An overview of our latest webinar series

EU4CHILD is coming to an end… and to highlight its milestones and achievements thus far we hosted a series of webinars between the 27th and 28th of September 2022, with a focus on explaining the three fundamental pillars driving the purpose, goals and results of EU4CHILD:

  • EU4CHILD Paediatric cancer Artificial Intelligence (AI) framework and Knowledge Base
  • Our Community Building and Policy Approach
  • EU4CHILD’s State of the Art and Roadmap for Implementation

Find out more about our results and recommendations by watching the full webinar recordings below!

Webinar #1 - EU4CHILD's pediatric cancer Artificial Intelligence (AI) framework and Knowledge Base

This first webinar focuses on our vision for an AI framework and how it can help in the fight against Childhood Cancer. It begins with a brief introduction of this preparatory action funded by the EU to set the path for further developments in the area of AI for Pediatric Cancer.

But what are these three main pillars? First, we define the state-of-the-art by studying in detail the resources and use cases already available around the topic of AI and pediatric oncology. Next, we plan out the necessary steps to build a solid Knowledge Base that can act as a single access point to boost knowledge exchange and collaboration opportunities. Third, we bring together an active and engaged community. Our main objective is to provide the starting guidelines to future projects and developers to create and operate a platform that allows for efficient data sharing within the field of childhood cancer, improving the healthcare sector through the implementation of AI.
In the end, it is all about data. As presented by our technical experts from UPM, data is the fundamental input required to design and execute AI algorithms that bring understandable and sophisticated results. There are many factors that influence this process, such as the source of the data, its quality, type and all that makes up the data environment per se. The discussion continued by UPM’s detailed walkthrough of our Knowledge Base and the process of EU4CHILD’s platform service definition, based on a ulti-dimensional decision tree designed to guide IT experts and researchers through the development of fit-for-purpose AI solutions.

 

Webinar #2 - EU4CHILD’s Community Building and Policy Approach

Our second webinar highlighted the community building efforts undertaken by EU4CHILD and our colleagues at IFIC. This is an essential part of the project that builds on the idea of enabling the free flow of information between all key stakeholders in pediatric oncology and AI. As such, community building has established three main goals: liaise with ongoing EU initiatives including ERN-PaedCan and other EU-funded initiatives for knowledge exchange in this area; connect with existing networks and to create a space for collective learning, sharing and influencing actions for our multidisciplinary community across the EU.

Our fieldwork activities (mainly through focus groups, interviews and stakeholder consultation) has been a fundamental part to collect insight on how to design this space. An important challenge was also to find and engage with experts and policy makers that had an established understanding on both AI and pediatric cancer.

EU4CHILD’s policy approach came up next. This is a field of complex political and ethical dimensions, deriving from a wide array of stakeholders with multiple interests and sensibilities. One key dimension is time. We need to understand the momentum and recognize that there is a great opportunity to make a real impact in this field, supported by the mandates of the European Union and its commitment to fight childhood cancer and increase the uptake of AI across the EU. We must recognize that policymakers have a critical role in mobilizing the broad community around AI and pediatric cancer towards a common goal.

 

Webinar #3 - EU4CHILD’s State of the Art and Roadmap for Implementation

The last in this webinar series focuses on the State of the Art and EU4CHILD’s roadmap. These are essential elements to investigate in order to establish the foundations from which future projects related to AI and pediatric cancer can build from.

The first topic in this discussion is the user and ecosystem requirements and methodology, led by partner USAAR. This is a fundamental part of the project itself, which is to set the stage for the successful implementation of a collaborative childhood cancer ecosystem. The definition of stakeholders in this field and therefore user research, was carried out starting from a comprehensive study of existing initiatives and projects around this field. In addition, there are many ethical and political issues to take into consideration, apart from all technical aspects of AI.

OPBG, another of our clinical experts, then walked the audience through EU4CHILD’s State of the Art investigation, which included looking through biometric analyses, review of reviews, registered clinical trials, existing repositories and registries, ethical challenges and existing AI applications for healthcare and telemedicine. A critical hurdle here is obtaining real world data, including also more unconventional sources such as social media, patient reported outcomes and more.

Our colleagues from ICH then introduced the vision behind our roadmap for the successful implementation of the EU Childhood cancer ecosystem. For the purpose of building a childhood cancer ecosystem, this roadmap is not a conventionally linear one, but rather a matrix that involves diverse stakeholders and disciplines. In this sense, clinical and quality-of-care, data quality and availability and the integration of ethical/legal perspectives have been singled out as priority areas of intervention to foster research with AI in pediatric oncology.

 

 

Photo by Piron Guillaume on Unsplash

A joint talk on ethics with PERSIST

EU4CHILD and the PERSIST project have met in a first insightful discussion about cancer and the ethics surrounding it, in an age where advanced technologies such as Artificial Intelligence are taking center stage.

What is PERSIST about? The PERSIST project works to provide a patients-centered survivorship care plan after cancer treatments, based on Artificial Intelligence and Big data technologies. Both in the field of oncology paired with technologies such as AI and big data, EU4CHILD and PERSIST hold deep potential synergies that must be explored in order to maximize results.

During the meeting, the power of ethics was discussed, especially when it comes to sensitive subjects such as childhood cancer and related data. Childhood cancer is a rare disease, of varied nature, requiring different approaches. When it comes to childhood cancer, the emotional distress among family members and parents is a relevant factor, affecting the treatment process, including how AI in medicine in general, and prediction models in particular, need to be understandable/explainable to a non-expert audience. The presentation was led by Norbert Graf, from the EU4CHILD consortium, and it discussed important questions related to the topic, such as:

  • How high is the risk of re-identification for anonymized medical data?
  • How freely can informed consent be given under these circumstances?
  • How can this be achieved for patients of different age groups and their parents as well?
  • Who can be responsible in the case of a diagnostic or treatment error?

A discussion highlight was how a central repository may not be suitable for AI applications, considering how medical experts have struggled to develop central solutions for the last 30 years and have so far been unable to make them work. This is because central repositories can only use fixed schemes and there are tremendous barriers against data sharing. In addition, real-time updates for AI algorithms should be considered instead of batch learning.

After a very successful first meeting, we now prepare for a second encounter in September with our PERSIST colleagues to further exploit our synergies and power towards a better approach in AI applied to cancer pathways.

 

 

Photo by Marcus Wallis on Unsplash