DataNexus connects clinical and genomic data

DataNexus is a project initiated by the Netherlands Cancer Institute–Antoni van Leeuwenhoek (NKI-AvL) to improve data availability by integrating clinical and genomic patient data in collaboration with Hartwig Medical Foundation. Through recurring requests for clinical data, a continuous data stream is created. Linking this to genomic data provides an important boost to research quality and personalized treatment. DataNexus thus forms a powerful hub for quality control, international collaboration, and faster implementation of new insights in oncology.

By Frank van Wijck, science journalist

With the DataNexus project, Antoni van Leeuwenhoek is taking an important step for both itself and Hartwig Medical Foundation. “In addition to genomic data, patients’ clinical data are also essential,” says Sophie Roerink, bioinformatician at Hartwig Medical Foundation. “Clinical data are needed to determine which treatment is appropriate for which patient. But we only received limited clinical data—mainly as part of clinical trials, and far less for routine care.”

For that reason, Hartwig was eager to participate in the AvLinitiated DataNexus project, as AvL had clearly expressed the need for improvements in datamanagement. “We had several reasons for this,” says pathologist Kim Monkhorst.

“We noticed how difficult it was to keep clinical data up to date. For a long time, updates were sent by email, which was very impractical. We need a constant data stream that keeps information such as subsequent therapy lines and overall survival current. That also means we don’t want to repeatedly request clinical data from Hartwig—we want direct access, so we can immediately add additional clinical information. We also want access to the NGS-panel data generated by Hartwig, which until now was only stored for five years. And we want to be able to share all data—not just wholegenome sequencing data—with international research groups we collaborate with.”

Recurring data requests
DataNexus was designed to make all of this possible.

“The idea is to place a kind of questionnaire in front of the database, containing clinical variables that we repeatedly retrieve from hospital information systems. This ensures that the data are updated whenever the source data change—such as a diagnosis—or when new information is added, like a treatment,”: Monkhorst

“Think of patient age, overall survival, tumor location, whether it’s a metastasis, and all the other variables you need for research. We linked this questionnaire to data from our own NGS-panels, WGS, and other sources.”

“It’s a hub where all data come together,” he continues. The questionnaire is requested on a recurring basis, and the data are provided to Hartwig monthly. “This creates a continuous update,” says Monkhorst. “It also immediately provides an option for quality control. If something in the data is incorrect, this approach gives us a feedback loop that allows us to quickly determine whether we selected the wrong source or need to adjust the query.”

A strong collaboration partner
Hartwig was eager to help bring this AvL initiative to life because of the special relationship between the two organizations. “AvL is the hospital in the Netherlands where WGS is most deeply embedded in routine care,” Roerink explains. “That makes AvL our largest collaboration partner. And what this project enables aligns perfectly with our philosophy: collecting data from today’s patients to better inform tomorrow’s patients.”

For Hartwig, the added value of DataNexus is that it provides highquality clinical data, Monkhorst says. “For us, the major benefit is that we can now share all our genomic data—combined in a structured way with all clinical data—with third parties, and use it ourselves for diagnostics, quality evaluation, validation, implementation of new tests, and research. We gain access to more and better data with strong justification. In diagnostics, we can use the raw data to answer the questions we need at that stage. It also accelerates the implementation of new tests. More broadly, it’s valuable for care evaluation. Large datasets allow us to see whether subgroups of patients respond better—or worse—to certain treatments. All research made possible by this enriched data will ultimately contribute to better patient care.”

Next steps
With the introduction of DataNexus, AvL has taken an important step in datamanagement, but certainly not the last. “We still have several ambitions for the next phase,” Monkhorst says. “First, we want to use the database to rapidly improve data quality. Once we’ve achieved that, we want to collaborate nationally. We’re already in discussions with several university medical centers, because this model could be very useful for them as well.”

Hartwig also sees room for next steps. “Further structuring of the enriched data so that it becomes comparable to the data we already have will make it possible to conduct even more targeted research using even more complete and richer clinical data,” Roerink says. She also gives an example. “We work closely with a team in Barcelona that has developed a system to systematically identify non‑responders. When we run an analysis based on our database, we now see that for certain biomarkers the number of patients becomes too small to draw statistically significant conclusions.  

So if we can expand the cohort based on genomic and clinical data, we can more precisely determine which patients will or will not benefit from a particular treatment. If the signal from such a biomarker is sufficiently robust, it can even make its way into a clinical guideline. A lot of research is already being done on response biomarkers. But non‑responders are just as interesting. And because the pharmaceutical industry is less interested in this group, we clearly see a role for ourselves here. If treatments can be avoided for patients who will not benefit from them, it spares them the burden of that treatment and saves healthcare costs for society.” 

You read an article in the category Hartwig Medical Database. You may also be interested in Learning healthcare system, Participating hospitals, Re-use data or Whole genome sequencing.
All news