World Cancer Day is held every February 4th to raise awareness, improve education and encourage action to advance treatment and increase government actions to help save more lives.  Created in 2000, World Cancer Day has expanded internationally into a positive movement for everyone, everywhere to unite, including the Scottish iCAIRD (Industrial Centre for Artificial Intelligence in Digital Diagnostics) consortium.  This year’s World Cancer Day’s theme is “Close the Care Gap”, which aims to raise awareness of the various barriers that can impact on the care pathway for patients with cancer.

Our work to improve cancer diagnosis through the application of artificial intelligence (AI) is vital as iCAIRD works to establish a world-class centre of excellence. iCAIRD aims to bring researchers, clinicians, health planners, industry & the public together, facilitating collaboration between researchers, clinicians and innovative SMEs (small & medium enterprises) to better inform clinical questions, and ultimately solve healthcare challenges more quickly and efficiently.

To mark World Cancer Day, we wanted to share some of the ongoing iCAIRD research that we hope will revolutionise diagnosis.

Case study: Mammography AI Evaluation

This team in iCAIRDs’ Aberdeen hub are working with the consortiums’ industry SME partner Kheiron Medical Technologies in the area of mammography AI research. Their current work is centered on evaluation & validation of Kheirons’ Mammography Intelligent Assessment (MIA) tool for deployment in clinical breast screening service programmes.  MIA has been developed to support radiologists in breast mammogram reading to enable accurate and rapid clinical decisions. It is intended that deployment of MIA will help manage the radiology workload, improving patient outcomes through facilitating faster screening, diagnosis & treatment.

Dr Gerald Lip, NHS Grampian Breast Screening Service Clinical Director & clinical lead on the project stated: “As someone working specifically in breast cancer screening, I’m excited by the possible opportunities that AI is presenting. Although still in evaluation stages we are already seeing where benefits such as reduction in waiting times and improving accuracy in diagnosis can potentially come from adopting this technology”.

Case study: Digitising Pathology at NHS GG&C

This project is focussed on converting the pathology service clinical workflow at NHS Greater Glasgow & Clyde into one which is fully digital. Currently, patient biopsy tissue samples are prepared onto histology glass slides & diagnosed by pathologists using physical microscopy tools. By changing to a digital service, all slides produced by the pathology service department will be digitised onto Whole Slide Images (WSIs) using a glass slide scanner.

The project team have been busy planning the expansion of their pathology digital archive, which will soon be able to provide the on-going capacity to store pathology sample slide images and data. This will be the main enabler to sustain the digital pathology service and allow it to transition from research into a fully operational system. The team are also developing the required integration mechanisms for the pathology AI research environment, providing secure mechanisms  for sharing of anonymised and approved research data. They are are also finalising their digital workflow integration tools which will link all the processes in the pathology laboratory.

Case study: National Pathology Research Image Archive Service

The ICAIRD project team have created a prototype pathology image archive service that can allow clinicians to upload anonymised image data to a secure, protected data access environment. The intention is for researchers to have capacity to access this archive to create and test machine learning (or other) algorithms to help classify images or indeed aid diagnosis. The archive system makes use of the consortiums’ industry SME collaborating partner Glencoe Software technologies and is hosted in the Edinburgh Parallel Computing Centre at the University of Edinburgh.

Moving forward, the project team hopes to use the experiences and developed solution to move from a prototype service to a fully active resource of large amounts of pathology image data available for clinical & academic researchers.

Case study: Gynaecological Cancer AI

Comprising NHS GG&C pathologists and University of St.Andrews data scientists, the team are working on development and training of AI which can detect cancer in gynaecological tissue samples. The AI is being trained using endometrial and cervical biopsy images from the national database and the aim is for the developed algorithms to be capable of integration into the clinical service workflow across NHS in Scotland to facilitate automated reporting for these biopsies. This has the potential to reduce time taken for gynaecological pathology reporting by up to 50%, allowing for quicker diagnosis and treatment for patients.

Dr David Harris-Birtill, Head of University of St Andrews Medical Technology Team (MedTech) and project research lead: “With this ground-breaking project, my team and I aim to make key advances in detecting cancers that will supercharge our global abilities to diagnose and treat disease far sooner to save lives. Harnessing the power of Artificial Intelligence, we are creating complex new deep learning algorithms to detect and classify endometrial and cervical cancer in histopathology images, enabling better triage of cases and speeding up diagnosis”.

“Working closely with pathologists from the NHS GG&C has allowed us to use real clinical expertise to refine the data that we use to train our algorithms, allowing for better performance in our cancer image classification models and ultimately creating a method of cancer classification that is far faster and more reliable than the human eye alone. We look forward to sharing the results of this extremely promising research in progress”.