This project advances breast cancer diagnostics through the development of robust artificial intelligence methods, with a focus on improving diagnostic accuracy using optimised machine learning approaches. Validated as a proof of concept using high-quality public datasets, the work establishes a strong scientific foundation for AI-assisted cancer detection with broad translational potential.

A defining strength of the project is its exceptional academic performance, which reflects both methodological rigour and international relevance. The research has produced six Web of Science Q1 publications, including four ranked within the top 5% globally, demonstrating that the proposed AI diagnostic framework meets the highest standards of scholarly excellence in medical AI.

Crucially, the project builds trusted and transparent AI foundations that support future clinical integration and wider diagnostic use. These validated methods provide a strong platform for progression towards real-world data analytics using PPUM’s clinical data repository, creating a natural pathway towards clinical workflow adoption and controlled sandbox evaluation. This staged approach ensures that future clinical applications are evidence-driven, reliable, and locally relevant.

Beyond breast cancer, the flexibility of the approach has been demonstrated through successful application to skin cancer diagnostics, highlighting its potential as a generalisable AI framework for enhancing cancer detection accuracy across multiple disease areas.

Human capital development has been a key outcome of the project. Two graduate students have been trained to conduct high-impact AI research with clinical significance, resulting in publications in leading journals. In addition, two intellectual property outputs have been generated and are currently in the copyright filing process, reflecting early steps towards knowledge protection and future translation.