Work Package 3

Management Team

Main Objectives 

The main objectives of WP3 are:
1. Establish and employ a sustainable node network organisation model responsible for mediating and collecting datasets from beneficiaries and third - parties and to contribute them to the Bigpicture data collection;

2. Define and prioritise datasets within each node from different sources to establish the Bigpicture data collection and to support the innovation of tools and algorithms developed by WP4;

3. Employ software that enables data collection, conversion, standardisation and harmonisation to ensure the highest quality and uniformity of the data, using the infrastructure provided by WP2;

4. Deliver mechanisms of security that guarantee optimal management and fair, regulated access and sharing of data by maintaining their confidentiality and privacy (honest broker concept).

Main Tasks

Phase I. Organisation establishment and start-up of node network

Task 3.1 - Establishing collector nodes. The tasks of collecting and managing Bigpicture’s large datasets will be distributed between 7 nodes that, as functional units act as coordinators, collectors, and distributors of datasets. The coordinating managers will establish a communication network amongst node contributors, functioning as a board, overseeing and coordinating the composition and diversity of the slide collection.

Task 3.2 - Prioritisation of datasets. The Bigpicture consortium has already identified a number of potential data contributors. During the project, the consortium, represented by the node coordinators will continue to work with these data contributors to gain alignment of interests and establish DAA. At the same time, in order for the repository to remain functional and valuable, the node coordinators’ board, in collaboration and alignment with WP1 and WP6, will have a mechanism in place to prioritize the most relevant data sets.

Task 3.3 – Development of ontologies and standardized metadata sets. Even though several previous studies focused on defining ontologies and standardizing meta-data for pathology, no commonly agreed upon structure is yet widely used. This task will address these issues, re-using as much as possible results from previous efforts in this area. A dedicated postdoctoral researcher will be hired to define a common ontology strategy, as well as develop specific extended metadata sets as required for different sets of WSI.

Task 3.4 - Defining and implementing datasets and metadata sets. A common mandatory metadata structure will be defined containing the minimal information required for making a slide usable. In addition, other metadata fields may be relevant for different sets of WSI, since the datasets have been collected for different purposes (e.g. determining the presence of cancer or toxicity of a compound). We therefore will define specific extended metadata sets in addition to common, mandatory metadata. 

Task 3.5 – Identification of appropriate value propositions for potential data contributors. The node coordinators, in collaboration with WP1 and WP6, will develop value propositions tailored to the needs of the different data contributors. 

Phase II. Data collection related tasks

Task 3.6 - Collection of clinical data sets. The histopathological case series cover a wide range of oncological and non-oncological human diseases from all organ systems, well documented with clinico-pathologic and molecular data, prognostic data, information on response to therapy.

Task 3.7 - Collection of nonclinical data sets. The Nonclinical node DECI, having expertise with management nonclinical images and data, will facilitate the collection of all nonclinical slides and metadata. DECI will collect nonclinical cases including relevant metadata from systemic toxicity studies and submit nonclinical images to the main repository.

Task 3.8 - Prospective data collection and scanning. Slide scanning, if needed, will be performed on the existing infrastructure. We have 9 partners with scanning facilities. The node data control officers will coordinate the scanning. 

Task 3.9 - Monitor, review and advise nodes on WSI upload metrics. At the onset of WSI upload into the Bigpicture database, metrics regarding the images and metadata can be obtained and reviewed. This monitoring system will be incorporated two levels of oversight. This will minimise the loss of images and metadata and to maintain a broad network of contributors.

Phase III. Quality control and data conversion, integration and storage related tasks

Task 3.10 - Quality control. A WSI quality assurance system will be run by quality control officers of the nodes. Critical parameters are focus, color spectrum, completeness of scan, resolution, and compliance to the applicable mandatory set and pseudonymization.

Task 3.11 - Data conversion. Many slides have already been scanned, resulting the majority of WSI to be available in scanners’ proprietary format. Common mandatory metadata will be provided for all slides to ensure format conversion and central storage and access.

Phase IV. Honest broker related tasks

Task 3.12 - Honest broker requirements gathering. Drawing from WP2 and WP5, the legal (under the GDPR and Member State law), ethical and contractual aspects of an honest broker model (both for data collection and for data sharing) need to be identified and mapped as well as technical and practical requirements related to, among others, security and ease of use.

Task 3.13 - Honest broker pilot system development. This task includes the development of the honest broker framework, the technical component in order to ensure that a catalogue describing the images and associated metadata in a standardised fashion is available, and the first iteration of honest broker documentation.

Task 3.14 - Honest broker mechanism pilot testing and operation. In this task, pilot users (partners) will be trained to use the catalogue to gain data access by DAC authorisation. At the end of this task, the functionality for downloading minimum level data access to a trusted target (e.g. secure cloud) will be confirmed.

Task 3.15 - Honest broker maintenance. After 24 months, the honest broker will be fully operational and scalable, after which the task includes maintenance and further development based on customer requests and potential regulatory changes.