Researchers at Newcastle University have developed a new AI tool, DeepMerkel, designed to predict the progression and severity of aggressive skin cancers, such as Merkel cell carcinoma (MCC). This innovation enables healthcare professionals to tailor treatments, ensuring patients receive the most effective and personalized care.
Image Credit: New Africa/Shutterstock.com
In two recent studies, the research team demonstrated how DeepMerkel uses advanced machine learning to analyze personal and tumor-specific factors. The tool generates personalized predictions of treatment outcomes, significantly enhancing clinical decision-making for both patients and doctors.
The international team, led by Newcastle University researchers, created DeepMerkel as a web-based system that combines machine learning with clinical expertise. By analyzing subtle patterns and trends in large datasets, the AI tool offers precision insights that inform treatment decisions. The team believes the system could be adapted to other aggressive skin cancers, further improving precision medicine and patient choice.
Understanding Merkel Cell Carcinoma
MCC is an extremely aggressive but uncommon type of skin cancer. Usually affecting older adults with compromised immune systems who present with advanced disease linked to poor survival, it can be challenging to treat.
DeepMerkel is allowing us to predict the course and severity of a Merkel cell carcinoma enabling us to personalize treatment so that patients are getting the optimal management. Using AI allowed us to understand subtle new patterns and trends in the data which meant on an individual level, we are able to provide more accurate predictions for each patient.
Dr. Tom Andrew, Plastic Surgeon and Study First Author, Newcastle University
“This is important as in the 20 years up to 2020, the number of people diagnosed with this cancer has doubled and while it is still rare it is an aggressive skin cancer which is increasingly affecting older people,” said Dr. Tom Andrew, a CRUK funded Ph.D Student.
Dr. Aidan Rose, a Senior Clinical Lecturer at Newcastle University and a Consultant Plastic Surgeon at Newcastle Hospitals NHS Foundation Trust, and Penny Lovat, a Professor of Dermatology and Oncology at Newcastle University, collaborated on the study.
Being able to accurately predict patient outcomes is critical when guiding clinical decision making. This is particularly important when treating aggressive forms of skin cancer in a complex patient group which typically results in difficult, and sometime life-changing, choices being made regarding treatment options. The developments we have made using AI allow us to provide personalized survival predictions and inform a patient’s medical team of the optimal treatment.
Dr. Aidan Rose, Senior Clinical Lecturer, Newcastle University
The team explains how they created the web-based prognostic tool for MCC using sophisticated statistical and machine-learning techniques.
Method
In their study published in Nature Digital Medicine, the team explained how they used explainability analysis to uncover new insights into mortality risk factors associated with the aggressive skin cancer, MCC. By combining deep learning feature selection with a modified XGBoost framework, they developed DeepMerkel, a web-based prognostic tool specifically designed for MCC.
Further analysis, detailed in the Journal of the American Academy of Dermatology, involved data from nearly 11,000 patients across two countries. The results showed that DeepMerkel could accurately identify high-risk patients at earlier stages of the disease. This capability empowers medical professionals to make better-informed decisions about when to implement radical treatment strategies and more intensive monitoring protocols, ultimately improving patient outcomes.
Patients First
The team hope that DeepMerkel will provide better information for patients to make decisions with their medical teams about the best treatment for them as an individual.
Dr. Andrew added: “With further investment, the exciting next step for our team is to further develop DeepMerkel so that the system can present options to help advise clinicians on the best treatment pathway open to them.”
The next step is to incorporate the DeepMerkel website into routine clinical practice and expand its application to other tumor types.
Journal References:
Andrew, W. T., et al. (2024) A hybrid machine learning approach for the personalized prognostication of aggressive skin cancers. Nature Digital Medicine.doi.org/10.1038/s41746-024-01329-9
Andrew, W. T., et al. (2024) A multivariable disease-specific model enhances prognostication beyond current Merkel cell carcinoma staging: An international cohort study of 10,958 patients. American Academy of Dermatology. doi.org/10.1016/j.jaad.2024.10.096