Skip to content

The future of melanoma and skin research is bright. MSCAN is pleased to showcase the work of Australia’s awesome researchers. In this series, the three finalists of the Morgan Mansell Prize provide a summary of their research.

Dr. Deval Mehta a Research Fellow within the Augmented Intelligence and Multimodal Analytics (AIM) Lab at the Faculty of IT, Monash University. His research combines the interdisciplinary domains of computer vision, machine learning, and medical imaging.

Deval’s research was nominated as a finalist for the 2023 Morgan Mansell Prize.

Deval Mehta
Dr. Deval Mehta

Using Artificial Intelligence (AI) alongside skin imaging technology, holds great promise for improving skin cancer care, especially in areas where there’s a shortage of dermatologists.

Current AI tools usually only give one diagnosis for a skin lesion, unlike human dermatologists who consider multiple possibilities. Future AI could offer a range of possible diagnoses to build trust with doctors and patients.

AI models can be biased due to limited data leading to inaccurate diagnoses for people with different skin tones and conditions. A possible solution would be to diversify the datasets to train AI by including a broader range of skin tones to ensure accuracy.

Lastly, the current AI tools mostly need high-quality, specialised images, which can take time and special equipment to capture. This isn’t always practical in everyday medical practice, especially for patients who live in remote/rural areas. Thus, there is a need for AI that can work with simpler, everyday clinical photos (taken from a smart phone camera), making it easier for both doctors and patients to use AI as a first step in diagnosing lesions.

Our innovative solution is the AI-driven Hierarchical Out-of-Distribution Clinical Triage (HOT) model. It is designed to tackle the challenges in skin lesion analysis. Our user-friendly system categorises skin conditions into simple groups like ‘harmless’ or ‘serious’ providing initial insights and offering detailed information as needed. A good analogy is it could be an AI-powered guide that provides a broad overview before delving into specifics.

Another key feature of our developed AI HOT model is its ability to recommend if a patient needs to visit a specialized centre for further diagnosis, just by analysing a photo of the skin lesion taken with a smartphone camera. This is particularly valuable for patients living in remote areas where access to specialised healthcare centres is challenging. It’s like having a preliminary check-up right from your phone, helping you decide if a trip to the doctor is necessary.


About the Morgan Mansell Prize

Morgan Mansell is lovingly remembered as a smart, independent woman treasured by her family and friends and bravely inspired many seriously ill people across the world through her frank and candid discussions about her health and her battle with melanoma at The Alfred Hospital. Morgan was passionate about supporting others who were struggling with health issues and used her beautiful sense of humour to give hope and positivity to those around her.

Morgan recently lost her battle with melanoma at just 25 years of age. This prize is a tribute to her passion for melanoma research and working towards a cure.

Check4Spots Day

an initiative for all of us to ‘be all over melanoma’ in memory of 25 year old Morgan Mansell, who sadly lost her life to melanoma.