University of Sheffield Featured PhD Programmes
Engineering and Physical Sciences Research Council Featured PhD Programmes
University of Kent Featured PhD Programmes
University College London Featured PhD Programmes
The University of Manchester Featured PhD Programmes

Standardised evaluation of complex prognostic models with application to risk of breast cancer by estrogen-receptor subtype

  • Full or part time
  • Application Deadline
    Thursday, January 02, 2020
  • Competition Funded PhD Project (European/UK Students Only)
    Competition Funded PhD Project (European/UK Students Only)

Project Description

Applications are invited from graduates with a BSc (First or Upper Second) or MSc (Distinction), or equivalent, to work within the Wolfson Institute of Preventive Medicine. This 3 year studentship will commence in Spring 2020 and will be based at the Charterhouse Square Campus. This is an exciting opportunity for a graduate from disciplines related to epidemiology, statistics, and behavioural sciences.

This PhD project will suit a student looking to mix applied and methodological statistics. The aims are (1) develop a new risk model for breast cancer subtype risk assessment and (2) develop methods to evaluate performance of this model, and complex algorithms more generally.

Part 1: applied statistics

Preventive therapy is recommended by UK and other guidelines for women at high risk of breast cancer, but it only prevents oestrogen-receptor positive breast cancer. Models that predict oestrogen-receptor subtype are unavailable. The first aim of this project will be to adapt a state-of-the-art breast cancer risk model (Tyrer-Cuzick model) in order to determine risk by oestrogen-receptor status. Analysis will use data from six existing cohorts of women aged 40-80y from the UK and US. These include data on questionnaire risk factors, genetic testing, breast density, screening history, and have information on more than 200,000 women aid 7,000 with breast cancer. Three external cohorts will be used to validate model performance based on more than 1M women. The student will first undertake a systematic review to refine the Tyrer-Cuzick model risk by oestrogen-receptor group based on the current literature. The model will be tested using data from cohorts, and potentially extended to include interaction terms and risk factors based on imaging (linking to an ongoing CRUK artificial intelligence project in our department). Software to implement the models will be developed and integrated into the Tyrer-Cuzick model algorithm.

Part 2: statistical methodology

In recent years there has been a proliferation of the development and application of complex artificial intelligence, machine learning and statistical models for disease risk and prognosis. Regardless of the approach used to develop the model, it is vital that they are tested in different settings than used for development. Current methods used to evaluate the calibration of absolute risks and their discrimination have the following limitations: (1) it is difficult to transfer the measure of performance between studies with different population structures, such as age; (2) if a model does not work as expected, there are few standard methods to help determine why not, particularly when the underlying model is a black box. The second part of this PhD project will focus on the development of statistical methods and software to address these issues. The model developed in part 1 will be used to demonstrate and motivate the methods. Methodology developed will apply ideas from elsewhere in statistics, including standardisation and regression. Software to enable other people to apply these methods using R will be developed.

Timeline (in months)
1. Data
◦ 0-6: permissions / DTAs will need to be set up for the student to access all datasets planned.
◦ 3-9: data transfer (all available by month 12, latest)
◦ 3-12: organise data for research
2. Applied research: risk model development
◦ 0-12: literature review
◦ 6-18: initial model development
◦ 12-24: data-driven refinement
◦ 24-30: validation analysis
◦ 12-30: software development
3. Methods development
◦ 0-12: literature review
◦ 12-24: development of methods
◦ 12-30: software development
4. Writing up: 30+

Informal enquiries can be made to via email: Dr Adam Brentnall

How to apply
Your application should consist of a CV and contact details of two academic referees. You must also include a personal statement (1,000 words maximum) describing your suitability for the selected project including how your research experience and interests relate to the project.

Please submit your application to: Patrick Mullan ().

Funding Notes

This 3 year PhD studentship is funded by the Wolfson Institute of Preventive Medicine and comes with a tax-free stipend of £21,000. It is open to UK Nationals, EEA/Swiss migrant workers and non-UK nationals with indefinite leave to remain in the UK who will have three years ordinary residence in the EU prior to the start of the studentship. University tuition fees (at UK/EU levels) will be met by the Institute.

Email Now

Insert previous message below for editing? 
You haven’t included a message. Providing a specific message means universities will take your enquiry more seriously and helps them provide the information you need.
Why not add a message here
* required field
Send a copy to me for my own records.

Your enquiry has been emailed successfully

FindAPhD. Copyright 2005-2019
All rights reserved.