Nasopharyngeal carcinoma (NPC) is a major cancer burden in Malaysia, a recognised endemic region globally. While early-stage NPC is often treatable, tumour recurrence remains a leading cause of mortality, and clinicians currently lack reliable tools to identify which patients are at highest risk after initial treatment.

 

This project directly addresses this gap by developing an AI-based system to predict tumour recurrence using diagnostic pathological slides that are already part of routine clinical care. By applying machine learning to histopathological images, the project moves beyond visual assessment and enables data-driven risk stratification that is not possible through conventional pathology alone.

 

The core innovation lies in a multi-model AI pipeline designed to identify image-based features associated with tumour recurrence. A key technical advance is the development of a stain augmentation algorithm that enhances the quality and consistency of pathological slides, including those stored over long periods. This allows existing pathology archives to be effectively repurposed for AI analysis and strengthens the reliability of model outputs.

 

The project is built on strong interdisciplinary collaboration, combining clinical and biological sciences expertise with advanced machine learning support from computer science collaborators at Sunway University. This integration ensures that algorithm development remains clinically grounded while meeting rigorous computational standards.

 

Looking ahead, the project is focused on clinical translation. The next phase involves external validation with regional and international cohorts, followed by integration into clinical workflows as a decision-support tool for identifying high-risk patients who may benefit from closer monitoring or early intervention.

 

In parallel, the project delivers high-impact capacity building, engaging biological sciences researchers in solving a real clinical problem with direct implications for patient survival. Overall, this initiative demonstrates how clinically anchored, collaborative AI can tackle a national cancer challenge with the potential to change patient outcomes and save lives.