UHB WMSDE Digital Histology Annotation Solution

Overview

The WMSDE, hosted by UHBNHSFT is a data storage and access platform that improves patient care by delivering innovation sooner (West Midlands Secure Data Environment (West Midlands SDE). It is part of the national Data for Research and Development programme and is one of several regional platforms in the NHS Secure Data Environment network (The NHS Research Secure Data Environment Network - NHS England Digital). The WMSDE brings together existing collective knowledge and provide conditions for further innovation to benefit people locally and right across the West Midlands region by helping to understand and improve healthcare. Committed to responsible and ethical use of health data, the WMSDE maintains stringent information governance and ethical frameworks. It operates with the necessary approvals from the Health Research Authority, ensuring compliance with all relevant standards and regulations. Digital pathology whole slide images (WSI) represent a rich source of data that can be harnessed to develop algorithms that assist pathologists in diagnosis, providing critical information for treating a wide range of diseases. In creating these algorithms Computer scientists (CS) require WSI data to be prepared to facilitate their work. A crucial step in this preparation involves defining the pathology present within the slide and pinpointing its location so that it can be cross-referenced with the pixel data used by CS to train and test their algorithms. This process, known as semantic annotation, encompasses various levels of detail, from slide-level labels such as tissue type and disease status (positive or negative) to the identification of specific cell types and even subcellular structures, including organelles or viral inclusions. This document specifies the precise requirements for effective and efficient WSI curation, to be deployed within UHBNHSFT technical infrastructure managed by the West Midlands Secure Data Environment (WMSDE), to support the development of artificial intelligence (AI) based technologies. The Supplier shall be responsible for processing deidentified digital pathology image data and the associated metadata for annotation purposes, utilising the techniques outlined below. The data is to be used exclusively for research and innovation activities. All processing must take place within an approved West Midlands Secure Data infrastructure (WMSDE); under no circumstances shall the data be transferred, stored, or processed outside of the intended requirements set out.