Datascience & AI in der personalisierten Medizin
Entwicklung präziser Modelle und Unterstützung klinischer
Entscheidungen für eine verbesserte medizinische Versorgung
Vorstellung des
Forschungsbereiches
Der Bereich Datascience & AI des MOLIT Instituts forscht in den Themenbereichen rund um die personalisierte Medizin. Schwerpunkte stellen dabei die Modellentwicklung sowie die klinische Entscheidungsunterstützung dar. Dabei soll es dank einem Fokus auf Interoperabilität möglich sein die Entwicklungen möglichst einfach in bestehende Systeme zu integrieren.
Als Teil der AI Community sollen Daten für AI-Training bereitgestellt und Modellsysteme abgeleitet werden, sodass eine digitale Entscheidungsunterstützung in der translationalen Forschung zur Klinikrealität werden kann. Mit dem Einsatz aktuellen Data Science Methoden und der Verbindung zur Biologie wird Evidenz- und Wissensgenerierung so möglich.
Forschungsteam
Chantal Bachschmid
Managerin Lion-App
Felix Edel
Software Entwickler
Kevin Kaufmes
Software Entwickler
Georg Mathes
Software Entwickler
Dr. Stefan Sigle
Bereichsleitung
Valeriya Vishnevskaya
Software Entwicklerin
Publikationen
Auszug unserer
Publikationen
3757376
W54IEIEG
1
apa
3
date
desc
21
https://www.molit.eu/wp-content/plugins/zotpress/
%7B%22status%22%3A%22success%22%2C%22updateneeded%22%3Afalse%2C%22instance%22%3Afalse%2C%22meta%22%3A%7B%22request_last%22%3A0%2C%22request_next%22%3A0%2C%22used_cache%22%3Atrue%7D%2C%22data%22%3A%5B%7B%22key%22%3A%22P4SIP576%22%2C%22library%22%3A%7B%22id%22%3A3757376%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22Mathes%20et%20al.%22%2C%22parsedDate%22%3A%222026-09-17%22%2C%22numChildren%22%3A0%7D%2C%22bib%22%3A%22%26lt%3Bdiv%20class%3D%26quot%3Bcsl-bib-body%26quot%3B%20style%3D%26quot%3Bline-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%26quot%3B%26gt%3B%5Cn%20%20%26lt%3Bdiv%20class%3D%26quot%3Bcsl-entry%26quot%3B%26gt%3BMathes%2C%20G.%2C%20Berger%2C%20S.%2C%20%26amp%3B%20Sigle%2C%20S.%20%282026%29.%20Automated%20Extraction%20of%20Genetic%20Eligibility%20Criteria%20from%20Clinical%20Trial%20Records%20Using%20LLMs%20%26%23x2013%3B%20A%20Technical%20Case%20Report.%20In%20U.%20Sax%2C%20T.%20Ganslandt%2C%20R.%20Breitschwerdt%2C%20K.%20Jung%2C%20J.%20K%26%23xF6%3Bnig%2C%20%26amp%3B%20A.%20Gro%26%23xDF%3Bhennig%20%28Eds.%29%2C%20%26lt%3Bi%26gt%3BStudies%20in%20Health%20Technology%20and%20Informatics%26lt%3B%5C%2Fi%26gt%3B.%20IOS%20Press.%20%26lt%3Ba%20class%3D%26%23039%3Bzp-DOIURL%26%23039%3B%20href%3D%26%23039%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.3233%5C%2FSHTI260998%26%23039%3B%26gt%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.3233%5C%2FSHTI260998%26lt%3B%5C%2Fa%26gt%3B%26lt%3B%5C%2Fdiv%26gt%3B%5Cn%26lt%3B%5C%2Fdiv%26gt%3B%22%2C%22data%22%3A%7B%22itemType%22%3A%22bookSection%22%2C%22title%22%3A%22Automated%20Extraction%20of%20Genetic%20Eligibility%20Criteria%20from%20Clinical%20Trial%20Records%20Using%20LLMs%20%5Cu2013%20A%20Technical%20Case%20Report%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22editor%22%2C%22firstName%22%3A%22Ulrich%22%2C%22lastName%22%3A%22Sax%22%7D%2C%7B%22creatorType%22%3A%22editor%22%2C%22firstName%22%3A%22Thomas%22%2C%22lastName%22%3A%22Ganslandt%22%7D%2C%7B%22creatorType%22%3A%22editor%22%2C%22firstName%22%3A%22R%5Cu00fcdiger%22%2C%22lastName%22%3A%22Breitschwerdt%22%7D%2C%7B%22creatorType%22%3A%22editor%22%2C%22firstName%22%3A%22Klaus%22%2C%22lastName%22%3A%22Jung%22%7D%2C%7B%22creatorType%22%3A%22editor%22%2C%22firstName%22%3A%22Jochem%22%2C%22lastName%22%3A%22K%5Cu00f6nig%22%7D%2C%7B%22creatorType%22%3A%22editor%22%2C%22firstName%22%3A%22Anika%22%2C%22lastName%22%3A%22Gro%5Cu00dfhennig%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Georg%22%2C%22lastName%22%3A%22Mathes%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Stephanie%22%2C%22lastName%22%3A%22Berger%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Stefan%22%2C%22lastName%22%3A%22Sigle%22%7D%5D%2C%22abstractNote%22%3A%22Introduction%3A%20Accurate%20interpretation%20of%20clinical%20trial%20eligibility%20criteria%20is%20essential%20for%20applications%20such%20as%20patient-trial%20matching%20and%20clinical%20decision%20support%2C%20particularly%20in%20precision%20oncology.%20However%2C%20relevant%20information%2C%20including%20genetic%20mutation%20requirements%2C%20is%20typically%20embedded%20in%20unstructured%20text%20within%20trial%20registries%20such%20as%20ClinicalTrials.gov%2C%20limiting%20accessibility%20for%20automated%20processing.%20Methods%3A%20This%20paper%20reports%20on%20development%2C%20integration%20and%20evaluation%20of%20a%20system%20for%20extracting%20and%20structuring%20mutational%20eligibility%20criteria%20from%20trial%20records%2C%20focusing%20on%20identifying%20mutated%20genes%2C%20distinguishing%20inclusion%20and%20exclusion%20criteria%2C%20and%20assigning%20them%20to%20individual%20study%20arms.%20To%20address%20challenges%20like%20ambiguous%20abbreviations%20and%20context-dependent%20meaning%2C%20we%20combine%20large%20language%20models%20%28LLMs%29%20for%20context-aware%20extraction%20with%20rule-based%20validation%20against%20HUGO%20Gene%20Nomenclature.%20The%20system%20was%20implemented%20using%20local%20LLMs%20and%20integrated%20into%20the%20Community%20Annotated%20Trial%20Search%20%28CATS%29%20platform.%20Results%3A%20Applied%20to%204%2C918%20clinical%20trials%2C%20the%20system%20generated%20structured%20representations%20of%20genetic%20eligibility%20criteria%20for%201%2C010%20studies.%20Expert%20review%20of%2042%20trials%20showed%20that%2088.1%25%20of%20studies%20were%20correctly%20annotated%2C%20with%20a%20precision%20of%2080%25%20at%20the%20level%20of%20individual%20eligibility%20criteria.%20Failures%20were%20mainly%20due%20to%20hallucinated%20genes%20and%20misinterpreted%20abbreviations%2C%20highlighting%20challenges%20in%20biomedical%20text%20processing.%20Discussion%3A%20The%20findings%20indicate%20that%20LLM-based%20extraction%20is%20a%20promising%20approach%20for%20structuring%20complex%20eligibility%20criteria%2C%20particularly%20when%20combined%20with%20strategies%20to%20improve%20precision.%20The%20integration%20into%20an%20operational%20system%20demonstrates%20practical%20feasibility%2C%20while%20the%20observed%20limitations%20emphasize%20the%20need%20for%20careful%20dataset%20design%2C%20error%20mitigation%20strategies%2C%20and%20continued%20refinement%20to%20achieve%20reliable%20automation%20in%20clinical%20applications.%22%2C%22bookTitle%22%3A%22Studies%20in%20Health%20Technology%20and%20Informatics%22%2C%22date%22%3A%222026-09-17%22%2C%22originalDate%22%3A%22%22%2C%22originalPublisher%22%3A%22%22%2C%22originalPlace%22%3A%22%22%2C%22format%22%3A%22%22%2C%22ISBN%22%3A%22978-1-64368-687-5%22%2C%22DOI%22%3A%2210.3233%5C%2FSHTI260998%22%2C%22citationKey%22%3A%22%22%2C%22url%22%3A%22https%3A%5C%2F%5C%2Febooks.iospress.nl%5C%2Fdoi%5C%2F10.3233%5C%2FSHTI260998%22%2C%22ISSN%22%3A%22%22%2C%22language%22%3A%22%22%2C%22collections%22%3A%5B%22W54IEIEG%22%5D%2C%22dateModified%22%3A%222026-09-29T12%3A37%3A45Z%22%7D%7D%2C%7B%22key%22%3A%22ND27Q3M3%22%2C%22library%22%3A%7B%22id%22%3A3757376%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22Pelka%20et%20al.%22%2C%22parsedDate%22%3A%222026%22%2C%22numChildren%22%3A0%7D%2C%22bib%22%3A%22%26lt%3Bdiv%20class%3D%26quot%3Bcsl-bib-body%26quot%3B%20style%3D%26quot%3Bline-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%26quot%3B%26gt%3B%5Cn%20%20%26lt%3Bdiv%20class%3D%26quot%3Bcsl-entry%26quot%3B%26gt%3BPelka%2C%20O.%2C%20Eil%2C%20J.%2C%20Manjunatha%2C%20K.%2C%20Singh%2C%20N.%2C%20Omeirat%2C%20J.%2C%20Sigle%2C%20S.%2C%20Mathes%2C%20G.%2C%20Schweizer%2C%20S.-T.%2C%20Stump%2C%20S.-R.%2C%20Girdziunaite%2C%20G.%2C%20%26amp%3B%20Nensa%2C%20F.%20%282026%29.%20Interoperable%20clinical%20data%20integration%20for%20distributed%20medical%20AI%3A%20The%20Open%20Medical%20Inference%20Gateway.%20%26lt%3Bi%26gt%3BGMS%20Medizinische%20Informatik%2C%20Biometrie%20Und%20Epidemiologie%26lt%3B%5C%2Fi%26gt%3B%2C%20%26lt%3Bi%26gt%3B22%26lt%3B%5C%2Fi%26gt%3B.%20%26lt%3Ba%20class%3D%26%23039%3Bzp-DOIURL%26%23039%3B%20href%3D%26%23039%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.3205%5C%2FMIBE000309%26%23039%3B%26gt%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.3205%5C%2FMIBE000309%26lt%3B%5C%2Fa%26gt%3B%26lt%3B%5C%2Fdiv%26gt%3B%5Cn%26lt%3B%5C%2Fdiv%26gt%3B%22%2C%22data%22%3A%7B%22itemType%22%3A%22journalArticle%22%2C%22title%22%3A%22Interoperable%20clinical%20data%20integration%20for%20distributed%20medical%20AI%3A%20The%20Open%20Medical%20Inference%20Gateway%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Obioma%22%2C%22lastName%22%3A%22Pelka%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Jan%22%2C%22lastName%22%3A%22Eil%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Kaushik%22%2C%22lastName%22%3A%22Manjunatha%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Nimesh%22%2C%22lastName%22%3A%22Singh%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Jazia%22%2C%22lastName%22%3A%22Omeirat%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Stefan%22%2C%22lastName%22%3A%22Sigle%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Georg%22%2C%22lastName%22%3A%22Mathes%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Simon-Tobias%22%2C%22lastName%22%3A%22Schweizer%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Shura-Roman%22%2C%22lastName%22%3A%22Stump%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Gerda%22%2C%22lastName%22%3A%22Girdziunaite%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Felix%22%2C%22lastName%22%3A%22Nensa%22%7D%5D%2C%22abstractNote%22%3A%22Hintergrund%3A%20Die%20Integration%20von%20KI-Inferenzdiensten%20in%20klinische%20Arbeitsabl%5Cu00e4ufe%20bleibt%20eine%20anhaltende%20Herausforderung%2C%20vor%20allem%20weil%20Inferenz-Backends%20und%20Standards%20f%5Cu00fcr%20Gesundheitsdaten%20von%20grundlegend%20unterschiedlichen%20Annahmen%20ausgehen.%20Der%20EU%20AI%20Act%20fordert%20zudem%20Transparenz%20und%20Nachvollziehbarkeit%20f%5Cu00fcr%20Hochrisiko-KI-Systeme%20im%20Gesundheitswesen.%5Cn%20%5Cn%20Methoden%3A%20Diese%20Arbeit%20stellt%20das%20Open%20Medical%20Inference%20%28OMI%29%20Gateway%20vor%2C%20eine%20Interoperabilit%5Cu00e4tsschicht%2C%20die%20zwischen%20FHIR-%20und%20DICOM-basierten%20klinischen%20Datenumgebungen%20und%20heterogenen%20KI-Inferenz-Backends%20vermittelt.%20Das%20Gateway%20wurde%20in%20Python%20mit%20FastAPI%20und%20asynchroner%20Auftragsverarbeitung%20%5Cu00fcber%20Celery%20und%20RabbitMQ%20implementiert.%20Die%20Architektur%20folgt%20einem%20serviceorientierten%20Plug-in-Modell%2C%20das%20gemeinsame%20Interoperabilit%5Cu00e4tsaufgaben%20von%20dienstspezifischer%20Transformationslogik%20trennt.%20Ein-%20und%20Ausgabedaten%20werden%20an%20mehreren%20Stellen%20der%20Pipeline%20gegen%20FHIR-Profile%20und%20DICOM-Spezifikationen%20validiert%20und%20erzeugen%20OperationOutcome-Ressourcen%20als%20maschinenlesbaren%20Audit-Trail.%20%5Cn%20%5Cn%20Ergebnisse%3A%20Die%20Ende-zu-Ende-Latenz%20wurde%20%5Cu00fcber%20100%20aufeinanderfolgende%20Inferenzanfragen%20gemessen%20und%20ergab%20eine%20mediane%20Latenz%20von%20263%20ms%20bei%20einer%20Erfolgsrate%20von%20100%25.%20Das%20Konzept%20wurde%20anhand%20von%20Interoperabilit%5Cu00e4tsanforderungen%20aus%20Datenintegrationszentren%20der%20deutschen%20Medizininformatik-Initiative%20%28MII%29%20evaluiert.%20Das%20System%20ist%20mit%20einem%20einzigen%20Befehl%20einsatzbereit%20und%20als%20Open-Source-Software%20verf%5Cu00fcgbar.%5Cn%20%5Cn%20Diskussion%3A%20Die%20direkte%20Einbettung%20von%20Validierung%20und%20strukturierter%20Artefaktverarbeitung%20in%20den%20Ausf%5Cu00fchrungspfad%20der%20Inferenz%20ergibt%20ein%20besser%20steuer-%20und%20auditierbares%20Integrationsmodell%20als%20ad-hoc%20entwickelte%20Dienstadapter.%22%2C%22date%22%3A%222026%22%2C%22section%22%3A%22%22%2C%22partNumber%22%3A%22%22%2C%22partTitle%22%3A%22%22%2C%22DOI%22%3A%2210.3205%5C%2FMIBE000309%22%2C%22citationKey%22%3A%22%22%2C%22url%22%3A%22https%3A%5C%2F%5C%2Fjournals.publisso.de%5C%2Fen%5C%2Fjournals%5C%2Fmibe%5C%2Fvolume22%5C%2Fmibe000309%22%2C%22PMID%22%3A%22%22%2C%22PMCID%22%3A%22%22%2C%22ISSN%22%3A%221860-9171%22%2C%22language%22%3A%22en%22%2C%22collections%22%3A%5B%22W54IEIEG%22%5D%2C%22dateModified%22%3A%222026-09-24T08%3A10%3A36Z%22%7D%7D%2C%7B%22key%22%3A%22UIMHJA4U%22%2C%22library%22%3A%7B%22id%22%3A3757376%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22Pelka%20et%20al.%22%2C%22parsedDate%22%3A%222025-09-29%22%2C%22numChildren%22%3A0%7D%2C%22bib%22%3A%22%26lt%3Bdiv%20class%3D%26quot%3Bcsl-bib-body%26quot%3B%20style%3D%26quot%3Bline-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%26quot%3B%26gt%3B%5Cn%20%20%26lt%3Bdiv%20class%3D%26quot%3Bcsl-entry%26quot%3B%26gt%3BPelka%2C%20O.%2C%20Sigle%2C%20S.%2C%20Werner%2C%20P.%2C%20Schweizer%2C%20S.%20T.%2C%20Iancu%2C%20A.%2C%20Scherer%2C%20L.%2C%20Kamzol%2C%20N.%20A.%2C%20Eil%2C%20J.%20H.%2C%20Apfelbacher%2C%20T.%2C%20Seletkov%2C%20D.%2C%20Susetzky%2C%20T.%2C%20May%2C%20M.%20S.%2C%20Bucher%2C%20A.%20M.%2C%20Fegeler%2C%20C.%2C%20Boeker%2C%20M.%2C%20Braren%2C%20R.%2C%20Prokosch%2C%20H.-U.%2C%20%26amp%3B%20Nensa%2C%20F.%20%282025%29.%20Democratizing%20AI%20in%20Healthcare%20with%20Open%20Medical%20Inference%20%28OMI%29%3A%20Protocols%2C%20Data%20Exchange%2C%20and%20AI%20Integration.%20%26lt%3Bi%26gt%3BR%26%23xF6%3BFo%20-%20Fortschritte%20auf%20dem%20Gebiet%20der%20R%26%23xF6%3Bntgenstrahlen%20und%20der%20bildgebenden%20Verfahren%26lt%3B%5C%2Fi%26gt%3B%2C%20a-2651-6653.%20%26lt%3Ba%20class%3D%26%23039%3Bzp-DOIURL%26%23039%3B%20href%3D%26%23039%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.1055%5C%2Fa-2651-6653%26%23039%3B%26gt%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.1055%5C%2Fa-2651-6653%26lt%3B%5C%2Fa%26gt%3B%26lt%3B%5C%2Fdiv%26gt%3B%5Cn%26lt%3B%5C%2Fdiv%26gt%3B%22%2C%22data%22%3A%7B%22itemType%22%3A%22journalArticle%22%2C%22title%22%3A%22Democratizing%20AI%20in%20Healthcare%20with%20Open%20Medical%20Inference%20%28OMI%29%3A%20Protocols%2C%20Data%20Exchange%2C%20and%20AI%20Integration%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Obioma%22%2C%22lastName%22%3A%22Pelka%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Stefan%22%2C%22lastName%22%3A%22Sigle%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Patrick%22%2C%22lastName%22%3A%22Werner%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Simon%20Tobias%22%2C%22lastName%22%3A%22Schweizer%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Alexa%22%2C%22lastName%22%3A%22Iancu%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Lucas%22%2C%22lastName%22%3A%22Scherer%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Nicolas%20Andreas%22%2C%22lastName%22%3A%22Kamzol%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Jan%20Horst%22%2C%22lastName%22%3A%22Eil%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Timo%22%2C%22lastName%22%3A%22Apfelbacher%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Dmitrii%22%2C%22lastName%22%3A%22Seletkov%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Tobias%22%2C%22lastName%22%3A%22Susetzky%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Matthias%20Stefan%22%2C%22lastName%22%3A%22May%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Andreas%20Michael%22%2C%22lastName%22%3A%22Bucher%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Christian%22%2C%22lastName%22%3A%22Fegeler%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Martin%22%2C%22lastName%22%3A%22Boeker%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Rickmer%22%2C%22lastName%22%3A%22Braren%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Hans-Ulrich%22%2C%22lastName%22%3A%22Prokosch%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Felix%22%2C%22lastName%22%3A%22Nensa%22%7D%5D%2C%22abstractNote%22%3A%22Abstract%5Cn%20%20%20%20%20%20%20%20%20%20%20%20The%20integration%20of%20artificial%20intelligence%20%28AI%29%20into%20healthcare%20is%20transforming%20clinical%20decision-making%2C%20patient%20outcomes%2C%20and%20workflows.%20AI%20inference%2C%20applying%20trained%20models%20to%20new%20data%2C%20is%20central%20to%20this%20evolution%2C%20with%20cloud-based%20infrastructures%20enabling%20scalable%20AI%20deployment.%20The%20Open%20Medical%20Inference%20%28OMI%29%20platform%20democratizes%20AI%20access%20through%20open%20protocols%20and%20standardized%20data%20formats%20for%20seamless%2C%20interoperable%20healthcare%20data%20exchange.%20By%20integrating%20standards%20like%20FHIR%20and%20DICOMweb%2C%20OMI%20ensures%20interoperability%20between%20healthcare%20institutions%20and%20AI%20services%20while%20fostering%20ethical%20AI%20use%20through%20a%20governance%20framework%20addressing%20privacy%2C%20transparency%2C%20and%20fairness.%5Cn%20%20%20%20%20%20%20%20%20%20%20%20OMI%5Cu2019s%20implementation%20is%20structured%20into%20work%20packages%2C%20each%20addressing%20technical%20and%20ethical%20aspects.%20These%20include%20expanding%20the%20Medical%20Informatics%20Initiative%20%28MII%29%20Core%20Dataset%20for%20medical%20imaging%2C%20developing%20infrastructure%20for%20AI%20inference%2C%20and%20creating%20an%20open-source%20DICOMweb%20adapter%20for%20legacy%20systems.%20Standardized%20data%20formats%20ensure%20interoperability%2C%20while%20the%20AI%20Governance%20Framework%20promotes%20trust%20and%20responsible%20AI%20use.%5Cn%20%20%20%20%20%20%20%20%20%20%20%20The%20project%20aims%20to%20establish%20an%20interoperable%20AI%20network%20across%20healthcare%20institutions%2C%20connecting%20existing%20infrastructures%20and%20AI%20services%20to%20enhance%20clinical%20outcomes.%22%2C%22date%22%3A%222025-09-29%22%2C%22section%22%3A%22%22%2C%22partNumber%22%3A%22%22%2C%22partTitle%22%3A%22%22%2C%22DOI%22%3A%2210.1055%5C%2Fa-2651-6653%22%2C%22citationKey%22%3A%22%22%2C%22url%22%3A%22http%3A%5C%2F%5C%2Fwww.thieme-connect.de%5C%2FDOI%5C%2FDOI%3F10.1055%5C%2Fa-2651-6653%22%2C%22PMID%22%3A%22%22%2C%22PMCID%22%3A%22%22%2C%22ISSN%22%3A%221438-9029%2C%201438-9010%22%2C%22language%22%3A%22de%22%2C%22collections%22%3A%5B%22W54IEIEG%22%5D%2C%22dateModified%22%3A%222025-10-13T12%3A02%3A16Z%22%7D%7D%5D%7D
Mathes, G., Berger, S., & Sigle, S. (2026). Automated Extraction of Genetic Eligibility Criteria from Clinical Trial Records Using LLMs – A Technical Case Report. In U. Sax, T. Ganslandt, R. Breitschwerdt, K. Jung, J. König, & A. Großhennig (Eds.), Studies in Health Technology and Informatics. IOS Press. https://doi.org/10.3233/SHTI260998
Pelka, O., Eil, J., Manjunatha, K., Singh, N., Omeirat, J., Sigle, S., Mathes, G., Schweizer, S.-T., Stump, S.-R., Girdziunaite, G., & Nensa, F. (2026). Interoperable clinical data integration for distributed medical AI: The Open Medical Inference Gateway. GMS Medizinische Informatik, Biometrie Und Epidemiologie, 22. https://doi.org/10.3205/MIBE000309
Pelka, O., Sigle, S., Werner, P., Schweizer, S. T., Iancu, A., Scherer, L., Kamzol, N. A., Eil, J. H., Apfelbacher, T., Seletkov, D., Susetzky, T., May, M. S., Bucher, A. M., Fegeler, C., Boeker, M., Braren, R., Prokosch, H.-U., & Nensa, F. (2025). Democratizing AI in Healthcare with Open Medical Inference (OMI): Protocols, Data Exchange, and AI Integration. RöFo - Fortschritte auf dem Gebiet der Röntgenstrahlen und der bildgebenden Verfahren, a-2651-6653. https://doi.org/10.1055/a-2651-6653
Kollaborationen
IT-Lösungen für wissenschaftliche Projekte
Wir sind offen für Kollaborationen zur Durchführung wissenschaftlicher Projekte und zur Bereitstellung von Forschungsinfrastruktur. Unsere Expertise hilft Ihnen, Projekte effizient umzusetzen und die nötige IT- und Datenbasis zu schaffen.
Aktuelle Projekte aus Datascience & AI
Open Medical Interference (OMI)
Die OMI-Plattform ermöglicht den Austausch und die Nutzung von Gesundheitsdaten und
KI-Diensten durch offene Protokolle
und Datenformate.
Lion-App für mehr Lebensqualität
Die Lion-App verbessert die Lebensqualität von Patienten durch Erfassung und Analyse von Wohlbefinden und Symptomen, um die Arztkommunikation zu unterstützen.
Studienplattform CATS
CATS ist eine Plattform zur Verwaltung und semantischen Filterung von Studien, die personalisierte Medizin durch moderne Datenmodelle unterstützt.
Variant Browser (VB)
Der Variant-Browser bringt relevante Datenquellen zu genetischen Varianten zusammen und bringt verteilte Abfragen in einer einheitlichen Übersicht zusammen.
Knowledge Artifacts (KA)
Behandlungsdaten werden mithilfe von KI-Technologien anonymisiert, extrahiert, strukturiert und im Forschungskontext gezielt für für die Entscheidungsfindung genutzt.
AI MTB-Assistant
Der AI Assistant ermöglicht eine Interaktion zwischen klinischen Anwender:innen und Large Language Models – mit Fokus auf zuverlässigen, evidenzbasierten Ausgaben.
Qualitätsmanagement in molekularen Tumorboards
Digitale Werkzeuge schaffen Transparenz über Zusammensetzung und Entwicklung und ermöglichen ein datenbasiertes, automatisiertes QM.
Prädiktionsmodelle in der personalisierte Medizin
Mit ML und KI werden öffentlich verfügbare Daten verarbeitet, um prädiktive Modelle für die klinische Entscheidungsunterstützung im Forschungskontext zu entwickeln.