Utilizing high-dimensional omics datasets of breast cancer linked with pathology imaging to inform public health, prevention and treatment

Additional Supervisor: Ass Prof Nick Tobin, Department of Oncology-Pathology, Karolinska Institute

Background

Breast cancer survival is improving in high-resource settings such as the UK and US, but in low income countries in Africa and Asia currently about half of women diagnosed with breast cancer will die within five-years of diagnosis. The implementation of screening mammography has led to controversy about potential over-diagnosis, but also facilitates effective treatment of early stage tumours. To design effective approaches to improve health-care, the development of high-dimensional ’omics population-based datasets are needed to inform prevention, screening and treatment strategies. Like many cancers, breast cancer is a heterogeneous disease arising from different aetiological factors with varied clinical outcomes. The finding of the oestrogen receptor (ER), a key molecular marker for responsiveness to anti-oestrogen therapy; varied with key risk factors (e.g. age, genetic markers and screen detection), provided a major breakthrough in understanding aetiology and targeted treatment.

Oestrogen receptor (ER) expression in tumours is a marker of aetiologic and treatment differences, and preliminary data from over 73,000 cases in Scotland, show ER-positive breast cancers rising and ER-negative breast cancers declining from 1997-2014. The underlying reasons for these divergent trends are lacking as are whether important subgroups of the population differ in these trends (e.g. age, socio-economic status, and screening). Further, recent advances in ’omics technologies have changed breast cancer treatment regimens beyond ER, identifying subgroups of cases that might benefit from different chemotherapy, radiotherapy and hormonal treatment regimens.

Gene expression data is now publicly available for over 10,000 tumors, however the majority of studies are relatively small, representing 40-300 patients. Little effort has been made to-date to consider how representative these studies are of the different populations and to what extent the incidence of tumours with particular molecular subtypes or prognostic signatures varies and whether this is changing over time. This project will bring together expertise in gene expression analysis (Sims) and breast cancer subtyping and prognostic signatures (Sims and Tobin) with digital pathology and epidemiology (Figueroa), leveraging publicly available data against existing high quality cohorts from Edinburgh and the Karolinska.

Aims

Compare how variable and representative the many publicly available primary breast cancer gene expression profiling datasets (microarrays, RNAseq, gene panel arrays) are with cancer registry statistics, relative to local institution specific cohorts from UK and Sweden together representing nearly 15,000 patients. These datasets will be interrogated to assess how the calling of molecular subtypes and prognostic signatures relates to epidemiologic risk factors, treatment response, metastasis and survival, varies across different countries over time. The amount of sample annotation information (such as clinicopathological features, length of follow-up, age, race, BMI) available for each cohort will vary, but we hope to gain an overall picture of general trends and differences between populations over time to inform public health, prevention and treatment.
Quantify at the population level how many women are set to benefit from improved stratification of breast cancer using improved molecular subtyping, with the goal of defining the populations that might most benefit from prevention, screening or treatment interventions in the UK and abroad. This will be achieved by analyzing immunohistochemistry and electronic medical records data for over 80,000 breast cancer cases diagnosed since 1997 in Scotland. Collation of this data has been funded with a recent Wellcome Trust Seed award. Ethics approval from tissue governance are in place and the dataset will be complete by December 2017. We aim to incorporate mammography imaging data from the Picture archiving system of Scotland to this in 2018/2019.
Data from specific aim two will also be used to develop a unique and powerful resource using 1600 samples from two five-year periods (1997-2001and 2007-2011) which have multi-level ‘-omic, pathology data allowing us to evaluate changes in subtypes and outcomes using current and potential molecular and image-based screening, diagnostic, subtyping and prognostic methods.

This MRC programme is joint between the Universities of Edinburgh and Glasgow. You will be registered at the host institution of the primary supervisor detailed in your project selection.

All applications should be made via the University of Edinburgh, irrespective of project location:

http://www.ed.ac.uk/studying/postgraduate/degrees/index.php?r=site/view&id=919

Please note, you must apply to one of the projects and you are encouraged to contact the primary supervisor prior to making your application. Additional information on the application process if available from the link above.

For more information about Precision Medicine visit:

http://www.ed.ac.uk/usher/precision-medicine