Our Impact

GEL

Over the last 4 years the BRC has supported the Genomics England 100,000 genomes (GEL) with UCLH and GOSH recruiting a quarter of all UK samples for genome sequencing.  We were the largest single contributor to the 100K genome (GEL) project (n> 6,000 patients and 4,000 families). We have established the data handling and analysis infrastructure to interpret and implement these data into clinical use. Thus we are very well positioned as the next phase of whole genome analyses are rolled out across the NHS (5M).

With 77% of families with no genetic cause identified and over 6,850 families at UCLH out of 13,100 neurology genomes in total there is a huge potential in identifying novel disease genes and risk factors for neurological disorders.

We will continue to build our rare disease cohort, for genome sequencing in the NHSE one million genomes project: this is a tremendous opportunity given our track record in the 100K and the large number of patients seen at UCLH. We are aiming to achieve a precise molecular diagnosis >90% over the next 5 years. This will deliver a hugely valuable resource for UCL/UCLH researchers.

Therapies

The theme works closely with LWENC to deliver early-phase clinical research projects Including the first in human ASO trial in Huntington’s disease (Tabrizi et al NEJM 2019) and BIOMAX-ALS: A study to establish and modulate the relationship between the microbiome- microglia axis in ALS (£1.2 million grant to Dr Sharma from Reta Lila Weston Trust, 100K from MND Association), the first trial of faecal microbiota transplantation (FMT) in any neurodegenerative disease world-wide.

  • We led the first multi-arm, neuroprotective trial in multiple sclerosis (MS) (Chataway Lancet 2020) and the first trial in Parkinson’s disease (PD) (Athauda Lancet 2017) with repurposed-drugs.
  • Phase III efficacy trial of arimoclomol in IBM (heat shock protein upregulation and protein misfolding strategies) currently recruiting - FDA orphan products division ($3M) and Orphazyme Denmark ($14M) new venture capital investment for this trial and now owners of this compound; UK Co-PIS Hanna and Machado; May 2018 signed Royalties share agreement between UCLB and Orphazyme Denmark in relation to possible arimoclomol licensing. May 2019 £256,000 royalties paid to UCLB.
  • We have created a new role to expand on the clinical trials portfolio with 1PA awarded from neuro/dementia themes, to Dr Cath Mummery – as “Director of Innovation and Translation”
  • We established an ION NHNN ALS board including all PIs in this NMD BRC theme along with Prof Nick Wood (BRC Neuro theme Director) and Dr Chris Turner (NHNN Hospital Clinical Director) to support and drive clinical research in this important disease.

AI and Data Science

Neuroimaging

  • We have deepened and widened the application of machine learning across neurology and healthcare generally at UCLH, developing the world’s most accurate predictor of scheduled non-attendance for radiology (published in npj Digital Medicine and widely discussed in the media and within government circles). An AI-enabled non-attendance predictor for outpatient appointments has now entered service at UCLH.
  • We have devised an AI-based method for detecting response to treatment in multiple sclerosis (MS) (published in npj Digital Medicine and discussed in the media (Financial times). Compared with conventional analysis of the traditional MRI measures that a radiologist can extract, AI-assisted modelling of the complex imaging fingerprints was able to discriminate between pre- and post-treatment MRI changes with much higher fidelity. We are extending this approach to predicting cognitive and motor response to treatments, to lead to an individual prediction of treatment response in multiple sclerosis using AI. This work has led to a NIHR Research Professorship (awarded to Dr Ciccarelli) and UK MS Society project grant: total £2.3M. We have detailed clinical and imaging data from 1861 MS patients currently on treatment at UCLH (which is the largest, deeply-phenotyped cohort of MS patients) and initiated a prospective cohort of patients, including children (so far N=140) initiating a new treatment for MS, with genetics, neuroimaging, deep clinical phenotyping, and serum neurofilaments.
  • We have built a pipeline for analysing historic MRI data - 232224 sessions from 134243 unique patients across ~2.8 million volume images by cross cutting with the BRC Neuroimaging Initiative, BRC Machine Learning Neurology Initiative and collaborating with High Dimensional Neurology.

AI and radiogenomics:

  • We developed a state-of-the-art-lesion segmentation algorithm applicable to heterogeneous clinical imaging and used in ~2000 MR studies of patient with glioblastoma, and other brain tumours including meningiomas and pituitary tumours. We recruited two academic clinical fellows (Dr James Ruffle (neuroradiology, pre-doctoral) and Mr Anand Pandit (neurosurgery, post-doctoral), and an engineering PhD student (with Dr Zhang).
  • We commenced a project to apply machine learning to digitised histopathology slides for the prediction of tumour genotype; Machine Learning (ML) for tumour probability mapping and ML-assisted histological diagnosis: this is a new initiative, to date our results suggest a prediction rate for IDH mutations with 85-90% accuracy using structural MRI and perfusion or diffusion techniques (1 paper published and 2 currently under review).
  • AI-driven algorithms for non-invasive assessment of the IDH mutation are under consideration for IP by the UCL Business office.
  • We have built a genetic variant database, Koios, with over 20,000 exomes (with 344 users and over 37 projects), allows researchers to look for novel variants across cases or validate findings, through a web interface. The pioneering platform, largest of its kind, has greatly enabled and accelerated gene/mutation hunting, and importantly shortened the time to clinical implementation, enabling discovery, and promoting open science.

Inter BRC Collaboration

The London UCL MRC Centre biobank supported by UCLH and GOSH BRC works closely with the Newcastle biobank supported by the Newcastle BRC - together this is a national resource of human myoblast and fibroblast lines available for preclinical research and therapy development (e.g. ASO in DMD and CMT and PCT124 in stop codon NMD e.g. currently in preclinical studies in channelopathies and repurposing serine for HSN1). Across the MRC Centre total cell lines in biobank has risen and is now >10674.

With a highly effective tripartite partnership between our 3 BRCs together providing funds to support the successful trials platform for children and adults with trial coordinator support from each BRC. Together to date over 10,987 patients (children and adults) have been recruited into natural history studies and clinical trials.

In addition we secured funding from the North Thames LCRN (£71,495) for staff salary (1 Data Manager and 1 Research Nurse) to support 11 NIHR Portfolio Research Studies carried out at the Centre for Neuromuscular diseases.

Commercial partnerships

  • Medtronic are testing EpiNav developed at UCLH for inclusion in future neurosurgical navigation software packages. UCL have just signed an External research support agreement with Medtronic for a 4 year project (PI: A. McEvoy). Over the last year, the iSYS1 concept robot used in the RCT we finished this year has been re-engineered by Medtronic into the Autoguide device, that incorporates a number of advances and improvements. Approval by FDA and CE marking is expected in the coming months. The aim of this proposal is to demonstrate the utility of the Autoguide device, directed by EpiNavTM, for placing SEEG electrodes and performing tumor biopsies. The study will enroll 75 SEEG patients and 60 brain tumour biopsies.
  • Vitaflo have licensed the IP for our epilepsy dietary therapy (Betashot), have funded an initial tolerability clinical study (completed) and plan to get approval for marketing later this year.
  • UCB have agreed a €2.5 million, 5 year grant with Epilepsy Society for personalised medicine using genomics.
  • The commercialised version of our prototype platform (DIADEM from Brain Miner PLC) is under evaluation on the UCLH clinical network.

EPINAV

The development of the EpiNav™ surgical navigation platform: a multimodal integration software platform EpiNavTM that provides computer-assisted planning for brain surgery and with superiority to manual planning. This software is now being used in over 20 Epilepsy surgery centres in Europe and USA.

We have continued to work to develop previous initiatives funded by Wellcome and The Department of Health. These grants combine the use of novel multimodality imaging techniques for epilepsy surgery, with the potential scope to expand this to neuro-oncology, deep brain stimulation and gene therapy. The PhD student funded on this project has successfully discussed his thesis this year, and a new PhD student has just been appointed to further develop the project over next 3 years.

The aim is to further develop the EpiNav™ platform and to disseminate this software to other neurosurgical units both nationally and internationally.

We focus on:

  • improving computer assisted planning and developing an automated planning system for intracranial electrodes.
  • developing visualisation of seizure onset and spread in the electrodes within the software
  • building resection models of the epileptogenic area
  • developing robotic assisted implantation of electrodes
  • performing a randomised controlled trial comparing manual implantation with robotic implantation of electrodes
  • improving accuracy during manual implantation

We have introduced into EpiNav™ the possibility to use previous trajectories to help the software perform automated planning. Over 100 trajectories that have been planned with EpiNav™ have been gathered into "spatial priors". We have used these priors, with the aid of a machine learning algorithm, to subsequently successfully plan more than 300 electrodes. The use of spatial priors has made planning faster and more consistent between patients. We are in collaboration with the Niguarda Epilepsy Centre in Milan, which is one of the largest European epilepsy units, to also use their trajectories to create “priors”.

The ultimate aim will be to be to disseminate these premade trajectories within EpiNav™ to help other units starting StereoEEG programs to plan in a safer and quicker way.

The same machine learning algorithm has been applied to trajectories to treat mesial temporal lobe epilepsy with Laser interstitial ablation. We have collected the experience of centres in the US, where this technique has been introduced clinically over the last 5 years. This will facilitate the planning of these complex cases when we introduce this surgical technique as a new procedure here at Queen Square over the next year.