Skip to main content
Pathology Update 2025

Using machine learning to uncover the predictive power of the tumour microenvironment.

The Ds Nelson Trainee Oral Prize

The DS Nelson Trainee Oral Prize

5:10 pm

21 February 2025

Plenary 3

DS Nelson & RCPAQAP Trainee Orals

Abstracts/Presentation Description

Jack Hywood1, Jeremy VanderDoes2, Gregory Rice2
1Department of Anatomical Pathology, Royal Children's Hospital, Parkville, Australia
2Department of Statistics and Actuarial Science, University of Waterloo, Canada

The tumour microenvironment, and in particular the dynamics between tumour cells and immune cells, has increasingly been shown to contain prognostic and potentially predictive information relevant to patient care. A common technique for analysing the tumour microenvironment is the use of multiplex immunostaining, in which multiple protein antigens are detected simultaneously within a single area of tissue. This allows for accurate localisation and phenotyping of the imaged cells. There is increasing need to develop statistical techniques for the analysis of this type of data. We present here a novel statistical method for comparing cell interactions in the tumour microenvironment between two different groups of patients (e.g. responder vs. non-responders to a therapy). The technique can also be used to predict which group a new patient belongs to based on the data from their tumour.
 
Author contribution statement:
J Hywood, G Rice: Conceptualisation, methodology.
J VanderDoes: Methodology, software.

Speaker/Presenting Authors

Authors

Submitting/Presenting Authors

Dr Jack Hywood - Royal Children's Hospital, Parkville (VIC, Australia)

Resources