To make progress in the study of blood cancers, there is an urgent need to have access to a massive amount of high-quality, high-content data. To address this need and improve patient care, public and private organizations across Europe are working together in a European consortium. HARMONY collects and harmonizes electronic health records on diagnosis, treatment, and outcomes of patients with blood cancers in order to identify new biomarkers, improve evidence-based treatment decisions, and thus improve patients' outcomes.
Using machine learning techniques and advanced algorithms, the HARMONY Alliance is orchestrating the consolidation of all databases into one common platform. Launched in January 2017, HARMONY is a public-private partnership currently comprised of 53 partners, 41 associated members, and 7 patient organizations from 17 European countries.
“Until now, we have been focused in specific projects collecting retrospective data. What we are working on now is downloading as many data sets as possible for the different diseases hoping to ‘tear down the walls’ separating national registries in Europe,” said Jesús María Hernández Rivas, MD, the coordinator of HARMONY.
Hernández Rivas is Professor in Hematology at the Medical School of the University of Salamanca, Spain, and Senior Physician of the Department of Hematology in the University Hospital of Salamanca.
“This project started as an idea about 5 years ago. At that time, a group of hematologists from different European countries decided to apply for a European Union grant from the Innovative Medicine Initiative conducted in collaboration with the pharmaceutical industry,” said Hernández Rivas. “We had to do a lot of negotiations while building up the consortium and then trying to include all of the relevant institutions in Europe. I was leading these initiatives, and this is how I became the coordinator of this large consortium.”
Hernández Rivas said the project has brought together groups from a broad spectrum of disciplines, including academic institutions, national clinical disease networks, patient advocacy groups, hospitals, regulatory agencies, health technology assessment bodies, technology specialists, and pharmaceutical companies.
HARMONY's partners from industry provide access to clinical trials data, showing how drugs work in controlled circumstances. Public registries, however, collect data on treatment and outcomes in the real world. By combining all that information, HARMONY is positioned to learn from best practices across Europe and discover underlying patterns that determine better treatment outcomes.
Four Big Questions
In 3 years, 45,000 data sets from patients with blood cancer have been identified. The goal of the data collection is to enable researchers to answer four questions:
- How can we diagnose patients faster and with greater precision?
- What are the best practices that could help doctors in making better treatment decisions?
- How can we tackle the unmet needs of patients?
- How can biomarkers and surrogate parameters help accelerate new drug development?
“I say as a researcher hematologist that we can never be comfortable or happy with our results. I will be happy the day all the hematological patients will be cured without any kind of secondary effects. It is our mission to provide new hopes, new tools, new ideas, to better treat the patients,” said Hernández Rivas.
Hernández and the collaborating institutions believe these key outstanding questions can only be answered by studying large numbers of patients. The HARMONY Alliance has developed a blended European clinical data platform, assembling data from thousands of European patients and aiming to expand the data pool to at least 100,000 patients during its lifetime.
This will enable the researchers to characterize the molecular landscape of the six main hematological malignancies: acute lymphoblastic leukemia (ALL), acute myeloid leukemia (AML), chronic lymphocytic leukemia (CLL), multiple myeloma (MM), myelodysplastic syndromes (MDS), non-Hodgkin lymphoma (NHL), and additionally, the special case of pediatric hematologic malignancies.
The researchers hope the platform will uncover a deeper understanding of the pathophysiology of these diseases, identify novel drug targets, and reliably predict disease course and drug response for subgroups of patients.
“Hematological malignancies are rare diseases and for a single institution, it is quite difficult to collect, let's say 500 samples, with the same diagnosis. For instance, if you are going to study patients with myeloma, it is quite difficult to collect this number of patients in a short period of time. You would have to wait 5 years and in 5 years everything, including the therapies, may have changed,” said Hernández Rivas.
“In HARMONY, we are focusing on most of the hematological malignancies, as previously mentioned, but I hope that by the end of the year, we will be able to cover all the different hematological malignancies: AML, CLL, ALL, MDS, NHL, and MM. Working groups focus on those individual indications. These are very different types of cancers, each one has its own special therapeutic regimes and their own pressing problems.”
The data revealed may also assist researchers in selecting outcomes for future trials, thereby promoting harmonization of blood cancer clinical trials and improving clinical management of the diseases. Researchers also hope that, ultimately, the project will result in more personalized medicine tools to rapidly select the most promising treatment strategy for each patient.
“The data that we already collected and the data that we are planning to collect over the next years will enable us for the first time to put together the molecular data, genomic data, and the clinical data. The only way in which we can achieve the challenge of personalized medicine or precision medicine is by collecting large amounts of relevant data, good quality data, to be analyzed and to start to take the clinical decisions based on as much evidence as we can have from previous patients.”
Big Data Complexities
In 2020, there is a focus on gathering real-world data in the medical community. Clinical trials provide critical data but don't always give a full picture because only a minority of patients are enrolled in clinical trials. Analyzing the multitude of data can help researchers capture the evolution of diseases in real time.
One challenge is data harmonization, or in other words, data from different databases must be transformed in the same format before it can be analyzed. The enormous size of big data means that new analytical and statistical tools need to be developed. According to Hernández Rivas, getting these fundamentals right was the main priority during HARMONY's initial phase.
“First of all, we have to use new tools for screening and for the analysis of the data. The big data analytical techniques have not yet been widely applied in health research. I am starting to see some papers in JAMA about the use of artificial intelligence, machine learning, big data approaches; however, big data so far has been largely and widely used for economics, for weather prediction, for traffic monitoring and so on—but not frequently in health research so far,” said Hernández Rivas.
“We will need to develop new algorithms or refine algorithms to be applied in health research. We must adapt the big data concept to health. We will run new algorithms and developing tools that we will be able to help the hematologists to establish better prognosis for each patient suffering from these diseases.”
Other complexities in bringing together so many institutions and researchers across Europe include patient confidentiality and data protection. Hernández Rivas said these concerns were addressed at the start of the project.
“First, it is a complex project, involving a lot of partners. Second, we have built a large repository, with many data sets coming from both the public and the private sector going together to the same platform. Now we are starting to extract the results. So, all the problems related to confidentiality, to data protection, have been resolved in the first year.”
The database must also comply with European and national regulations on data exchange, privacy, and ethical rules. According to HARMONY, the consortium requires the use of standard language in the patient's informed consent form, which provides that explicit consent be given for the further use of data. Hernández Rivas said the consortium also actively works with advocacy agencies to ensure the data collection is fair to the patients.
“In the project, there is an active participation from patient advocacy groups. There is also an important participation from regulatory agencies. We are in contact with the European Medicines Agency and national country level-based regulatory agencies in Europe are HARMONY partners,” said Hernández Rivas.
“Furthermore, HARMONY has the active participation of several Health Technology Assessment bodies. In addition to collecting data, HARMONY initiates discussions in an open frame from the different perspectives on how we can improve health for the patients and how precision medicine really can contribute it to the sustainability of the public health systems in the different countries. More organizations are welcome to provide insights to the project to improve the quality of life for patients.”
Hernández Rivas said the project always aimed to share data and open borders and hopes the impact will spread beyond Europe and that other countries and institutions outside the EU can participate.
“I think it's important to state that HARMONY is a European project, but also an open project with a global vision. We have received some feedback from other continents to be in touch and to start a collaboration—such as the United States and Australia,” said Hernández Rivas. “Our idea is to provide the proof of concept that this initiative is really important for the patients. And it's important for the clinicians who treat the patients.”
Currently, the first sets of data are being uploaded onto the platform and the HARMONY community is working to make their databases available. In the coming months and years, the next steps are to reach agreement on the set of outcomes for each disease, finish the data collection and harmonization process, and start analysis.
“Our objective in HARMONY is to share data. By sharing more and more high-quality data, we will improve the analytical power and we will detect new genes impacting the prognosis of hematological diseases. In that way, we will arrive at better diagnosis and improve the survival of our patients,” said Hernández Rivas. “We have demonstrated that it is possible to do it. Some people told me it is not possible to work with data from the private and the public sector together. It is possible to do it. HARMONY is doing it.”
Sarah LaCorte is associate editor.