INSPIRE: Imaging-based Nasopharyngeal Slide Prognostics Integrated with RNA-seq Evaluation

Introduction

Nasopharyngeal carcinoma (NPC) is a highly heterogeneous cancer with distinct geographic and ethnic prevalence, particularly in South-East Asia. The INSPIRE project aims to develop advanced diagnostic and prognostic models for NPC by integrating artificial intelligence (AI) analysis of whole slide histopathology images with molecular profiling, including RNA-seq and spatial transcriptomics. By combining imaging biomarkers derived from routine histology and immunohistochemistry with transcriptomic information, INSPIRE seeks to enable more accurate risk stratification and personalised treatment planning for patients with NPC.

The project brings together expertise from the UK and Malaysia in computational pathology, cancer biology, and clinical translation, and is supported by the British Council through the International Science Partnerships Fund (ISPF) to strengthen international research collaboration and capacity in AI-enabled precision oncology.

Method

INSPIRE adopts a multi-modal, imaging-driven methodology combining computational pathology and transcriptomics:

• Whole Slide Imaging (WSI): Collection of paired H&E and immunohistochemistry (IHC) slides (including LMP1 and EBER) from nasopharyngeal carcinoma samples, scanned at high resolution.
• WSI Registration and Alignment: Development and validation of a novel WSI registration algorithm to accurately align serial histology and IHC sections. This method has been integrated into TIAToolbox to allow interactive visualisation and analysis.
• AI-based Biomarker Development: Training deep learning models on aligned H&E and IHC images to discover prognostic imaging biomarkers associated with disease progression and treatment response.
• Integration with Transcriptomics: Planned integration of spatial transcriptomics and RNA-seq data to link imaging phenotypes with underlying molecular pathways (subject to data availability).
• Validation and Bias Mitigation: Model validation across gender and ethnic groups, with stratified analyses to minimise bias and improve generalisability.

Alignment with NHS Long Term Plan and wider government priorities

The INSPIRE project aligns with key NHS Long Term Plan and UK government priorities by:

• Supporting early and more precise cancer diagnosis through digital pathology and AI.
• Advancing data-driven and personalised medicine, a core theme of UK life sciences strategy.
• Contributing to AI and digital health innovation, addressing national priorities in artificial intelligence, health data science, and translational research.
• Strengthening global health research partnerships, particularly with regions disproportionately affected by NPC.

Project partners

This UK–Malaysia partnership combines strengths in AI algorithm development, computational pathology, and access to clinically relevant NPC datasets.