Mathematical Modelling of Tumour Molecular Characteristics in Triple- Negative and ER+ HER2+ Breast Cancer using Bulk and Spatial Omics to Predict Disease Outcomes
Loading...
Embargo End Date
2027-02-09
ICR Authors
Authors
Zhu, X
Document Type
Thesis or Dissertation
Date
2026-02-09
Date Accepted
Abstract
Breast cancer (BC) is heterogeneous, both based on the statuses of hormone
receptors and human epidermal growth factor receptor 2 (HER2) and molecularly.
Patients with triple-negative breast cancer (TNBC) and estrogen receptorpositive
HER2-positive breast cancer (ER+ HER2+ BC) tend to have worse
outcomes. For TNBC, chemotherapy and the emerging immunotherapy are the
current treatment strategy. For ER+ HER2+ BC, patients are generally treated
with endocrine and anti-HER2 therapies. Despite the use of tailored systemic
therapies, tumours with the same receptor statuses often exhibit differential
responses to treatment. We hypothesised that this variability is driven by
molecular heterogeneity within the tumour and its microenvironment. In this
thesis, I aimed to unveil the molecular characteristics of TNBC and ER+ HER2+
BC using multi-omics data generated from tumour samples in clinical trials, while
leveraging computational methods to predict survival outcomes and treatment responses. In TNBC, I identified four biological subgroups that were associated
with differential prognosis and showed potential in predicting pathological
complete response. Subsequently, I developed a robust biology-informed
machine learning-based TNBC classifier to assign TNBC into these subgroups.
In ER+ HER2+ BC, my analyses revealed that changes in molecular
characteristics after short duration of endocrine therapy provided better
prognostic information than their baseline characteristics. Finally, an integrative
bulk and spatial multi-omics analysis on selected cases in ER+ HER2+ BC
suggested that TP53 mutations and greater intra-tumour heterogeneity were
associated with endocrine therapy resistance. In summary, my PhD study has
demonstrated the potential of integrating multimodal molecular data to uncover
biomarkers for treatment response in TNBC and ER+ HER2+ BC. This thesis
contributes a foundation for advancing personalised treatment for BC patients,
while pointing to exciting opportunities for future exploration.
Citation
2026
DOI
Source Title
Publisher
Institute of Cancer Research (University Of London)
ISSN
eISSN
Collections
Research Team
Genomics in trials