ACED Funded Research Projects at UCL

ACED Pilot Awards 

Early Detection of Prostate Cancer Progression in Active Surveillance 

• Prof Dean Barratt, UCL 
• Prof Geoffrey Sonn, Stanford University 

Prof Barratt and Prof Sonn will develop software that analyses changes in serial multiparametric MRI images to determine radiological and pathological progression of prostate cancer. In conjunction with other clinical data, it will allow for the early detection of disease progression in prostate cancer patients on active surveillance. 

Dynamic predictive model for baseline early detection and follow-up re-evaluation of the risk of prostate cancer progression on active surveillance (PROGRESS Prostate) 

• Prof Alexey Zaikin, UCL 
• Dr Tristan Barrett, University of Cambridge 

Using cutting edge modelling methodologies, Prof Zaikin and Dr Barrett aim to develop a predictive model that will be able to provide a baseline estimate as well as continuous re-evaluation of the risk of prostate cancer progression. Their ultimate aim is to produce an open-source clinical decision support tool to be used by both patients and clinicians.

REPRESENT: A community engagement roadmap to improve participant representation in cancer research early detection

• Dr Ignacia Arteaga, University of Cambridge
• Prof Jackie Shannon, OHSU
• Prof Nora Pashayan, UCL
• Dr Bella Starling, University of Manchester

Prof Pashayan and Drs Arteaga, Shannon, and Starling aim to address the uneven rates of participation in cancer research across sociodemographic and protected characteristics by carrying out engagement activities. They will draw together their findings to design a community engagement roadmap for the improved representation of underserved communities in cancer early detection research.

Real-time cancer analytics (REACT)

• Dr Neal Navani, UCL
• Dr Thomas Callender, UCL
• Prof Mihaela van der Schaar, University of Cambridge

Large-scale, high-quality datasets from research cohorts and real-world healthcare services are rare and access is tightly controlled and subject to delay. To remedy this, Prof van der Schaar and Drs Navani and Callender aim to develop a privacy-preserving platform allowing for the generation of bespoke synthetic datasets from real-world healthcare records and research cohorts.

Delineating the microenvironmental contexture of the radiologically progressing prostate cancer lesion

• Dr Vasilis Stavrinides, UCL
• Dr Ece Eksi, OHSU

Recognising which actively surveyed prostate tumours will transition to clinical significance requires understanding of their radiological, microstructural and molecular evolution. To better understand the underpinning biology, Mr Stavrinides and Dr Eksi aim to systematically describe microstructural, immunohistochemical and molecular information from lesions sampled before and after radiological progression. The outputs from this work will inform the design of larger studies and the development of longitudinal models that predict progression to clinically significant disease.

ACED Project Awards  

Risk-representative targeted biopsy for early prostate cancer detection 

• Dr Yipeng Hu, UCL 
• Prof Geoffrey Sonn, Stanford University 

Early stage prostate cancers can often be too small to visualise on medical images, and some identified cancer targets can be missed or under-graded with a biopsy. Dr Hu and Professor Sonn will use deep learning-based approaches to capture the links between prostate cancer imaging and retrospective histopathological evidence allowing personalised targets to be predicted without the need for biopsy. The team ultimately aims to develop an artificial intelligence agent to be trained to target the most risk-representative regions, allowing new cancer targets to be generated for individual patients in real-time. By developing a novel, real-time biopsy sampling strategy, they hope to maximise biopsy efficiency and better stratify personalised risk in prostate cancer. 

Real-world risk-stratified early detection and diagnosis using linked electronic health records data 

• Prof Georgios Lyratzopoulos, UCL 
• Prof Antonis Antoniou, University of Cambridge 
• Dr Ruth Etzioni, Oregon Health & Science University 

Many patients, particularly those with harder-to-suspect cancers, present to GPs with non-specific symptoms. Professors Lyratzopoulos, Antoniou and Etzioni and their collaborators will produce novel algorithms using both state-of-the-art and novel computational approaches to improve the calculation of risk of symptomatic-but-as-yet-undiagnosed cancer. They will analyse longitudinal primary care data, including on prescription history, comorbidities, diagnostic tests and underlying susceptibility to cancer. This work aims to inform referral guidelines for specialist investigation or assessment, and to improve decision tools supporting doctors in their assessment of cancer risk. 

Computational modelling of tumour growth kinetics to inform early cancer detection 

• Prof Nora Pashayan, UCL 
• Prof Paul Pharoah, University of Cambridge 
• Dr Johannes Reiter, Stanford University 
• Prof Sylvia Plevritis, Stanford University 

Cancer is a heterogeneous disease with varying rates of growth and malignant potential. To reduce death from cancer it is important to base early detection initiatives on an understanding of the growth rate of a tumour and the relationship between growth rate, tumour size and its metastatic potential. Professors Pashayan, Pharoah, Reiter and Plevritis will construct mathematical models to simulate growth and metastatic spread of a range of cancers based on comprehensive time-course data from patients that have not received active treatment. This will identify the most promising early detection initiatives that can improve prognosis without risking overtreatment.

Luminal Index MRI Identification of Treatment critical Prostate Cancer (LIMIT PCa) 

• Prof Shonit Punwani, UCL 
• Prof Fiona Walter, University of Cambridge 

Profs Punwani and Walter will undertake a multi-site, non-randomised study to evaluate the feasibility and acceptability of a novel and potentially game changing MRI technology, Luminal Index MRI (LI-MRI) in the early detection of prostate cancer. LI-MRI is simple, cheap and has good performance making it amenable to be delivered as part of a community-based testing strategy. This will allow the opportunity for earlier detection of prostate cancer in men that would not otherwise be scanned. This project will provide the first step to a larger national trial using this new MRI technique to allow earlier detection of prostate cancer in the 16% of men that present with late stage metastatic disease.  

Establishing optimal methods for cancer early detection using methylation markers (EPITOME): An integrated comparison of three leading-edge platforms for cell-free DNA epigenetic analysis and application to multiple cancer types 

• Dr Charlie Massie, University of Cambridge 
• Dr Andy Feber, UCL 
• Dr Olivier Gevaert, Stanford University 
• Dr Parag Mallick, Stanford University 

To allow the detection cancer at an early stage and improve the diagnostic pathway requires minimally invasive tests which are cost effective, accessible and accurate (e.g. high positive-predictive values for enriched screening populations). Drs Massie, Feber, Gevaert and Mallick together with Dr Chunxiao Song from University of Oxford will directly address this challenge in the context of cell-free DNA (cfDNA) analysis, focussing on DNA methylation markers as the most promising target to overcome cfDNA detection limits and determine tissue-of-origin. This ACED Project will directly compare three new methods using blood/urine/proximal samples and identify the method(s) with the best sensitivity and specificity for a given application. This will generate new SOPs and data analysis workflows that can be utilised in future early detection studies across the ACED network. 

REPRESENT: A Community Engagement Roadmap to Improve Participant Representation in Cancer Research Early Detection

• Dr Ignacia Arteaga, Dr Perveez Mody, University of Cambridge
• Prof Jackie Shannon, OHSU
• Prof Nora Pashayan, UCL
• Dr Bella Staring, University of Manchester

Members of the public are at the heart of any early cancer detection effort; they are pivotal to make early cancer detection approaches socially acceptable and to increase the uptake of detection technologies. Authentic and sustained community engagement between health research centres and those communities is therefore necessary to improve diversity in cancer early detection research. Informed by published research toolkits (e.g. NIHR INCLUDE, EoE CLAHRC, THIS PPI), Prof Pashayan and Shannon and Drs Arteaga, Mody, and Starling aim to address the uneven rates of participation in cancer research across sociodemographic and protected characteristics by carrying out engagement activities. They will draw together their findings to design a community engagement roadmap for the improved representation of underserved communities in cancer early detection research.

MRI signatures of tumour-promoting microenvironments as an early warning system for prostate cancer progression

• Dr Vasillis Stavrinides, UCL
• Dr Ece Eski, OHSU

Advancing age, infection and injury cause inflammatory changes in the prostate. Unanticipated, these microenvironmental perturbations promote oxidative damage, enable malignancy through a series of molecular alterations and determine prostate MRI phenotypes targeted with a biopsy needle. Until recently, exploiting the sensitivity of the immune system for prostate cancer early detection was uniquely difficult, largely because standard prostate sampling fails to detect significant cancers in up to 50% of cases or to serially capture tumour-infiltrating immune cell populations within evolving tumours. Cancer initiation and growth are linked to subverted immune balance, and early tumour-promoting prostatic microenvironments have characteristic MRI signatures that predict subsequent transitions to clinically significant cancer. Drs Stavrinides and Eksi aim to systematically categorise MRI-targeted prostate lesions according to their microenvironmental topography and derive an early detection immunoradiological signature that identifies aggressive prostate cancers before they become untreatable.

Comprehensive genomic analysis of tumour and host interactions in the genesis of kidney cancer

• Dr Sarah Welsh, Prof Grant Stewart, Unviersity of Cambridge
• Dr Maxine Tran, UCL

Renal cell cancer (RCC) is the 7th commonest cancer and the most lethal urological malignancy with 50% of patients dying from their disease. Small renal masses represent a stage of RCC where progression and prognosis varies, so there is a need to identify which are likely to progress and which are unlikely to cause future harm and avoid overtreatment. In this project, Dr Welsh and Miss Tran aim to determine the relationship between tumour progression and the timing of key genomic events by mapping molecular archaeology with tumour growth rates from 3D CT images. This will determine whether key driver mutations predict progression of small renal masses, and whether other features of the tumour microenvironment or host immune system predict progression of small renal masses. Ultimatley this will enable development of a multiparametric test to allow clinicians to tailor treatment according to how we think their tumour will progress and, if detectable in blood, addresses the urgent need to develop an RCC screening test.

ACED Programme Awards

Exploiting the Immune System for Early Cancer Detection

• Dr Jamie Blundell, University of Cambridge
• Prof Robert Bristow, University of Manchester
• Dr Evan Lind, OHSU
• Dr Parag Mallick, Stanford University
• Prof Benny Chain, UCL

Early‐stage tumours are small, often measuring <1 cm across. Detecting such small tumours is hard because the cancer‐specific signals are weak and difficult to spot using imaging and traditional biomarkers. Improving early cancer detection requires new approaches that have the ability to amplify these weak cancer signals. The cells of our adaptive immune system (B‐ and T‐cells) are exquisitely sensitive detectors of disease. They can detect tiny quantities of diseased cells by sensing the presence of rogue proteins using specialised receptors on their surface. Once rogue signals are received, immune cells rapidly divide in number thus naturally amplifying an initially weak signal. B‐ and T‐ cells are now known to play a central role in recognising and responding to early cancers. This raises the prospect that the body’s own immune system could be used as a powerful early warning system for cancer. However, to achieve this, we need to know (i) which specific immune receptors are responsible for cancer‐surveillance, and, (ii) whether these cancer‐specific immune signatures are detectable in a blood sample. To answer these questions we have assembled a multidisciplinary team of experts in immunology, clinical cancer genomics and computational biology/AI from acrossthe5ACEDcentres.Ourprogrammeofresearchdrawson unique clinical cohorts in which blood samples are collected in people serially before cancer develops. Our team will use tumour samples from these studies to understand which immune cells are responsible for cancer‐surveillance and will then use the matched blood samples collected before diagnosis to determine whether these cells could have been detected years before clinical diagnosis using a simple blood test.

ACED Pathway Awards

Developing and evaluating Bayesian multilevel models to improve cancer risk prediction for patient groups with rarer presenting features and patient characteristics

• Dr Matthew Barclay, UCL
• Mentors: Prof Georgios Lyratzopoulos, UCL and Prof Antonis Antoniou, University of Cambridge

Diagnostic investigations for cancer are targeted at patients at higher risk. This targeting is appropriate, as diagnostic investigations can be harmful. But it relies on accurate assessments of the risk of cancer for individual patients. Currently we can assess risk of cancer well for large groups, but it is difficult to produce accurate risk estimates when there are only a few patients in a group. Many small groups exist, including patients presenting with rarer symptoms and patients belonging to minority ethnic groups. Such small patient subgroups are underserved because they are assumed to have the same risk as groups which are more common. This results in some patients not receiving diagnostic investigations that would be helpful or being investigated more often than is appropriate.

One approach to trying to improve upon this problem is to try to account for relationships between different groups by using more complicated statistical methods. These methods take advantage of inherent similarities between certain rarer and more common features. For example, different symptoms relating to the same part of the body may be expected to have similar predictive value for cancer, other things being equal. Accounting for this allows ‘borrowing’ of information between similar characteristics, partly addressing small sample size limitations to allow more accurate risk predictions. This proposal aims to develop prediction models for risk of cancer following primary care consultations for new symptoms. We will then examine how much these methods improve the accuracy of risk estimates for rarer symptoms in different data sources

ACED Skills Exchange and Development Awards 

Dr Tanveer Tabish 

Tumour-derived exosomes may be used as non-invasive biomarkers for early detection through their ability to transport proteomic and genetic material around the body before the subsequent establishment of metastatic lesions. Quantification of HER dimers within blood harvested exosomes has added significant information towards the early detection of lung cancer. Quantification of HER dimers within blood harvested exosomes has added significant information towards the early detection of lung cancer. This Award will allow Dr Tabish to spend time in the laboratory of Dr Roisin Owens at the University of Cambridge to facilitate necessary training in relevant technologies of lab-on-a-chip bioelectronics technologies for single exosome measurements. Dr Tabish will also share expertise in fluorescent imaging microscopy for the quantification of HER dimers within blood harvested exosomes and their role in early detection of cancer to support future collaborative research with the Cambridge ACED Centre.

Dr Vasilis Stavrinides

Epigenetic changes have been found in negative MRI-targeted prostate biopsies, which often precede a subsequent cancer diagnosis. Biopsy-negative prostate MRI lesions could therefore represent pre-malignant states that precede the manifestation of clinically significant cancer. Understanding lesion dynamics requires a fresh approach towards the systematic molecular characterization of MRI-targeted biopsy prostate tissue, particularly when serially derived from the same man. The analysis and interpretation of composite genomic and epigenetic data that are both highly dimensional and longitudinal requires advanced computational approaches and sophisticated statistical modelling. Mr Stavinides will learn from world-class cancer evolution experts at the Blundell lab at Cambridge and apply the acquired knowledge in the prostate cancer domain, with a view to a new UCL-Cambridge ACED collaboration.

ACED PhD Studentships

Miss Lucie Gourmet 

"Investigating emergent properties of cancer cells and their environment to make sense of early carcinogenesis"

Lucie's PhD project, supervised by Prof Simon Walker-Samuel, involved studying the ecological interactions occurring within cancer cells (including competition and cooperation) and in the tumour microenvironment (blood vessels and the immune system). Lucie's research interests are in the impact of cancer cell mutualism on immune predation during early tumour formation using High Resolution Episcopic Microscopy to visualise the impact of angiogenesis on tumour morphology.

Mr Callum Oddy

Callum's research, focuses on dissecting the evolutionary processes that are driving precursor lesions to become cancer. Using normal human gastric tissue and intestinal metaplasia tissue (a known precursor lesion to gastric cancer), he will be developing a pipeline to create organoids and evaluate their genetic profiles and epigenetic landscapes. Using these profiles, Callum plans to interrogate individual clones’ fitness effects to determine how precursor lesions encourage a cancer phenotype, under the supervision of Dr Marnix Jansen. Together this research, will be used to help identify and quantify genetic and epigenetic patterns that can be applied to a patient’s risk of cancer progression over time.