Job:CRUK Cambridge Centre MRes + PhD in Cancer Biology (Cross-Programme Quantitative Projects)
Entering edit mode
Last seen 2.6 years ago
United Kingdom

Artificial Intelligence assisted integration of imaging, pathology and genomics data to identify prognostic features and improve prostate cancer diagnosis (Early Detection Programme)

Application Deadline: 1st December 2018
Full details and application link at

Eligibility and Funding:
We invite applications from UK, EU and non-EU (international) students for these non-clinical studentships. Please ensure you meet the University of Cambridge entrance requirements. 


Dr Charlie Massie

Dr Shamith Samarajiwa

Aims and objectives:

Multi-parametric magnetic resonance imaging (mpMRI) is the emerging gold standard for prostate cancer diagnosis. Men with suspected prostate cancer are classified using mpMRI, blood tests (PSA) and clinical features to determine who should receive trans-rectal diagnostic biopsies. However, 1-in-5 patients with highly suspicious lesions on mpMRI have benign disease and there is a similar risk of missing significant cancers. These data highlight limitations in the current diagnostic pathway and the need for more accurate tools to support clinical decision making. The long-term aim of the project is to spare tens of thousands of men each year from unnecessary invasive diagnostic procedures while minimizing the risk of missing significant cancers. Specifically, this project will develop AI-based data science approaches to achieve the following aims:

• Integrating genomic data sets together with rich clinical and pathology data to mechanistically classify tumour samples
• Develop image analysis methods, training on 100s of MRI scans with paired pathology and genomics data
• Apply these methods to validation cohorts in the diagnostic pathway to test the impact of AI approaches on diagnostic accuracy
• Explore the added value of AI approaches to existing patient prognostication models

Experimental plan:

This project will apply deep neural network (DNN) approaches coupled with crossmodal machine learning (ML) methods and computational genomics, using imaging, pathology and genomic data collected from patients in the prostate cancer diagnostic pathway. Convolutional Neural Network analysis of multi-parametric magnetic resonance imaging (mpMRI) data will be used to segment and classify diagnostic/prognostic features and these will be integrated with paired pathology diagnostics (ground truth) and genomics data with the aim of improving diagnostic accuracy and informing the design of new imaging sequences and companion diagnostic tests. This exciting cross-disciplinary project will enable the selected student to interact with an active clinical research team and two laboratories; the Massie lab, which will provide data sets and expertise in prostate cancer biology/bioinformatics and the Samarajiwa lab, contributing expertise in Data Science (including AI and ML approaches), Systems Biology and Computational Genomics.

Main Techniques:

• Deep Neural Networks, Machine Learning (CNNs, Variational Autoencoders etc. Cross-modal representation learning & Transfer learning methods)
• Computational Genomics and Bioinformatics
• Data Science: Data Integration and Visualization, Statistical Modelling 

How to apply:

Applications, including a CV and a Reasons for Applying statement (2,500 character limit), must be made via the University of Cambridge Graduate Admissions website.  Applicants must select up to two different Programmes (a first and second choice out of: Breast; Cell and Molecular Biology; Haematological Cancer; Neuro-oncology/Paediatric Cancer; and Cross-Programme Quantitative Projects) from the five listed, and this should be clearly stated in your Reasons for Applying statement. 


PhD cancer deep learning computational biology Job • 587 views

Login before adding your answer.

Traffic: 226 users visited in the last hour
Help About
Access RSS

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6