3-years PhD studentship on Statistical learning-driven outcomes for clinical trials of glaucoma therapies.


Collaborating Institutions:

- UCL Institute of Ophthalmology, London, UK (Prof David Garway-Heath)

- UCL Translational Imaging Group, Centre for Medical Image Computing, London, UK (Dr. Marco Lorenzi, Prof. Sebastien Ourselin)

- School of Health Sciences, City University London (Prof D P Crabb)

- Asclepios Research Group, INRIA Sophia Antipolis, France (Dr. Marco Lorenzi).

The project will be carried in collaboration with the Asclepios Research Group, and part of the project will require travel and stay to INRIA.


Duration of Studentship – 3 years

Stipend – £16,296 per annum + fees (£4,728)

Closing Date – 30th April 2017


Studentship Description:

The general concept to be developed is the combination of diagnostic imaging results with results of visual/cognitive function to improve the power of clinical trials to evaluate the benefit of new treatments. The main disease areas to be studied are glaucoma and cognitive impairment; unique large glaucoma datasets are available and datasets on cognitive impairment are also available.

Glaucoma is one of the most common neurodegenerative diseases in the world and the leading cause of irreversible blindness. It affects more than 70 million people, of whom more than 7 million are blind.  Glaucoma prevalence increases with age; therefore, with an increasingly ageing population, glaucoma is becoming a much more common condition. The extent of damage to the visual field (the area of our vision in which objects can be seen) relates to the reduction of vision-related quality of life in glaucoma patients.   Early detection is important for blindness prevention, and regular monitoring for deterioration in vision is a fundamental aspect of glaucoma management. At present, raised intraocular pressure is the only known modifiable risk factor for glaucoma.

A huge obstacle in the clinical management of glaucoma patients is the variability of measurements from tests used for measuring glaucoma deterioration. The variability is such that it takes many measurements and/or a long period of time to distinguish true deterioration from measurement variability. The most important indicator of glaucoma deterioration is progressive loss. Change in measurements of structural damage to the optic nerve may also be used to track glaucoma deterioration, but structural parameters have not yet been accepted by the regulatory authorities.

Imaging technology is rapidly evolving and new models relating structural measurements and visual function need to be developed, including the prediction of visual fields from images of the optic nerve and retinal nerve fibre layer. To mitigate the impact of future advances in imaging technology, there is a need to provide general model relating degree of correlation between structure and function measurements to the benefit of adding imaging measurements to visual field outcomes to measure deterioration in clinical trial settings.


Person Specification and PhD Objectives:

The project aims at developing an in silico high resolution image-based statistical model of the natural history of Glaucoma, and of the relationship between imaging and visual fields over time. The outcomes will be time varying maps of significant anatomical changes observed during the disease time-span, and the associated clinical measurements. The relative disease staging will represent a specific tracking marker of the pathology, and will provide a quantitative reference for assessing the efficacy of modulators of the disease progression in subgroups with specific trajectories.

The candidate will focus on novel formulations of high-dimensional Bayesian methods, such as Gaussian processes (GP), for spatio-temporal analysis of longitudinal changes in clinical trials data. The model will be estimated from short-term clinical data, and will rely on the identification of appropriate covariance models for relating spatio-temporal and clinical features. Particular focus of the project will be in the modelling and simulation of clinically plausible evolution of biomarkers, represented by clinical and imaging data.


Applicants are expected to have a first degree in computer science, statistics, informatics or relevant physical sciences based subject passed at 2:1 level (UK system or equivalent) or above.


Funding will be for 3 years, with a tax free stipend of £16,296 per year and UK/EU-level university fees (£4,728). Only UK/EU students are eligible to apply for this studentship.

The closing date is 30th April 2017 and the latest anticipated start date is October 2017. However, ealier starting dates can be accommodated.


How to Apply:
If you have any scientific queries please contact Dr. Marco Lorenzi at

Applications (including a covering letter, CV and names of two referees) should be sent to CDT Administrator Miss Rebecca Holmes (Medical Imaging CDT) ( who will also be happy to handle any informal enquiries.


Further information on the research is available at: