Using patient data to enable personalized healthcare, ATHENA shows us how!

Share this article

Personalized healthcare is characterized by tailoring treatments to each individual patient’s needs. Despite evident benefits, implementing such a system is not straightforward. The ATHENA project consortium has successfully devised a set of building blocks to address the challenges associated. Both technological and governance tools were developed that can now be used on a larger scale to catalyze the transition towards personalized medicine and care.

The road towards personalized medicine

Although we share more than 99% of our DNA with each other, we have very distinct physical and psychological characteristics. We can be more or less susceptible to certain diseases, respond differently to treatments, and have distinct priorities and preferences in our care pathway. While in traditional medicine, this is rarely taken into account, personalized medicine focuses on providing the best possible care to each individual patient.

“Getting the right therapy to the right patient, at the right time, that is what personalized medicine is all about.” – Bart Vannieuwenhuyse, Johnson and Johnson

“In ATHENA, we aim to develop essential building blocks that are needed to transition towards such a care system based on personalized medicine,” says Bart Vannieuwenhuyse, Senior Director Health Information Sciences at Johnson and Johnson and co-lead of the ATHENA project. ATHENA, short for Augmenting Therapeutic Effectiveness through Novel Analytics, is a collaborative partnership, funded by the Flemish government (VLAIO), bringing together world-class research institutes and medical data partners. Project partners Janssen Pharmaceutica, Illumina, Inovigate, Robovision, imec, KU Leuven, and UGent lend their expertise to the project. The hospitals partners include AZ Groeninge, CHU de Liège, OLV Aalst, and ZOL Genk. Currently, most of the progress in personalized medicine has been applied to oncology. Therefore, ATHENA specifically focuses on two oncological disease cases: multiple myeloma and bladder cancer.

Building block 1 – Integrating data

In personalized medicine, proper patient stratification is essential. This can only be accomplished by linking patient characteristics (age, gender, etc.) to clinical data, for instance, responsiveness to treatment. Treatment efficiency is influenced not only by the clinical status of the patient, but also by the specific driver mutations present in the patient’s genome. This is especially prevalent in oncological cases. “Machine learning can help us to discover such patterns in patient properties, yet requires a lot of data,” emphasizes Vannieuwenhuyse. “One of the goals of ATHENA is to combine clinical patient information, typically recorded in electronic health records, with genomic data.” 

Integration of clinical and genomic data, together with other sources of information, such as imaging material and patient reported outcomes, should be used to stratify patients and match them with their best possible care pathway. “Unfortunately, at the moment, all these valuable data sources are still very fragmented and isolated,” says Tine Lewi, Senior Director, Global Data, Platforms & Partnerships – Global Commercial Data Science at Johnson and Johnson and ATHENA co-lead. In addition to issues regarding data integration, the insufficient availability of high quality data poses a challenge. Yet improvements are underway. “Project ATHENA has succeeded in overcoming some of these challenges by introducing innovative technological and governance solutions,” Lewi adds. “With what we have learned during the project, we can perform more comprehensive and faster analyses of large sets of patient data,” says ATHENA partner Michel Delforge, professor of hematology at UZ Leuven and KU Leuven, chair of the Leuven Cancer Institute.

“ATHENA offers a platform to use real-world data in our clinical practice.” – Frank Van der Aa, UZ Leuven and KU Leuven

Building block 2 – Federated data management

Even when data is available, patient privacy poses another challenge. Privacy must be maintained and respected through all stages, from collection and integration to analysis. ATHENA’s federated data management approach provides a technical solution that effectively mitigates complex governance and legal challenges. “Today, the dominant model for data analysis remains to centralize all data,” says Yves Moreau, professor of bioinformatics at KU Leuven and ATHENA partner. “In doing so, you actually face an increased risk of privacy loss”. By using a decentralized − or federated − approach, patient data remains secure within the hospital’s IT environments. Analyses are performed at the hospital level and only aggregated insights are collected. “By conducting analyses on-site and only sharing patterns and insights, we ensure security and confidentiality while minimizing the risk of data misuse,” states Griet Verhenneman, professor privacy law at Ghent University and ATHENA partner. “Moreover, as no data transfer is needed, the legal processes involved are less complex.”

Building block 3 – Collaboration and inspiration

Federated learning allows us to access more data sources without compromising on patient privacy, thereby increasing the possibility of collaboration and cross-fertilization. This becomes particularly evident in the case of rare diseases, where the data collected within a single medical center, or even a country, are insufficient to extract valuable information. Much more can be achieved by joining forces across borders. “Collaboration is vital,” says Frank Van der Aa, professor of urology at UZ Leuven and KU Leuven, chair of the Department of Urology and ATHENA partner. “We need to work together across medical centers to benchmark our activities and learn from each other.” Delforge agrees, “The opportunity to cooperate and build future networks together is central to the ATHENA project.” ATHENA stands out as a multi-stakeholder initiative, setting it apart from various single-stakeholder projects. Project partner Ingrid Maes, Managing Director and Co-founder of Inovigate, emphasizes, “Instead of reflecting the perspective of a single entity, ATHENA integrates insights from hospitals, patient organizations, academic institutes, the life sciences industry, and pharmaceutical companies. The solutions proposed by the ATHENA project garner support from all these stakeholders.” – Ingrid Maes, Inovigate

Building block 4 – An overall data governance model

To support the reuse of real-world data, it is essential to focus on the governance aspects as well. “Next to an updated data infrastructure, a solid data governance framework needs to be decided upon and implemented,” says Maes. “Today, such a framework is missing in Belgium.” The ATHENA consortium has elaborated a governance model to ensure that patient data can be reused in a safe way. By implementing the key building blocks and recommendations devised in the ATHENA project, the government can construct appropriate legislative frameworks, allowing Belgium to reclaim its leadership in digital health and real-world data.

Read this article to learn more about how stakeholders can unite and put Belgium on the map again for real-world evidence studies.

ATHENA has delivered successful building blocks, laying the foundation for future real-world data projects. These achievements will allow us to collaboratively move forward on our path to personalized medicine and care. “Towards the future, both the bottom-up and top-down approach designed within the project need to be continued,” concludes Vannieuwenhuyse. “We need to promote local initiatives and stimulate emerging networks within and beyond hospitals. The tangible use cases and successful outcomes stemming from these initiatives will serve as evidence to encourage governing bodies to start utilizing real-world data on a larger scale. This, in turn, will pave the way towards more sustainable and structural investments in the sector, fostering the advancement towards a data-driven healthcare system.”

To celebrate the successful conclusion of Project ATHENA and recapitulate key findings, the ATHENA Symposium on Data-Driven Innovation in Personalized Medicine and Care took place on November 23, 2023. Catch the highlights from the symposium here.