Check4Cancer launches skin cancer AI project with Essex researchers

Artificial intelligence (AI) could be used to speed up diagnosis and treatment of skin cancer for NHS and private patients thanks to a partnership between award-winning cancer screening and diagnosis healthcare provider Check4Cancer and researchers at the University of Essex.

The new Knowledge Transfer Partnership (KTP), supported by a grant from Innovate UK, will use artificial intelligence (AI) to analyse images to detect skin cancer at an early stage. The project will develop an AI powered model based on Check4Cancer’s vast bank of skin lesion images which are all labelled with a wealth of clinical information. Skin lesions are any skin area with different characteristics from the surrounding skin, including colour, shape, or texture, so can be an indicator of various conditions including skin cancer.

Run by Innovate UK, KTPs are part-government-funded project-based collaborations between universities and businesses created to help UK companies innovate and grow. Projects are designed to address a specific challenge or business need and give companies access to academic expertise and a grant of up to two-thirds of the project cost.

This award of over £180,000 funds a data scientist to work full-time on the project for two years. The project will build on the work of the Check4Cancer’s nurse-led skin diagnosis pathway for insured and self-pay patients that tackles long waiting times and helps to avoid unnecessary biopsies and surgical procedures. More than 80% of patients are offered an appointment within five days and only 14% of patients need to be referred on for biopsy or treatment.

Professor Gordon Wishart, Chief Medical Officer of Cambridge-based Check4Cancer, commented: “By working with AI experts in image recognition, machine learning and algorithm development from the University of Essex, we plan to harness both academic excellence and industry expertise to build an AI model that replicates the clinical excellence of our current clinical skin cancer model. We have more than 70,000 images, each tagged with significant clinical data, which gives us a powerful advantage in developing a model built on images as well as clinical data. I am delighted to be launching this project to tackle major healthcare issues, with our local academic partners at the University of Essex.”

He added: “We hope that the final model will not only be used in the private sector, but will also benefit the NHS by allowing patients with suspicious skin lesions to be triaged in primary care so those who require further investigation or biopsy are referred for hospital care sooner.”

The project is a collaboration with the School of Computer Science and Electronic Engineering at Essex, it aims to identify major cost reductions in the assessment of skin cancer across both private and public sector health care.

Dr Haider Raza will lead the project with the support of Dr Alba Garcia Seco De Herrera and Professor John Gan. A KTP associate has been appointed to start working on the project.

He said: “This project is providing us with a great opportunity to collaborate with medical experts from Check4Cancer. Deep learning-based computer vision models are highly accurate and can achieve great performance in well-defined balanced datasets.

“We want to use the power of these types of models in a healthcare setting by using Check4Cancer’s data and image bank to detect melanoma skin cancer at earlier stages.

“We will also be looking to minimise false negatives and ensure our models can be easily understood and interpreted by health professionals in clinical settings.“

Dr Raza said his team would be aiming to publish academic articles based on their work to show the impact of the project.

Sarah Taylor, Head of Specialist & Practitioner Relations at AXA Health, advised that: “Since working with Check4Cancer our insured members have benefited from fast access to this streamlined pathway, with the unnecessary biopsy rate reduced by more than 50%. We look forward to the launch of Check4Cancer’s skin cancer AI model to increase the network capacity and provide our members with a choice of traditional or digital healthcare.”

This KTP grant award is in addition to the Innovate UK Smart Grant also awarded to Check4Cancer this year. The UK Smart Grant is supporting the development of a ‘Skin App’ to take high-quality images to assist the assessment of skin lesions. This will be integrated with the company’s Patient Management System and, the funding will also support commercialisation of the digital skin platform.

Check4Cancer enables people to take ownership of their health and that of their families, by providing cancer services that are innovative, validated and effective. The company’s team of cancer specialists has extensive clinical expertise, enabling Check4Cancer to provide award-winning services trusted by leading UK insurance providers and corporate sector clients.