Architects of a new IT research landscape

© dkfz.de

Medical data generated during diagnosis or treatment contains valuable information for researchers and physicians. Frank Ückert and his team develop concepts and IT tools that will make this data resource usable.

A computer and sometimes a whiteboard - the employees around Frank Ückert do not need much more for their daily work. The more multifaceted are their tasks. Persons who exclusively program are here in the minority. Most of the staff of the Department of Medical Informatics in Translational Oncology develop concepts - they create project plans, develop strategies and communicate with cooperation partners. Their goal: data from the laboratory and clinic will form the basis for new therapies or diagnostic procedures.

Although the notion of medical informatics has existed since the 1970s, the Internet and even more powerful computers fueled this discipline. Many experts see this as a great opportunity: The extensive medical data that is already available today could help to find a customized personalized treatment for each patient in the future.

Representative for the development of this specialization is also the department itself. After stations in Münster, Erlangen and Mainz, Frank Ückert joined the DKFZ in 2016. At the beginning, his team consisted of five members. It has now grown to around 40 scientists - a comparatively rapid growth for a research group. The majority of the group does not consist of computer scientists but of physicians and biologists. Also Ückert himself studied medicine. At the same time he completed a second degree, that of mathematics. But what do the medical informatics do in detail?

© dkfz.de

Creating added value

"We first think about which data are interesting and how it can be approached," explains Ückert, who helps research institutes and clinics save and integrate their data with existing ones - for example genome analyzes with radiological images or other clinical data. "It's all about adding value from the bundled information." Ultimately, the work of IT specialists should help physicians make the most accurate diagnosis possible and select the appropriate therapy. Ückert sees his team on a good path: "In some cases, we are already so far that we can support the doctors in their decision-making."

But even the merging of the data is a challenge: even unstructured data, such as medical reports, should be included in the databases and later be searched and filtered together with laboratory values ​​and other information. Alexander Knurr would like to do just that. Within the department, he leads the DataThereHouse working group, whose goal is to establish a data infrastructure that stores the clinical and scientific data of the cancer patients of the National Center for Tumor Diseases (NCT) Heidelberg and the research data of the DKFZ. Even complex queries should be able to handle the database at high speed. First of all the scientists have to prepare the data in a way that is compatible with each other, because often different standards are documented, so besides programming skills, it is important for the employees Knowing the importance of the data. Knurr emphasizes we need to understand the complexities of oncology. In order to assist a doctor later in his daily work, the team must ensure that the data are presented comprehensibly. As a result, researchers often play a mediating role between the people involved, such as doctors, researchers and software developers.

© dkfz.de

Mediation role

One such facilitator is the biotechnologist and computer scientist Sophia Stahl-Toyota. "My job is primarily to bring together the different professions. They want to see the added value of bringing the data together, "says the researcher, who benefits from the fact that she understands both the computer scientists' and scientists' perspectives in the laboratory." Stahl-Toyota works the AG Collaborative Research led by Martin Lablans and Tanja Höpker, which aims to establish networks where data are not simply collected - they should also bring tangible benefits."In particular, we want to break the boundaries between research and treatment - to both translate research findings into clinical application as well as to bring clinical results into research"says Stahl-Toyota.

The working group now cooperates with partners throughout Germany. "It is often no longer enough to analyze data from just one location," emphasizes Ückert, because the molecular causes of a cancer can vary greatly from patient to patient, and each case must be considered individually. To be able to form sufficiently large groups in which the patients have the same or at least similar characteristics, the researchers must be able to select from a correspondingly large number of cases. The rarer a particular type of cancer occurs, the more difficult it is to ensure.

"Some scientists, however, regard the data as a kind of treasure that they are reluctant to share with others," says Ückert. This it is understandable: After all, everyone has to reap the laurels of his work at some point. So the team has a lot of convincing: "To do this, one has to explain to those responsible in research institutes and clinics: The value of this treasure does not decrease when it is shared with others, it even increases," explains Ückert. In addition, one must create confidence. The group succeeds by guaranteeing the respective locations the control over their data. You can always decide for yourself which information you really want to share and which you do not.

© dkfz.de

The bridgehead

For this purpose, the team has developed the so-called bridgehead. This is basically a computer with specially designed software that resides at every network location. The computers stores and standardize the data and brings it together - so the people responsible give their consent. In addition, a specially designed data protection concept ensures that the procedure complies with the strict German data protection requirements.

Meanwhile, the group operates the necessary computer technology for very different cooperations. Outstanding is the German Cancer Consortium (DKTK), in which the DKFZ cooperates with research institutions and clinics in eight locations. Other partners include biobank networks or the network of German oncological centers sponsored by the German Cancer Aid. The department also cooperates with the National Network Genomic Medicine (nNGM). The nNGM aims to ensure that all patients with advanced lung cancer in Germany receive access to molecular diagnostics and new therapies. "Through these networks, we basically get the opportunity to access larger databases," emphasizes Ückert, but he and his colleagues are still looking for new cooperation partners: "We try again and again to encourage all those who work with data collections, to contact us. "

Because the scope and quality of the data decide on their benefits. In particular, the use of artificial intelligence (AI) makes sense only with a sufficiently large amount of harmonized data. Self-learning algorithms then examine the huge amounts of data for patterns and reveal relationships that would not have been found using conventional methods. "In most cases, the amount of information is currently too low to be meaningfully analyzed by AI." However, no one in his team doubts that this technology will rapidly advance the development of medical informatics. Nobody should worry about a job in the future. (Jd)

DataThereHouse - make research data accessible quickly, easily and securely

© dkfz.de

The analysis of medical datasets helps researchers to develop new diagnostic, therapeutic and prevention methods. Often this project fails due to the lack of IT infrastructure. The DataThereHouse project at NCT Heidelberg and DKFZ is establishing a digital platform that ensures the integration of the data and facilitates their evaluation.

"The data are not somewhere, but they are right here - so there." On this original idea, remembers Professor Frank Ückert, based the naming of the DataThereHouse. The idea of ​​such a system is to actually integrate and consolidate medical and scientific data in one place, reports Ückert, who heads the department "Medical Informatics in Translational Oncology (MITRO)" at the German Cancer Research Center (DKFZ). After all, the manifold medical data that are generated every day is a valuable asset for physicians and researchers, but it is difficult to access without appropriate tools. Medical informatics therefore develop appropriate analysis methods. The insights gained can be used to develop new diagnostic procedures and innovative therapies.

However, many clinics and facilities on the Heidelberg campus have their own documentation systems and data storage. "The data is usually stored in very different forms and formats and the access to the systems is not centrally regulated," describes Ückert the problem. "Consequently, we must first harmonize data," as he calls it. For this purpose, researchers from the DKFZ and the NCT Heidelberg together with developers of the software company SAP started in 2013 to build a first analysis platform using the so-called Hana database technology from SAP as the foundation stone of the DataThereHouse. In 2016, the MITRO department took over the project and continued to develop it. With the inclusion of new and established tools, this resulted in a comprehensive platform for the support of clinical and scientific projects.
In addition to clinical data from patients at the NCT Heidelberg and information on biological samples from the NCT Tissue Bank, the platform now also integrates the molecular findings generated in the DKFZ and provides information on oncological relevant mutations in the genetic profiles of the tumors. Sharing this information is a crucial step toward personalized oncology and is used in studies such as the Molecularly Aided Stratification for Tumor Eradication Research Program, or MASTER for short. As a result, physicians will increasingly be able to make therapy decisions based on the multiple mutations of a single tumor or patient. However, the number of patients with comparable mutations in a hospital is too small to obtain reliable statements on the success of treatment and thereby secure insights for the treatment of future patients.

Heidelberg as interface to international partner

Therefore, according to Alexander Knurr, employee of Ückert and leader of the group DataThereHouse, it is important to bring the clinical and scientific world outside of Heidelberg on board for joint projects and clinical studies.

Knurr cites the Athens Comprehensive Cancer Center (ACCC) as an example of a recent international collaboration. The DKFZ is supporting five hospitals in Athens to build a new center based on the model of the NCT. A central aspect of this collaboration is the merging of both clinical data and bio-sample data from both sites into the DataThereHouse. This is urgently needed, says Knurr, in order to broaden the database. It will begin in 2019 with data and bioprobes from the areas of colorectal carcinoma, multiple myeloma and pediatric oncology.

Challenges for the DataThereHouse


In order to prepare the platform for new challenges in personalized oncology, which was previously used predominantly in Heidelberg, and the associated cross-location cooperation, the existing platform has been adapted, Knurr reports. For example, he and his colleagues needed to expand and integrate systems for documenting clinical and scientific-oncological traits and interviewing patients in the course of treatment for non-German-speaking users. In order to ensure the data protection compliant processing of the data, the team had to further develop the pseudonymization of the data and adapt the data protection concept in view of the EU data protection regulation.

In addition, the partners would often deal with other data and issues, explains Knurr. In order to support them, the computer scientists, physicians and biologists of MITRO have to intensify their efforts

Support of physicians and researchers

The DataThereHouse enables physicians to examine the data with the help of intuitive analysis tools and gain a quick overview of a patient's collected information: blood samples, gene analysis, physician's letters, pathology findings, radiological examinations, genetic alterations, and so on. Comparison with information available on the platform on similar patients, current research, or clinical trials will help physicians diagnose, develop personalized therapy, or search for appropriate clinical trials for which a particular patient would be appropriate.

In addition, scientists in the DataThereHouse can discover new diagnostic and therapeutic options.
For this purpose, the platform offers them the following possibilities:

· Identifying biomaterials from patients with medically interesting profiles to investigate the efficacy of potential drugs in laboratory and subsequent clinical trials.

· Analyzing the large volumes of data generated by genetic analysis with innovative methods and finding new approaches to innovative therapies.

· Automating the detection of tumors or metastases in radiological images and thus helping radiologists and oncologists to make therapeutic decisions.

Overall, the importance of such powerful IT infrastructures in medicine will continue to increase. On the one hand, this is due to the ever better methods of data analysis and, on the other hand, to the rapid increase in the data amount. "All these medical data are in total much more valuable than their individual parts," Ückert sums up. The DataThereHouse increases the value of the data and becomes a tool for oncologists to get the most out of the patient. (Jd)

to top