Check4Cancer Skin Cancer AI Model

• Funding Amount: £9,906.28
  
• Funder: Innovation Voucher, University of Essex
  
• Role: Principal Investigator (PI)
  
• Academic Team: Dr Haider Raza and Professor John Q Gan (PI)
• Associate: Mr Suraj
  
• Year: 2020

Project Summary

This project aims to develop an AI model to partially automate the reporting process for skin lesion assessment, enabling the rapid and accurate identification of cases classified as NOT SUSPICIOUS. This capability will provide reassurance and facilitate prompt discharge, while also supporting the scalability of the business model.

The AI model will utilise digital and dermoscopic images, alongside metadata such as the patient’s age, skin lesion history, and lifetime melanoma risk score, to deliver reliable predictions. As part of the development process, IBM has partnered with C4C to offer a limited Proof of Concept (POC) using their hardware and Visual Inspector software. This POC, planned for October-November 2020, will involve building a machine learning model that integrates images and metadata.

The POC will provide valuable early insights into the model’s potential accuracy, particularly in evaluating the predictive capability of digital images combined with clinical data versus dermoscopic images. These findings will help refine the final project and ensure that the AI model achieves optimal performance, advancing the automation of the reporting process and enhancing operational efficiency.