PhD Studentship: Optical coherence elastography: using touch and artificial intelligence to detect cancer

Optical coherence elastography: using touch and artificial intelligence to detect cancer

Primary Supervisor: Professor Peter Munro

Secondary Supervisor: Dr Sami Sahmed, Professor Simon Arridge

A four-year fully funded PhD studentship, co-funded by the Royal Society, is available in UCL’s Computational Optics Group within the Department of Medical Physics. The successful candidate will be part of the UCL i4health CDT and will benefit from the activities and events organised by the centre. The project will involve close interaction with researchers from UCL Hospital.

Background

Oesophageal cancer is one of four cancers of “unmet need” (Cancer Research UK) mainly due to its poor survival rates and late diagnosis. Significant deficiencies exist at all stages of surveillance, staging, and treatment. We aim to overcome this problem using optical Coherence Elastography (OCE), which images tissue mechanical properties, in three-dimensions, with very high sensitivity, at a spatial resolution approaching that of single cells. OCE is yet to be applied to oesophageal cancer.

OCE is inspired by the sense of touch that clinicians have used for centuries when diagnosing disease. OCE uses optical coherence tomography (OCT) to measure how tissue deforms when a compressive force is applied. Simplistically, the more tissue deforms, the softer it is. It is because of OCT’s nanometer scale displacement sensitivity that OCE has an unrivalled sensitivity to small variations in stiffness. Despite its success in breast cancer imaging, OCE is still limited by its ability to retrieve tissue stiffness from the raw OCT images, largely due to the speckled nature of OCT images.

Research Aims

The overarching aim is to improve how oesophageal cancer is diagnosed and treated by developing a novel in room, real time, functional OCE-based imaging system based on existing OCE system developed by Prof. Munro. The key innovation in this project will be a combination of hardware improvements and the development of a novel approach to tissue stiffness retrieval based upon the underlying physics of optical coherence tomography combined with deep learning.

Our specific aims are:

• Adapt an existing OCE imaging system (hardware and software) to the specific requirements of imaging oesophageal tissue oesophageal biopsy samples.
• Evaluate and adapt an existing model-based approach to tissue-stiffness retrieval technique for oesophageal tissue.
• Develop a deep learning approach to stiffness retrieval and compare with the model-based approach. We will initially use simulated OCE data in a supervised training approach and then investigate unpaired and unsupervised training schemes as well as generative adversarial methods and transfer learning.
• Investigate combining learned and model-based approaches to stiffness retrieval using a learned iterative scheme.
• Evaluate the ability of the system (ex-vivo in the lab) to differentiate between normal and dysplastic/malignant tissue using conventional histopathology as the reference standard.
• Perform a preliminary examination of the whole range of tissue types (normal, low and high grade dysplasia arising in Barrett’s oesophagus, intramucosal and invasive cancer) in order to understand where OCE performs best.

This successful student will participate in all aspects of the translation of this technique from the lab to the clinic, including hardware, software and pre-clinical imaging. Results permitting, this project may also lead to the creation of a spin-out company to translate the developed technique into routine clinical use.

Further details including how to apply can be found here

Application Deadline: 4^th January 2023.

Advert information