Research Assistant / Associate In Ai-powered Digital Pathology

Salary: Research Assistant £32,546 to £34,132 per annum. Research Associate: £35,116 to £45,413 per annum.

Newcastle University is a great place to work, with excellent benefits. We have a generous holiday package; plus the opportunity to buy more, great pension schemes and a number of health and wellbeing initiatives to support you.

Closing Date: 27 May 2025

The Role: This new position provides an exciting opportunity to develop a project in applied, multimodal artificial intelligence. The role is offered as part of the Medical Research Council (MRC) ‘Impact accelerator award’, a scheme that focuses on developing novel biomarkers to practically improve patient care and treatment. The project will leverage a unique dataset from patients with Mantle cell lymphoma, a rare and difficult to treat cancer.

The candidate will use high-resolution, gigapixel images from biopsies of patients diagnosed with Mantle cell lymphoma (MCL). We will scan and analyse almost 500 patient biopsy slides, and integrate these data with clinical information, other indicators of disease behaviour, and information about how patients responded to treatment. The ambition is to develop a standardised and effective model that can predict how lymphoma will behave in future patients, and to aid the effective choice of treatment.

Mantle Cell Lymphoma (MCL) is a rare lymphoma subtype, characterised by an often-aggressive nature and the lack of curative treatment, despite the impressive advances in new drugs like ‘BTK’ inhibitors and cellular immunotherapies. However, there exists a subgroup of MCL that progresses slowly, where immediate chemotherapy is unlikely to prolong a patient's life. There is currently no robust biomarker to thus differentiate ‘aggressive’ from ‘indolent’ MCL, which exposes some patients to unnecessary treatment and stretches medical resources.

Technological advancements in artificial intelligence (AI)-based computational pathology allow networks to extract defining histomorphological features from tumour sections and represent them digitally in AI architectures; when fused with clinical data and molecular determinants of tumour biology, these methods are anticipated to result in more generalisable and reproducible predictors of disease behaviour. We will develop a clinical risk model that accurately differentiates MCL phenotypes at diagnosis, personalising treatment and optimising resource management.

The candidate will be well supported within a multidisciplinary environment, comprising computer scientists and healthcare professionals. This is a collaboration between Newcastle University and the national MCL biobank (Liverpool). The co-supervisors in the school of computing have the required expertise and equipment to work with large volumes of anonymised data, and this work will complement that of a current Newcastle University PhD student (2024-), who is developing a broad classifier of lymph node pathology. Upon development of the MCL model, we will collaborate with an industrial partner to commercialise and deploy the model across health systems. There will be the opportunity to continue this work, contingent upon additional funding, including in the academic and commercial environment.

The multimodal data for this project comprise digitised images, textual pathology reports, patient metadata, and metrics of clinical outcome. Haematoxylin and eosin (H&E) diagnostic tissue sections from the MCL biobank (-500) will be scanned into high-resolution gigapixel images (40x magnification) at Novopath using whole slide imaging platforms (3DHistech P1500 & Roche). Pathology reports from diagnostic biopsies, along with fully anonymised metadata will pair image and text-metadata for each unique patient.

A multimodal deep learning framework will be developed by the RA under the supervision of Dr Xin (School of Computing). Programming will be in Python, with TensorFlow / Pytorch framework. Each data type will be processed using specialised pre-trained models from the public research community, such as ResNet or Vision Transformer (ViT) for images, BioBERT or ClinicalBERT for text, XGBoost or CatBoost for metadata. Feature representation from each data type will be aligned using methods like Contrastive Language–Image Pretraining (CLIP) to integrate features in a shared latent space. Cross-validation and weakly-supervised learning will help mitigate overfitting, while confidence scores (the model output) and intermediate visualisation will enhance model explainability and clinical usability.

The post is available on a full time, fixed term basis for 12 months.

For informal enquiries regarding the role, Chris Carey (christopher.carey@newcastle.ac.uk).

Find out more about the Faculty of Medical Sciences here: ncl.ac.uk/medical-sciences/

Find out more about our Research Institutes here: ncl.ac.uk/medical-sciences/research/institutes/

As part of our commitment to career development for research colleagues, the University has developed 3 levels of research role profiles. These profiles set out firstly the generic competences and responsibilities expected of role holders at each level and secondly the general qualifications and experiences needed for entry at a particular level.

Goals of the Study:

• Develop a streamlined process to efficiently process and combine these multimodal data.
• Build and validate an advanced deep learning framework to classify MCL phenotypes using the partitioned dataset (training, test, validation). A secondary endpoint is to assess the ability to predict disease progression following treatment.
• Pilot the model as user-friendly software for clinical use. Incorporate explainability features for supporting hypothesis generation and clinical decision-making.
• Validate the model in an independent dataset.

Key Accountabilities:

• Carry out research within agreed timelines, meeting project milestones and to an appropriate standard.
• Undertake research within defined area of expertise (e.g., bioinformatics, programming). You will be required to contribute to all aspects of the work, from the development of protocols through to production of the final report/paper.
• To ensure that knowledge and bioinformatics skills in your own, and related, areas of scholarship are extended and inform research activities and to develop collaborative networks to support analytical method development.
• Attend meetings and other events appropriate to the projects.
• Prepare papers for publication and disseminate findings at conferences as required.
• Contribute to grant applications submitted by others and in time develop own research objectives and proposals for funding.
• Begin to write, with appropriate support, proposals for individual research funding or, where funders do not permit this, contribute to the writing of collective bids.
• Provision of bioinformatic advice for researchers in proposal development and grant writing.
• To present results from your own and your team’s research activity, including information on progress and outcomes and input into the research project’s dissemination, in a variety of formats.
• Play an active role in project management, including the preparation of proposals for research funding, and monitoring research budgets.
• May be involved in the supervision of and providing support to postgraduate research students or Research Assistants.

The Person:

• Experience of conducting research in an academic environment.
• Knowledge and experience in bioinformatics and programming.
• A significant publication record for the discipline, showing evidence of work of at least national quality.
• Theoretical and practical proficiency with Python and/or R, and command line tools; we expect most the programming to be in Python, with either the TensorFlow / Pytorch framework.
• Practical experience of analysis of genomic data (ideally including single cell -omics).
• Practical experience of package development and using publishing tools such as Github.
• Ability to work independently and as part of the research project team.
• Ability to prioritise key tasks.
• Good coordination, time management, analysis, and presentation skills.
• Involvement in drafting manuscripts.
• Experience of project management, including contribution to the preparation of proposals for research funding, and monitoring research budgets.
• Commitment to own personal development.
• Experience with analysis of datasets from computational images (such as digital pathology).

• Demonstrates bioinformatics skills applied to a broad range of computing and biological sciences.
• Keeps up to date with new/improved bioinformatics knowledge, skills, techniques and research.
• Ability to work to deadlines and manage conflicting priorities.
• Excellent organisational, communication and interpersonal skills.
• Commitment to development to include extensive reading and literature/bioinformatics methods review to continue development as a future research leader in the specific field.
• Expectation and willingness to work outside normal office hours including weekends and to work away from main laboratory site including national and international collaborative visits.

• A PhD in Computer Science, Statistics, Mathematics, Biomedical Sciences, Physics or Engineering or a similar discipline with an interest in Life Sciences and experience in Bioinformatics and Programming.

Newcastle University is a global University where everyone is treated with dignity and respect. As a University of Sanctuary, we aim to provide a welcoming place of safety for all, offering opportunities for all.