Work Package 5

Management Team

Main Objectives 

The main objectives of WP5 are a) to ensure ethical and legal compliance of work performed within Bigpicture in a consistent and coherent manner, b) to establish a regulatory dialogue working group (RDWG) liaising with regulatory bodies for accelerating the utilisation of digital pathology and AI-based approaches, and c) to create, coordinate, and implement quality control for the generation of whole-slide imaging data in an “AI-ready” manner.

Main Tasks

Regulatory framework and standardisation: The current landscape of relevant guidelines (e.g. from medical societies), standards (e.g. data standards such as SEND, DICOM and ISO) and regulatory requirements (e.g. from OECD/FDA/EMA/notified bodies) will be assessed and summarised. Based on this report and interactions with regulatory authorities, standardisation activities will be initiated aiming to contribute to new EU and/or International standards. Furthermore, regulatory issues related to WSI from toxicological studies will be addressed in context of OECD principles of GLP in the nonclinical space. To this end, WP5 will interact with partner FDA and with EMA using existing working groups (e.g. the Alliance for Digital Pathology: digitalpathologyalliance.org, but also use new interaction opportunities (e.g. the EMA Innovation Task Force, Critical Path Institute Meeting for FDA).

QA-QC: One of the known obstacles to adoption of AI is algorithm robustness when using live data from centres not used in the training set. WP5 will include a discrete workflow to create, and coordinate, quality controls and quality assurance for the generation of digital whole slide imaging data sets. This will include developing standards and guidelines for generating WSI data and establishing a European Digital Pathology Quality Control Centre. This centre will distribute reference slides to all labs contributing digital data sets for the assessment of key performance metrics, such as colour reproduction. The centre will then advise on quality issues so the participating labs produce more consistent digital datasets which will add to the ‘value’ of the archive and is a unique element in Bigpicture.

Legal framework of the central repository of WSI: The legal framework will monitor and control the optimal reuse of the digital slides from patients and previous clinical trials to develop AI in pathology in an ethical way and in compliance with applicable regulations such as the GDPR. The project will build on the expertise from Timelex lawyers specialised in GDPR compliance monitoring in the European health sector and from ELIXIR/BBMRI-ELSI, including the ELSI Helpdesk to service our stakeholders in addressing legal and ethical issues regarding sensitive data sharing.

Ethical implications of sensitive data sharing and ethical AI: We will incorporate the ethical guiding principles (lawful, ethical and robust) as described in the ‘Ethics Guidelines for Trustworthy AI’ by the European High-Level Group on AI and adapt these to Bigpicture.