (Medical University of Vienna, Austria), Professor Soyer and Dr Lee will combine these advances into an artificial intelligence (AI) powered clinician assistant.
This tool will examine 3D total body photographs of patients at the pixel level to pick out skin features that are associated with molecular markers of melanoma prone skin. Doing so will allow the team to assess them alongside more traditional markers of risk such as skin colour, chronic sun damage and the number and distribution of moles.
The research team will also make this tool a module of PanDerm, a dermatology AI model, which is designed to incorporate new discoveries as they arise so that they can be efficiently integrated into clinical practice.
This tool will also allow clinicians to use these advances together to focus on melanoma prone areas of skin, so they can detect and remove melanomas as they arise.
This NHMRC Ideas Grant also allows the team to continue partnering with deeply dedicated study participants, many of whom they have collaborated with for over 10 years. They will donate biopsies of healthy moles and normal skin along with their time, DNA, photos and access to their medical history.