Optimising Diffusion Models for Histopathology Image Synthesis


Victoria Porter (The Queen's University Belfast), Richard Gault (The Queen's University Belfast), Stephanie G Craig (The Queen's University Belfast), Jacqueline James (The Queen's University Belfast)
The 35th British Machine Vision Conference
PDFPosterVideo (Right click to download)

Abstract

Oropharyngeal Squamous Cell Carcinoma (OPSCC) is a sub-type of head and neck cancer linked to human papillomavirus infection (HPV). HPV-positive OPSCC patients have an improved prognosis compared with HPV-negative OPSCC patients however, the reasoning for this is unknown. Visualising the clinical and molecular differences in HPV status would be highly interpretable and could aid our understanding of the impact these distinguishing features have on patient prognosis. A generative model trained to delineate features of HPV status provides both a synthetic visualisation of HPV-related OPSCC and a classification of HPV status. Conditional diffusion models (CDMs) have been shown to produce state-of-the-art quality and fidelity in the conditional image synthesis domain. Furthermore, they can generate representative Haematoxylin and Eosin (H&E) stained histopathology images of cancerous tissue. Despite this success, the performance of these models on histopathology images remains sub-optimal when compared to state-ofthe-art CDMs trained on natural images. This paper proposes novel weighting schemes that prioritise spatial features during training enabling important pathological markers associated with HPV-related OPSCC tissue to be learnt. Through experimental analysis of histological data, we demonstrate that our proposed approach improves the performance of existing CDMs and provides insightful, interpretable features that aid our understanding of HPV-related OPSCC.

Citation

@inproceedings{Porter_2024_BMVC,
author    = {Victoria Porter and Richard Gault and Stephanie G Craig and Jacqueline James},
title     = {Optimising Diffusion Models for Histopathology Image Synthesis},
booktitle = {35th British Machine Vision Conference 2024, {BMVC} 2024, Glasgow, UK, November 25-28, 2024},
publisher = {BMVA},
year      = {2024},
url       = {https://papers.bmvc2024.org/0727.pdf}
}


Copyright © 2024 The British Machine Vision Association and Society for Pattern Recognition
The British Machine Vision Conference is organised by The British Machine Vision Association and Society for Pattern Recognition. The Association is a Company limited by guarantee, No.2543446, and a non-profit-making body, registered in England and Wales as Charity No.1002307 (Registered Office: Dept. of Computer Science, Durham University, South Road, Durham, DH1 3LE, UK).

Imprint | Data Protection