To identify relevant participants for clinical trials, preconditions are formulated as long text document. With more than 30,000 active clinical trials worldwide, it is hard to keep pace with newly released trials. Thus, the identification of appropriate clinical trials is a time-consuming and manual task for clinicians today. While investigating genomic variants of a particular case, our cloud services automatically check all open clinical trials for relevant preconditions, such as age restrictions, affected genes, and pharmaceuticals. As a result, clinicians and researchers receive a list of ranked trials. Thus, only this very short list of possible trials needs to be checked for applicability. Since the extraction of entities is performed instantly, new clinical trials are directly included in the search once they are published.
The Medical Knowledge Cockpit shows how in-memory database technology combines international research data with patient specifics to find most relevant details for the treatment of individuals in course of personalized medicine in real-time. The video provides you with a brief walk through of an interactive use of the Medical Knowledge Cockpit.
Experience our Medical Knowledge Cockpit yourself and experience the advantages of in-memory technology interactively.
You might also be interested in
The following links provide additional content-related information.
- Slides of the 2019 Bio Data World congress available
- Slides of the 2019 Clinical Trials Conference available
- How AI Provides Medical Assistant System for Oncology
- In-Memory Apps for Precision Medicine (Slides)
- Analyze Genomes: In-memory Apps supporting Precision Medicine (Slides)
- 2016 Healthcare Information and Management Systems Society Conference (Slides)
- Quarks & Co on Big Data (German Video Footage)
- BioNRW: Personalisierte Medizin – im Spannungsfeld von Technik, Budget und Anwendung (Slides)
- Vernetzte Ärzte: Bessere Therapie durch neue Datenbanken
- Industrial Users Conference 2014 (Slides)
The genome browser enables you to interactively explore arbitrary genome locations on base pair level. It supports the comparison of selected genomes with each others. For example, you can compare a certain genome with a reference, e.g. diseased vs. normal tissue, or you can compare any number of genomes with each other, e.g. during patient cohort analysis.
Your exploration tasks is supported by a number of assistant information. The latest content of international annotation databases is automatically checked for relevant annotations. For example, by clicking on a single nucleotide polymorphism location the in-memory database systems scans all for relevant data, such as associated diseases, translocation partners, and known tumor types. Since links to external databases are provided, you can navigate there easily to acquire additional details information about the source data.
Furthermore, you can navigate through the genome on different levels of detail using the genome browser. For example, you can switch between base-pair-, amino-acid-, and gene-level. Thus, you can zoom-in and -out of a certain genome in accordance to your requirements.
The windowing system helps you to work on multiple cases at the same time without loosing your information. Thus, you can keep results from multiple analyses and compare them by switching between windows.
If you consider a set of genomes, you are typically interested in comparison to identify similarities and differences between individual genomes. Consider a cohort of patients where you observed that a certain therapy works for 80% of them. The question arises, how does the genotype of the 20% differ from the 80% and what do they have in common.
We support various clustering algorithms, such as k-means or hierarchical clustering, to group individual genome data. If you want to verify your hypotheses, you can set the gene and locus coordinates to build the clusters. If you do not know, why these cohorts differ, you can start an automatic discovery. Location permutations are calculated on the fly and ranked by relevance.
Thus, the cohort analysis helps to discovery new coordinates to form clusters. It supports you to obtain new insights and to build new hypotheses. The results of the clustering are visualized as interactive diagrams.
The Oncolyzer project is an interdisciplinary cooperation with Charité — Universitätsmedizin Berlin and SAP started in 2011. It combines individual competences in software engineering, IT systems, and medicine and defines a major key in our strategy to provide real-time data analysis cloud services for clinicians and researchers. The objective of the Oncolyzer project is to improve the treatment process of patients suffering from cancer diseases by leveraging optimized IT-aided software components for clinicians and researchers. As a result, doctors are enabled to select best available cancer therapies much faster by having all relevant information at hand.
The ‘heart’ of this innovation builds the in-memory computing platform, which supports the combined processing of structured and unstructured medical data in real-time. Thus, it is enables the integration of heterogeneous data sources within the clinical environment without the need for long-running and complex Extract Transform Load (ETL) processes to unify data. The organizational changes in the healthcare sector require increasing support by proper IT-aided tools and processes. Data needs to be instantaneously available at any location a doctor requires the data — even worldwide. The immense increase of knowledge about cancer requires the detailed analysis of biological and genetic details acquired during diagnosis of cancer cells to address only harmful cells as the target of future treatments while keeping side effects at a minimum. Until recently, common therapies that were not individually targeted were applied during the treatment of cancer. Meanwhile, treatments for specific genetic mutations are available, which enable treatment of cancer based on individual genetic dispositions. Nowadays, data processing and analysis becomes a time-consuming challenge due to improved and more detailed diagnostic approaches, such as next-generation sequencing of tumor DNA.
In the near future, the tumor’s DNA of all cancer patients will be sequenced to support individualized patient-specific cancer therapies, which result in diagnostic medical data in amount of multiple terabytes. The analysis of these data required optimized software tools that enable graphical exploration of data, their real-time analysis, and the identification of therapy-relevant details to support clinical decision taking.
Features of the Hana Oncolyzer iPad Application
The Oncolyzer iPad application provides clinicians and researchers access to relevant patient data while in the secured network of the clinic’s campus. The in-memory technology builds the backend of the application performing relevant data processing and analyses. Selected features of the Oncolyzer application are described in the following.
Combined Search in Structured and Unstructured Data
The Oncolyzer combines data from various data sources — structured and unstructured.
Structured data are, for example, biopsy results, size of tumor or tissue regions, blood concentration, etc. They are stored in a relational database format and can be accessed via defined attributes. Unstructured data are, for example, text documents, diagnosis, notes, etc. They are stored in text file and they neither consist of a predefined structure nor use standardized text paragraphs. As a result, unstructured data has the following drawback: typos, usage of pseudonyms, abbreviations, etc. However, the majority of clinical medical data is unstructured. Thus, it is important to analyze them in a systematic way and to extract relevant details in real-time. Further details about combined processing of structured and unstructured data can be found on the corresponding feature page.
Visualization of Patient Details for Personalized Medicine
All patient related information need to be available for decision taking by medical doctors. Individual specifics of patients need to be analyzed and evaluated to for personalized medicine. The Oncolyzer combines current as well as historic data of a selected patient on a single screen and performs automatically analysis of the available data, e.g. to highlight patient specifics compared to patients with similar diagnosis or anamnesis. Thus, characteristics and important differences are highlighted, i.e. medical doctors can use these additional information as indicator assessment of the individual reason for the disease and impact of the selected treatment. All information are visualized on an interactive time line, which enables clinicians to move back and forward through the patient’s anamnesis while having access to all relevant data with a single click.
Analytical Exploration of Patient Cohorts
Analysis data of all patients or a patient cohort is a time-consuming and often manual task. The Oncolyzer app enables analysis of patient data on mobile devices, i.e. there is no longer a need for a desktop PC to perform the analysis. Furthermore, freely definable filters can be applied to explore patient details. Thus, the Oncolyzer app enables identification of individual patients, e.g. for participation in specific clinical studies to provide individualized treatment. The use of always up-to-date data bridges information gaps and supports fast analysis and evaluation of patient cohorts. Further details about analysis of patient cohorts can be found in the corresponding app description.
You might also be interested in
The following links provide additional content-related information.
- F.A.Z.: Neue Dimension für Datenanalysen: Aus Wochen werden Sekunden!
- Berliner Kurier: Wettlauf gegen den Krebs
- Oncolyzer on SAP TV (German Footage)
- Oncolyzer – Medical Records on a Tablet PC
- 2012 Innovation Award of the German Capital Region
- Oncolyzer on SAP TV (English Footage)
- Presentation of Oncolyzer to German Chancellor Dr. Angela Merkel and Brazilian President Dilma Rousseff (German Video)
- Prof. Dr. Hasso Plattner on How Oncolyzer Improves Cancer Treatment (English Video)