We have assembled a team from academia, industry, charity and the NHS to integrate data from diagnostic technologies in novel ways using artificial intelligence algorithms. This will enable the earlier diagnosis of lung cancer for increased patient survival and large time and cost savings to the NHS. Lung cancer is the biggest cause of cancer death in the UK and worldwide, with £307M/year cost to the NHS England. Earlier diagnosis is critical for increasing survival, however the current diagnostic pathway is flawed. Randomised controlled trials show that screening programmes can reduce mortality by 20-26%, and detect co-morbid disease and has led to the establishment of a new £70M NHS England lung cancer screening programme, launching this year and running for four years. There are at present no commercially available tools that are proven to improve patient care compared to the current screening guidelines. To address this need, we will use our established data infrastructure to collect and transfer clinical data, CT scans, digitised images of stained tissue sections (digital pathology) and blood-derived data from the consented participants of the lung cancer screening programme to our secure data 'lake' based at the University of Oxford. For the first time, we will integrate these diverse data types using our Artificial Intelligence algorithms to enable further and improved characterisation of disease, than is possible by a radiologist alone. By linking the additional information available at diagnosis to outcome data, we will also be able to refine the lung cancer treatment guidelines. We will also link to data from primary care to better define risk in the general population. With this approach, we aim to (1) more accurately diagnose lung cancer with enhanced prognostic information; (2) reduce the occurrence of harmful invasive procedures in the diagnostic pathway; (3) improve patient selection for lung cancer screening and reduce costs; (4) improve assessment of risks from co-morbidities; (5) generate and store a large amount of data that can be used for future research. We will define a new set of standards for lung cancer diagnosis that will improve patient care and generate large cost and time savings to the NHS, and also help patients with lung cancer elsewhere in the world.

"We have formed a consortium of collaborating hospitals and commercial companies with public and patient involvement and input from charities to enable us to work with the public and the NHS to best develop and test the use of Artificial Intelligence programmes to improve health care. We are working with university academics and NHS hospitals across the country, from the south-west into London and including the North of England and Scotland. Our programme includes predominantly UK based companies ranging in size from a few employees to hundreds. We aim to learn about the difficult areas of ethics and public acceptance for machine learning and decision making in health care, and the use of patient data by commercial companies. We will put in strict safeguards to ensure patients and all involved in health care are confident that their data is managed securely and in a manner that would meet with their approval. Our ambition is to develop a relationship with patients and the NHS that will enable improved development and use of Artificial Intelligence to enhance the delivery of safer, better care, with clinical and economic gains to the NHS. Our programme of work will include the analysis and testing of algorithms on data for which we already have patient consent, and the collection of new patient data for which we will gain consent. This data will be securely stored and made available under strict rules to our collaborating academics and universities, and to commercial companies. Once we have developed and tested the Artificial Intelligence algorithms we will then test them in the NHS following applications to, and approvals from Research Ethics Committees."

Optellum’s vision is to enable earlier and more confident cancer diagnosis and treatment by using Artificial Intelligence (AI) and Deep Learning to unlock new insights in huge image databases. Our first product that will target early detection of lung cancer, the World’s most common and lethal cancer. The US market holds the biggest potential due to established reimbursement for lung-cancer screening. However, it presents special challenges, due to 1) Complex FDA regulatory regime 2) High data variability, combined with legal hurdles for accessing patient data for algorithm training 3) Different customer needs. This project will enable us to establish partnerships with US hospitals that will help us accelerate US market entry by 1) Developing FDA regulatory strategy 2) Creating a roadmap for the training and validation of our algorithms on US patient datasets 3) Developing product specifications that fulfil US customer needs.