, Design of PET Image Database and Retrieval for Diagnosis Support

Transcription

1 DEWS C-01 * {d03hc006 b *tommy}@ynu ac jp PET(Positron Emission Tomography) DBMS PET Design of PET Image Database and Retrieval for Diagnosis Support TetsuyaISHIE Kazunori SUNAKO Takashi TOMII* Department of Information Media and Environment Sciences Graduate School of Environment and Information Sciences Yokohama National University Division of Electrical and Computer Engineering School of Engineering Yokohama National University *Faculty of Environment and Information Sciences Yokohama National University 79-7 Tokiwadai Hodogaya-ku Yokohama Japan {d03hc006 b *tommy}@ynuacjp Abstract In recent years diagnosis by the whole body PET (Positron Emission Tomography) Images attracts attention as an effective diagnostic method of cancer The diagnostic method which a doctor performs at the time of the conventional interpretation of radiogram is matching the medical knowledge of the position of anatomical internal organs and physiological accumulation and the specific position in images in the doctor's head and describing the report The semantic information which pixel group in images corresponds to internal organs is not clearly recorded in that case Reuse of the diagnostic information at the time of interpretation of radiogram is difficult In this paper it aims at the diagnostic support which used DB Medical knowledge and image data are associated by defining clearly the domain corresponding to internal organs and accumulation which the doctor extracted at the time of interpretation of radiogram and accumulation and reference are possible For that purpose we design the schema which has 3 layers---knowledge notional existence at the time of interpretation of radiogram and image data Moreover we have implemented this schema on practical DBMS -- the usefulness of this schema and the practicality of this system were shown by actually inquiring the database Keyword Medical Information Processing Medical Image Database PET Positron Emission Tomography Diagnosis Support 1 PET mm PET(Positron Emission Tomography MRI CT PET ) [1] PET

2 PET Server PET Manager System PET PET PET DB PET Raw data 1PET PET 2PET CBIR Content based Image Retrieval [2] CBIR [3] FDG SUV(Standardized uptake value)[7] SUV FDG SUV SUV CT MRI PET CBIR (Conceptual data model) [4][5] PET [6] PET PET CT MRI [1] FDG FDG PET PET DICOM [8] PET 128*128Pixel16bit DBMS PET 2 PET 3PET PET PET 31 2 (Positron Emission TomographyPET) PET PET PET (F-18) FDG PET 1

3 Process-4 Process-2 Process3 DB server FDG CT [9] (4)(5) PET () SUV () SUV SUV ROI(Region of ROI Interest: ) SUV SUV 3 PET [10] FDG [11] PET FDG SUV 4 PET PET 41 SUV raw data ( ) Process-1 PET DB serverpet SUV 15 Process-2 SUV Process-1 SUV Process-3

4 () [12][13] [14] (raw data ) PET DICOM [16] PET 128*128(Pixel)CT [16] () 2 raw data 1 a raw data () raw data 3 42PET 4 3 DBMS DBMS DB server Microsoft SQLserver2000 [ ] SQLserver2000 (1) (2) raw data Image1 a 1 Image b c Image 3 Raw Data 4 PET

5 (3) 5 (Raw Data ) DBMS PET 5 [ ] 4 2 PET Ont_Region 5 1 SUV DBserver Dell Computer Dell Ont_Accumulation Precision530 (Xeon2 4GHz 1GB) [Raw data ] raw data Process-1 DICOM Raw Data PET Raw Data 1 [ ] Process-4 ID:6 1 ID ID Region Segmentation Categorization R000 A000 Accumulation Identification Definition R001 R002 A001 A002 R003 Inclusion R004 GUI R005 [11] R006 R007 2 [17] ID A B C D E F G H I J Raw Data A R Raw Data A

6 5 2 : Rawdata 3 Query Query DB [ ] PET (1) Query: 8 ID Q SUV 5 2 ROI SUV (2) Query: SUV MAX MIN raw data (3) Query3: SUV raw data SUV [ ] 1 SQL DBserver Query Query SQL 7 ] 7 AccumulationID A A A A AcucumulationID 5 ID 7 QuerySQL 9 Query3 SQL 8 Query2 SQL

7 B F J J 3030 B F G H [ ] PET B F J (1) B J B J Query1-1: RegionID: R F Query1-2 (2) SUV :B F Query2-1: PatientID : 0001 Query2-2 PET SUV (3) Query3-1: SUV Query3-2 (4) DB : 80 : 20 Query4-1: 9 Query4-2 RegionID R R ) Query4-3: AccumulationID:A A Query4-4 [ ] DBserver 2 PET ID Query Query1-1 MAX Query1-2 AVG 5 3 PET PET DB-1 DB-2 DB Query2-1 MAX Query2-2 AVG Query3-1 MAX Query3-2 AVG Query4-1 MAX Query4-2 AVG Query4-3 MAX Query4-4 AVG

8 [ ] Query [1] HP SQL JOIN SQL PET PET 54 PET DICOM DICOM D Viewer[18] CG 6 [6] DBS-131(II) pp33-40 Aug 2003 [7] Vol 23 No 10 pp October2003 [8] JIRA/DICOM HP [9] [10] 3D-ACTIT MIPS [11] PET DE pp7-12 October 2003 [12] JOHN F SOWA Knowledge Representation: Logical Philosophical and Computational Foundations Brooks/Cole DBMS 2000 [13] Gruber Thomas R A translation approach to portable ontology specifications In Knowledge Acquisition vol5 pp [14] Unified Medical Language System(UMLS) [15] Aconomy SQL [16] Tianqiu Wang Simone Santini Amarnath Gupta An Interpolated Volume Model for Database ER2003 LNCS 2813 pp [17] VolPRMU pp [18] Amira PET ( )