Скачать презентацию Eu So MII 25 September 2014 Warsaw Poland Скачать презентацию Eu So MII 25 September 2014 Warsaw Poland

2ec35449b8ef20edd3537a9ef8ff2b97.ppt

  • Количество слайдов: 37

Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers and PACS ontologies Bernard Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers and PACS ontologies Bernard Gibaud Medi. CIS, LTSI, U 1099 Inserm Faculté de médecine, Rennes bernard. gibaud@univ-rennes 1. fr 1

Eu. So. MII, 25 September 2014, Warsaw (Poland) Goal of the presentation • To Eu. So. MII, 25 September 2014, Warsaw (Poland) Goal of the presentation • To highlight possibilities that future image data sharing infrastructures should provide to optimize the production, sharing, use and reuse if imaging biomarkers 2

Eu. So. MII, 25 September 2014, Warsaw (Poland) Overview • Introduction (definition of imaging Eu. So. MII, 25 September 2014, Warsaw (Poland) Overview • Introduction (definition of imaging biomarkers) • Part 1. Change of paradigm (led by imaging biomarkers) • Part 2. Infrastructures for producing, sharing and using imaging biomarkers • Part 3. Imaging biomarkers modeling using ontologies • Conclusion 3

Eu. So. MII, 25 September 2014, Warsaw (Poland) Introduction Definition 4 Eu. So. MII, 25 September 2014, Warsaw (Poland) Introduction Definition 4

Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers Paper prepared by the Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers Paper prepared by the ESR Subcommittee on Imaging Biomarkers (chairperson: Bernard Van Beers) 5

Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers • Definition of biomarkers Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers • Definition of biomarkers (Atkinson 2001)* – « characteristics that are objectively measured and evaluated as indicators of • normal biological processes, • pathological processes, • pharmaceutical responses to a therapeutic intervention » • Definition of (quantitative) imaging biomarkers – Derived from medical images – Quantitative, objective, reproducible – « qualified » for specific clinical uses * Clin Pharmacol & Ther. 2001 Mar; 69(3): 89 -95. Biomarkers and surrogate endpoints: preferred definitions and conceptual framework. Biomarkers Definitions Working Group. 6

Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers • Used for – Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers • Used for – Early detection of disease – Staging and grading – Predicting response to treatment – Assessing response to treatment 7

Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers • Of critical importance Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers • Of critical importance in research – Focused clinical research (e. g. controlled clinical trials) – Translational research • Link/correlate results obtained in various domains • Need to share them at a broad scale Key aspect of federated imaging biobanks • Of critical importance in individual patient management (decision criteria) – Diagnosis, prognosis, treatment Key aspect of a structured EHR / tasks planning 8

Eu. So. MII, 25 September 2014, Warsaw (Poland) New paradigm led by imaging biomarkers Eu. So. MII, 25 September 2014, Warsaw (Poland) New paradigm led by imaging biomarkers 9

Eu. So. MII, 25 September 2014, Warsaw (Poland) General framework Acquisition Processing Reality Images Eu. So. MII, 25 September 2014, Warsaw (Poland) General framework Acquisition Processing Reality Images Human suject Animal subject Specimen etc. MR image CT image PET image etc. Decision Imaging biomarkers Volume of anatomical structure Fractal dimension Mean reg. blood volume Lesion load (MS) etc. Facts Plans, etc. Diagnosis of AD Diagnosis of MS Resp to treatment etc. 10

Eu. So. MII, 25 September 2014, Warsaw (Poland) Change of paradigm • In many Eu. So. MII, 25 September 2014, Warsaw (Poland) Change of paradigm • In many clinical situations, obtaining imaging biomarkers may become a primary goal of the imaging procedure • The clinical goal (clinical question) determines what imaging biomarkers are needed 11

Eu. So. MII, 25 September 2014, Warsaw (Poland) Change of paradigm Acquisition Detailed subject/spec Eu. So. MII, 25 September 2014, Warsaw (Poland) Change of paradigm Acquisition Detailed subject/spec imen preparation Processing Detailed imaging protocol Set of required imaging biomarkers Decision Scientific question to be answered or Clinical question 12

Eu. So. MII, 25 September 2014, Warsaw (Poland) Consequence (1/2) • Consensus building – Eu. So. MII, 25 September 2014, Warsaw (Poland) Consequence (1/2) • Consensus building – What imaging biomarker in which situation – Conditions of image acquisition (acquisition protocol, image reconstruction) – Precise definition of image processing • Work in progress in RSNA’s QIBA initiative – Quantitative Imaging Biomarkers Alliance – QIBA Mission: « Improve the value and practicality of quantitative biomarkers by reducing variability across devices, patients and time » – Development of « profiles » 13

Eu. So. MII, 25 September 2014, Warsaw (Poland) Consequence (2/2) • Information modeling – Eu. So. MII, 25 September 2014, Warsaw (Poland) Consequence (2/2) • Information modeling – Imaging biomarkers – As part of Structured Reports • To be managed in clinical PACS and « clinical research supporting systems » – Intelligent task management systems (workflow) – Decision support systems – Quality management systems 14

Eu. So. MII, 25 September 2014, Warsaw (Poland) Part 2. Infrastructure for producing, sharing Eu. So. MII, 25 September 2014, Warsaw (Poland) Part 2. Infrastructure for producing, sharing and using imaging biomarkers 15

Eu. So. MII, 25 September 2014, Warsaw (Poland) Infrastructure for producing, sharing and using Eu. So. MII, 25 September 2014, Warsaw (Poland) Infrastructure for producing, sharing and using image biomarkers • Two aspects – Data structures / information models – System infrastructure • Two application domains – Care delivery / clinical routine – Clinical and translational research 16

Eu. So. MII, 25 September 2014, Warsaw (Poland) Data structures for imaging biomarkers clinical Eu. So. MII, 25 September 2014, Warsaw (Poland) Data structures for imaging biomarkers clinical application • Mostly – Measurements included in free text reports – and results (e. g. of RECIST) available in separated unstructured documents (e. g. pdf, DICOM secondary capture) • Sometimes – Structured reports – DICOM structured reports Still often non-structured, because no subsequent automated 17 processing

Eu. So. MII, 25 September 2014, Warsaw (Poland) System infrastructures for imaging biomarkers clinical Eu. So. MII, 25 September 2014, Warsaw (Poland) System infrastructures for imaging biomarkers clinical application • Production of imaging biomarkers – General purpose image processing software of PACS WS – Specific software (plug-in, remote cloud server) • Sharing of imaging biomarkers – PACS archive • Use and re-use of imaging biomarkers – Mostly: human reading (e. g. clinician reading for medical decision, radiologist reading for comparison with baseline) Limited automated reuse because not-structured 18

Eu. So. MII, 25 September 2014, Warsaw (Poland) Data structures for imaging biomarkers research Eu. So. MII, 25 September 2014, Warsaw (Poland) Data structures for imaging biomarkers research application • Always – Measurements in structured files / databases • Entered via – Structured forms or e. CRF (by end-user) – Structured document produced by image processing software (specific XML structure, but also generic data structures, e. g. XCEDE*) – DICOM structured reports (sometimes) • e. g. Quantitative Image Informatics in Cancer Research (QIICR) Always structured because of subsequent automated processing (statistical analysis) * XCEDE: XML-based Clinical and Experimental Data Exchange 19

Eu. So. MII, 25 September 2014, Warsaw (Poland) System infrastructures for imaging biomarkers research Eu. So. MII, 25 September 2014, Warsaw (Poland) System infrastructures for imaging biomarkers research application • Production of imaging biomarkers – Mostly: specific image processing software provided by CRO (accessible as a web app in mono or multi-center studies) • Sharing of imaging biomarkers – Clinical research system are often limited to collecting and processing the results according to the study protocol – Sometimes, complementary « open image archives » (OIA) to disseminate original and derived image data (e. g. ADNI, Human Connectome Project, Cancer Imaging Archive) • Use and re-use of imaging biomarkers – Little reuse since data considered « non public » – OIA often limited to original image data, no online metadata 20 based data query/retrieve

Eu. So. MII, 25 September 2014, Warsaw (Poland) Part 3. Imaging biomarkers modeling using Eu. So. MII, 25 September 2014, Warsaw (Poland) Part 3. Imaging biomarkers modeling using ontologies 21

Eu. So. MII, 25 September 2014, Warsaw (Poland) Semantic web technologies • Ontologies and Eu. So. MII, 25 September 2014, Warsaw (Poland) Semantic web technologies • Ontologies and ontology languages • Ontology editors, e. g. Protégé (Stanford Univ. ) • Query languages, e. g. SPARQL (W 3 C recomm) • Reasoners, e. g. Fa. CT++, Pellet, Hermi. T 22

Eu. So. MII, 25 September 2014, Warsaw (Poland) Ontologies • Definition (informatics and AI) Eu. So. MII, 25 September 2014, Warsaw (Poland) Ontologies • Definition (informatics and AI) – « a formal, explicit specification of a shared conceptualization » (Gruber 1993) • Two basic aspects – A shared vocabulary – Formal semantics : axioms expressed in a logical language 23

Eu. So. MII, 25 September 2014, Warsaw (Poland) Formal semantics • Definitions of classes Eu. So. MII, 25 September 2014, Warsaw (Poland) Formal semantics • Definitions of classes of objects – Taxonomy of classes: subsumption (i. e. « is a » relation) – Instanciation (relation between an individual and a class) • Definitions of properties – Taxonomy of properties – Domain and range, inverse properties, etc. • Processing by a reasoning engine – Assess satisfiability (consistency) – Classification of ontologies – Classification of instances Reasoners are not application-specific 24

Eu. So. MII, 25 September 2014, Warsaw (Poland) Ontological modeling of imaging biomarkers • Eu. So. MII, 25 September 2014, Warsaw (Poland) Ontological modeling of imaging biomarkers • Note: It is important to distinguish… – Imaging biomarker as result of a measurement – from its role in some medical decision (e. g. diagnosis, prognostic) • Main aspects an ontological model of biomarkers should address – – Measure Relation to reality Provenance Context 25

Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers: measure Relies on long Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers: measure Relies on long and deep experience in metrology 26

Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers: measure • Is the Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers: measure • Is the result of some measurement process (manual or implemented in image processing software) • Indirectly involves a physical object under study, and / or a process under study (dynamic process or longitudinal process) in which this object participates • Concerns a specific quality of this object, or of the process under study) – Note: This quality may be a complex human construct (e. g. modelbased: fractal dimension, gyrification index) • Values chosen from a predefined scale of measurement – interval, ratio, ordinal, nominal (categorical) 27

Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers: relation to reality • Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers: relation to reality • A Measurement of a quality beared by an object • Or a Measurement of a temporal quality of the process under study • (Simple) Examples – Volume of hippocampus (in cm 3) – Speed of brain atrophy process – neuronal loss (in cm 3/year) – Mean Fractional Anisotropy over uncinate fasciculus 28

Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers: provenance • Execution of Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers: provenance • Execution of a program implementing some conceptual action (e. g. , a segmentation) • Resources used in this execution (e. g. , user, date, software tool, platform) • Input data (e. g. , datasets, ROIs, imaging biomarkers) • Input parameters (if any) 29

Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers: context • Case 1: Eu. So. MII, 25 September 2014, Warsaw (Poland) Imaging biomarkers: context • Case 1: Relation to a research question – Measurement process is part of the execution of research protocol – Context is provided by the research goal and protocol • Case 2: Relation to a clinical question – Measurement process is part of the actions performed to answer the clinical question (possibly detailed via a protocol, and/or a report template) – Context is provided by the clinical question and associated clinical information 30

Eu. So. MII, 25 September 2014, Warsaw (Poland) Principal ontologies to start from • Eu. So. MII, 25 September 2014, Warsaw (Poland) Principal ontologies to start from • Foundational ontologies: BFO or DOLCE – Provide a common modeling framework as well as the major top-level entities • • Qualities: PATO Provenance: PROV Measurement and information artifacts: OBI / IAO Imaging: Rad. Lex Imaging datasets: Onto. Neuro. Log Imaging biomarkers: QIBO Medicine in general: SNOMED, ICD, NCIT But of inequality and completeness 31

Eu. So. MII, 25 September 2014, Warsaw (Poland) Added value of ontologies for imaging Eu. So. MII, 25 September 2014, Warsaw (Poland) Added value of ontologies for imaging biobanks (1) • Standard vocabulary for images and imaging biomarkers, facilitating data sharing in large image repositories (imaging biobanks) • More explicit and more formal representation, enabling – intelligent querying (SPARQL) – rich inferencing capabilities implemented in generic (i. e. domain-agnostic) reasoners 32

Eu. So. MII, 25 September 2014, Warsaw (Poland) Added value of ontologies for imaging Eu. So. MII, 25 September 2014, Warsaw (Poland) Added value of ontologies for imaging biobanks (2) • Complementarity to existing infrastructures – e. g. a semantic implementation of imaging biomarkers might complement regular DICOM SR files – Possible implementation: RDF files accessible as a SPARQL endpoint • Allows enhancing data sharing infrastructures with new and powerful search capabilities 33

Eu. So. MII, 25 September 2014, Warsaw (Poland) Agenda (1/2) • Develop/extend the relevant Eu. So. MII, 25 September 2014, Warsaw (Poland) Agenda (1/2) • Develop/extend the relevant domain ontologies – Needs to involve relevant domain experts • the different radiological specialities • the DICOM standard community • the editors of major image processing packages – as well as ontologists • difficult task 34

Eu. So. MII, 25 September 2014, Warsaw (Poland) Agenda (2/2) • Deploy experimental systems Eu. So. MII, 25 September 2014, Warsaw (Poland) Agenda (2/2) • Deploy experimental systems in research infrastructures and imaging biobanks first. • Then, consider deployment into clinical PACS, to support – Intelligent task management systems (workflow) – Decision support systems – Quality management systems 35

Eu. So. MII, 25 September 2014, Warsaw (Poland) Conclusion / Summary • We underlined Eu. So. MII, 25 September 2014, Warsaw (Poland) Conclusion / Summary • We underlined the importance of imaging biomarkers in research (both clinical and translational research) and care delivery • We underlined the importance of structured reports as a convenient way to produce and share them for both research and clinical applications • And finally we introduced ontologies and we discussed their added value in both imaging biobanks and clinical PACS 36

Eu. So. MII, 25 September 2014, Warsaw (Poland) Thank you for your attention 37 Eu. So. MII, 25 September 2014, Warsaw (Poland) Thank you for your attention 37