Applying Semantic-Web Techniques to Translational Clinical ...

manyfarmswalkingInternet and Web Development

Oct 21, 2013 (3 years and 9 months ago)

102 views

Applying Semantic-Web Techniques to Translational Clinical Research
The rapid increase in the volume of electronic medical records (EMR) available for research purposes provides new op-
portunities to create semantically interoperable healthcare applications and solutions for individualized and evidence-based
medicine.Discovery,analysis and dissemination of clinical information,however,requires a robust and semantically crisp
model to represent the clinical data.The Semantic Web and the Web Ontology Language (OWL) provide a suitable envi-
ronment for modeling clinical data and reasoning about them.The use of OWL to represent clinical information brings many
benefits.First,it provides a standard mechanism with explicit and formal semantic knowledge representation.Secondly,
the Semantic Web offers automated reasoning capabilities.Thirdly,once we have an ontology that can represent semantic
assertions in the clinical domain precisely,we can annotate clinical data with respect to the ontology and store the instances
as RDF triples.The information then become “machine-understandable”.Tools and services such as reasoners,editors,
querying systems,and storage mechanisms that have been developed by the Semantic Web community can be directly
applied to the clinical data.
While the Semantic Web and the Web Ontology Language (OWL) provide a suitable foundation,not all the biomedical
ontologies and terminologies are in the OWL format.In addition,certain guidelines for representing commonly-used meta-
data need to be fulfilled,even within the OWL ontologies,in order to reach the level of harmonization desired for a shared
set of semantics and operational rules.Toward this direction,I have initialized and amcurrently leading the LexRDF project.
LexRDF is designed based on the LexGrid project,which provides a terminology model to represent multiple vocabulary and
ontology sources as well as a scalable and robust API for accessing such information.While successfully used and adopted
in the biomedical and clinical community,an important requirement is to align the existing LexGrid model with emerging
Semantic Web standards and specifications.The first step of LexRDF was to create a set of mapping specifications from
LexGrid to the W3C Semantic Web standards.The LexRDF model has also been presented in different meetings such
as the National Cancer Institute (NCI) Cancer Biomedical Informatics Grid (caBIG) face to face Meetings and the National
Center for Biomedical Ontology (NCBO) face to face meeting.It will potentially be adopted by the NCBO and caBIG as the
back end data model.Our next step for LexRDF is to generalize the model so that it can be applied beyond the LexGrid
environment.I am currently working on designing a set of common terminology guidelines,which help ontology designers
and developers to generate OWL ontologies or translate/convert ontologies and terminologies in other formats to OWL.
In addition,a LexRDF triple store is being implemented for representing biomedical ontologies according to the LexRDF
mapping specifications.I will further evaluate the benefits of using the common ontology guideline system by evaluating
different query cases and examples from heterogenous biomedical ontologies.
One application of using the Semantic-Web technologies in clinical research is the temporal relation reasoning project.
The ability to answer temporal-oriented questions based on clinical narratives is essential to clinical research.The temporal
dimension in medical data analysis enables clinical researches on many areas,such as,disease progress,individualized
treatment,and decision support.Existing temporal models are either not compatible with the powerful reasoning tools de-
veloped for the Semantic Web,or designed only for structured clinical data and therefore are not ready to be applied on
natural-language-based clinical narrative reports directly.We have developed a Semantic-Web ontology which is called
Clinical Narrative Temporal Relation ontology.Using this ontology,temporal information in clinical narratives can be repre-
sented as RDF (Resource Description Framework) triples.More temporal information and relations can then be inferred by
Semantic-Web based reasoning tools.Experimental results show that this ontology can represent temporal information in
real clinical narratives successfully.In addition,we have finished the preliminary implementation of a framework for tem-
poral reasoning from annotated patient data.This framework combines DL-based reasoning,SWRL-based reasoning,and
SWRL Built-Ins library.It provides a query API for users to query temporal data in clinical narratives.My next step for the
project is to:1) investigate on how to handle temporal uncertainties,which is quite common in clinical narratives;2) design
an ontology-based system that allows users to ask free form queries and translate the queries in computer languages;3)
investigate on better harmonizing the CNTRO ontology with standards;and 4) explore the possibilities to use CNTRO to
represent concept with temporal relations in common medical ontologies such as SNOMED CT.We will also extend and
improve cTAKES and use it as an automatic annotator for temporal information and annotate information with respect to the
CNTRO ontology.
Intellectual Merit.
This research will 1) provide meta-level guidelines for representing biomedical knowledge in OWL and
harmonize heterogenous biomedical ontologies into a unified model;2) support time-oriented question answering from
clinical narratives using semantic-web techniques;and 3) explore mechanisms of using Semantic-Web representations and
techniques for disease phenotyping.
Broader Impact.
This research work will (1) help clinical researchers harvest and make available facts from the wealth of
available heterogeneous digital data;(2) save critical time and effort in biomedical research and thus benefit patients who
could receive treatments/medications sooner.
Cui Tao,PhD,Division Of Biomedical Statistics and Informatics,Mayo Clinic1