Inspiring kick-off meeting of BlueBerry! Towards a blueprint for the EURACAN registry for rare cancers

The BlueBerry Kick-off meeting took place on November 22nd. Over the next two years, European medical centers and organizations will work together to develop a blueprint for a sustainable, scalable and impactful EURACAN registry for rare cancers. EURACAN is the European reference network for adult patients with rare solid cancers. To better understand and treat such rare cancers, European cooperation is essential. That is why KWF Cancer Control funded BlueBerry to set up a European register for rare cancers.

At the Kick-off meeting, four guest speakers gave a mix of inspiring talks and additional insights about the patient perspective and the clinical and technological aspects of the project. Furthermore, representatives of the four work packages presented the progress and objectives of their work package and asked for input from the participants. This resulted in valuable perspectives from a highly multidisciplinary group.

Guest speakers

Asih GolozarDirector of Data Science at Odysseus Data Services and Professor at the OHDSI Center, Northeastern University.

In her presentation Impactful Observational Cancer Research at scale with a common data model Golozar explains the benefits of using the OHDSI approach to enable observational cancer research in a global network setting. Within the OHDSI community, all participating data partners use standardized data, while analytics can be done on site, without data sharing. Using this common data model, the community conducts observational cancer research in a networked setting. The common data model forms the basis for answering a range of questions about cancer, such as disease burden, patient characteristics or treatment patterns. In BlueBerry, all participating centers will harmonize their data according to the OHDSI Common Data Model.

Building the future of observational cancer research together – Asieh Golozar

Eric PostmaProfessor of Artificial Intelligence (AI) at Tilburg University and at the Jheronimus Academy of Data Science in Den Bosch

Postma discussed the use of AI in different settings and zoomed in on the possibilities for healthcare in general and rare diseases in particular. He is very optimistic about the use of AI in healthcare in the future. The breakthrough in AI is deep learning; deep learning can be applied to problems with scarce healthcare data, such as data on rare cancers. Transfer learning can work for rare cancers, assuming cancers that have many similarities. AI algorithms are powerful tools that can surpass human pattern recognition capabilities, but require guidance from human experts who are aware of AI’s strengths and limitations. The challenge is to find a good balance between the best of both worlds: narrow AI (focused on a specific task) and human expertise.

I’m not interested in creating people with AI. We are already good at making people, we don’t need computers for that. But I am interested in getting people to work together meaningfully with powerful and dumb pattern recognition systems (aka AI) – Eric Postma

Judith TaylorePAG EURACAN, Chairman of Thyroid Trust UK, Co-Director and Secretary of the Thyroid Cancer Alliance (TCA)

Taylor sees the establishment of (patient) registries as a global priority for rare diseases. To promote the development of registries and their adoption by patients, we must ensure a careful informed consent procedure. By involving patients in the BlueBerry project, we should be well prepared for their questions:

• How can the registry benefit me and/or future patients?
• What kind of data will be collected.
• How safe is my data; who has access to my data and can the data be linked to me?

Clinical expertise and enthusiasm are clearly not the bottlenecks: We need a registry now, not tomorrow! – Annalisa Trama, Director Fondazione IRCCS Instituto Nazionale dei Tumori (INT), Milan

Dirk Grunhagensurgical oncologist Erasmus MC, with a special interest in sarcomas, melanomas and colorectal liver metastases.

Grünhagen emphasizes the need for cooperation in networks and registries in the field of sarcomas. The incidence of sarcomas in the Netherlands is approximately 700 adults per year; that is 1% of all malignancies. There are 7 Dutch centers of expertise for sarcomas in which 45% of sarcomas are diagnosed and treated. Collaboration in a network ensures faster uniform diagnostics for early recognition of rare cancers. Treatment should be given in centers of expertise by expert teams; sometimes treatment can be done at referral centers. This has implications for the way we collect data in and around the Reference Centers in Europe. Data must be collected for care, auditing and research.

We can improve the prevention, early detection, treatment and quality of life of patients with rare cancers by focusing now on data exchange and interoperability within Europe – Mai Tran, Project Coordinator Infrastructure Dutch Cancer Control (KWF)

Work packages

Four work packages have been defined within BlueBerry.

WP1 Governance and legal framework

The aim of WP1 is to establish an effective governance and legal framework to simplify the use, reuse and sharing of data between different institutions and countries within Europe. The aim of BlueBerry is to provide a register based on the principle of federated learning, in which the data remains at the source. But to properly design such a register, the choices will depend on European legislation (such as the GDPR) and additional national legislation. In order to provide good advice on this, we need input on various aspects of anonymous and personal data from the participating medical centers.

WP2: Data technology and data inclusion

The four challenges of WP2 are:
1. Data harmonisation: using the same data model
2. Connecting individual registers successfully
3. Transparency about data availability and user-friendly analytics
4. Data Protection

Biomeris will work with the participating centers to harmonize the data according to the OHDSI OMOP Common Data Model. The harmonized data will form the basis of the federated learning registry, which will be implemented using the Vantage6 open source technology. At the same time, TNO and partners will continue to develop new developments in the field of health data monitoring in Europe. With the arrival of EDHS (European Health Data Space), it is essential for EURACAN to make sustainable and scalable technology choices. Insights from WP2 are incorporated into the BlueBerry blueprint for the EURACAN registry. The first blueprint will be delivered in May 2023 and the second in February 2024.

WP3: Valorisation and financial sustainability

In WP3, a transparent business model is developed for sustainable financing of building and maintaining a future-proof data infrastructure and data collection of rare cancers. Within WP3, the focus is on exploring what the value proposition could be for paying entities. The business models with the most promising options will be further elaborated in the near future.

WP4: Sarcoma Use Cases

The aim of WP4 is to test the blueprint in practice with three sarcomas use cases, where each use case is of value to a specific target group:

1. a simple use case: identifying the number of patients with a specific (ultra) rare sarcoma in different datasets.
2. a clinically relevant use case: STRASS II is a randomized phase III trial of chemotherapy followed by surgery versus surgery alone to improve disease control and survival in patients with high-risk retroperitoneal sarcoma. Can the BlueBerry data help answer this question?
3. a sustainable use case: can we add value to existing Sarcomen apps (Sarculator and Persarc).

The proof of the pudding is in the eating, so the blueprint should be tested in practice as much as possible- Olga Husson, epidemiologist, Netherlands Cancer Institute

More information

The BlueBerry project started in mid-2022. The project will come to an end in 2024; the intended follow-up is that a registry will also be set up for all other rare cancers in addition to sarcomas. For more information about BlueBerry, please contact Eva Polk, BlueBerry project manager.