Don't miss our weekly PhD newsletter | Sign up now Don't miss our weekly PhD newsletter | Sign up now

  Personalised surveillance schedules for detecting hepatocellular carcinoma

   Faculty of Health and Life Science

This project is no longer listed on and may not be available.

Click here to search for PhD studentship opportunities
  Dr David Hughes, Dr M García-Fiñana, Prof Philip Johnson  No more applications being accepted  Funded PhD Project (UK Students Only)

About the Project

Early diagnosis of cancer allows early intervention which is expected to improve outcomes for patients. Surveillance is a tool that can aid early detection through regular assessment of patients. However, regular screening places a substantial burden, both on the individual required to attend clinic appointments, and on the NHS, who must pay for both clinic time and any required tests.

Risk prediction models are tools which can aid clinicians in deciding how likely it is a particular patient will develop cancer within a specific timeframe. These tools often utilise biomarkers which often change as a patient’s risk of cancer increases. Joint models for longitudinal biomarkers and time-to-event outcomes, such as time-to-cancer, are state-of-the-art statistical tools which can model the changes over time in biomarker values and the association with risk of cancer.

Joint models offer the potential for a personalised surveillance schedule where the time of a patient’s next surveillance visit is determined based on their individual risk predicted by the model. This provides a way to reduce the overall number of surveillance visits required whilst still identifying cancer cases as soon as possible.

This project aims to develop joint models for longitudinal biomarkers and risk of hepatocellular carcinoma (HCC). The project will investigate whether using multiple biomarkers increases performance and will provide a proof-of-concept of how personalised surveillance schedules could be introduced based on the model. The project will involve development of novel statistical techniques, and their application to large cohorts of patients with liver disease.

This PhD will consist of four work packages that will interlink and build on each other throughout the project.

WP1: Systematic review of existing risk prediction tools for HCC

WP2: Preparation of a Cheshire and Merseyside liver disease cohort dataset

This project will utilise the CIPHA (Combined Intelligence for Population Health Action) data source, the Clinical Practice Research Datalink (CPRD), and Japanese surveillance data for the purposes of validation.

WP3: Developing statistical models for linking longitudinal biomarker values to risk of developing HCC

In this WP, the student will develop a joint model to predict risk of HCC and derive personalised screening intervals from it. These models are computationally intensive so we will develop a Variational Bayes framework to make inference scalable. Several statistical novelties will be required for this including incorporating multiple longitudinal biomarkers and competing risks, and informative observation schedules.

WP4: Validation of the performance of personalised surveillance intervals

The PhD will develop novel statistical methods for scalable Bayesian inference and work collaboratively with liver cancer clinicians to develop tools that are useful in clinical practice.

Interested candidates should email a CV and cover letter to Dr David Hughes ([Email Address Removed]), by 31st August. Please note that if we find a suitable candidate before that date, we will end the search early, so applicants are encouraged to apply as soon as possible.

Geography (17) Mathematics (25) Nursing & Health (27)

Funding Notes

Funding is provided to cover tuition fees and an annual stipend in line with UKRI rates (£18,180 per year) for 3.5 years, and an annual research support budget of approximately £2500. The studentship is funded by North West Cancer Research. This position is for Home/UK students only due to the funding available.


1. Ormerod, J.T. and Wand, M.P., 2010. Explaining variational approximations. The American Statistician, 64(2), pp.140-153.
2. Hughes, D.M., Garcia-Finana, M. and Wand, M.P., 2021. Fast approximate inference for multivariate longitudinal data. Biostatistics.
3. Hughes, D.M., Berhane, S., de Groot, C.E., Toyoda, H., Tada, T., Kumada, T., Satomura, S., Nishida, N., Kudo, M., Kimura, T. and Osaki, Y., 2021. Serum levels of α-fetoprotein increased more than 10 years before detection of hepatocellular carcinoma. Clinical Gastroenterology and Hepatology, 19(1), pp.162-170.

Where will I study?

Search Suggestions
Search suggestions

Based on your current searches we recommend the following search filters.