10035053Digital skin cancer diagnosis modelClosedCollaborative R&DInnovate UKCheck4Cancer(C4C) is an award-winning cancer screening&diagnosis provider, and the current skin diagnosis pathway is based on consultant-led telemedicine reporting of digital and dermoscopic images taken in a clinic setting. Rates of skin cancer having been rising significantly during the last three decades and, waiting times for assessment of a suspicious skin lesion are long in both NHS and private sectors in the UK and, these waiting times have been amplified during Covid. C4C wants to shorten waiting times for skin cancer assessment in public and private sectors, giving people with non-suspicious skin lesions a rapid reassurance, and clearing the way for suspicious lesions to be treated quicker. We therefore intend to build a digital skin cancer detection model that supports the current clinical model, with development of a front-end Skin App to capture high quality images of suspicious skin lesions and clinical data on mobile devices for analysis from home rather than clinic. This model will provide a digital solution for patients who cannot access our UK network clinics or would simply prefer this approach. Patients with suspicious lesions will be advised to attend a face-to-face consultation for further investigation, including biopsy, as happens with the clinical model. Following a successful short Proof of Concept (PoC) project with IBM in Nov-20, where a primitive model was able to detect all the cancers and correctly identify 50% of the non-cancerous skin lesions, C4C is currently working with the University of Essex to develop an AI model using C4C's bank of skin images and associated key clinical data to partly automate classification of skin lesions into "suspicious" or "non-suspicious" and this will be integrated with the Skin App. Check4Cancer was recently successful in an application to Innovate UK to fund a Knowledge Transfer Partnership (KTP) 21\_22 R5 (Project number 10029141) to build the AI model. **The main area of focus for this project is to:** * Research&develop a front-end Skin App to capture high quality images of skin lesions and data on mobile devices for analysis from home. * Compare the Skin App images with traditional digital and dermoscopic images taken in clinic * Test the suitability of the Skin App images for telemedicine reporting * Develop an AI model to partly automate the skin lesion classification * Integrate the Skin App with the PMS and the AI model, to allow rapid reassurance of patients whose lesions are in the not suspicious group following initial assessment.