Docstoc

Outcomes and trends of Lithuanian Swedish telemedicine collaboration fundus

Document Sample
Outcomes and trends of Lithuanian Swedish telemedicine collaboration fundus Powered By Docstoc
					    APPLICATIONS OF INFORMATION
    TECHNOLOGIES FOR RESEARCH
     IN CLINICAL OPHTHALMOLOGY
     A.Paunksnis1,2, S.Kurapkienė1,2, V.Barzdžiukas1,2,
              A.Lukoševičius3, G.Dzemyda4

1Telemedicine   Center of Kaunas University of Medicine, Lithuania
2Dept. of Ophthalmology of Institute for Biomedical Research of Kaunas
University of Medicine, Lithuania
3   Kaunas University of Technology, Lithuania
4Vilnius   Mathematics and Informatics Institute, Lithuania



                                                                     1
                              18-          Luxexpo,
                        April 18-20, 2007, Luxexpo, Luxembourg
• This report covers the results of the research
  work performed in frame of the project
  “Information technologies for human health –
  clinical decision support “IT Health” during
  the period 2003-2006, financed by Lithuanian
  State Science and Studies Foundation.




                                                      2
                   18-          Luxexpo,
             April 18-20, 2007, Luxexpo, Luxembourg
The eHEALTH NETWORK for Integrated Support of Clinical Practice, Education and
Research connecting KAUNAS eHEALTH CLUSTER created. The CLUSTER
consists of Kaunas Medical University Hospital (http://www.kmu.lt), Telemedicine
Center of Kaunas University of Medicine (http://tmc.kmu.lt), Telemedicine Support
Center of Kaunas University of Technology (http://www.bmii.ktu.lt), Institute of
Mathematics and Informatics (http://www.mii.lt)




        Telemedicine Center of Kaunas           Telemedicine Support Center of
            University of Medicine              Kaunas University of Technology




                                            Institute of Mathematics and Informatics
                                                                                       3
       Kaunas Medical University Hospital
Purpose
To create a portal for:
1. Clinical decision support for ophthalmolgists including
   automatic and interactive eye fundus images
   parameterization and recognition;
2. Clinical decision support for family doctors including
   automatic and interactive signal and data parameterization
   and recognition;
3. CME – continuous medical education for ophthalmologists
   and family doctors;
4. Patients education.

                                                             4
                          18-          Luxexpo,
                    April 18-20, 2007, Luxexpo, Luxembourg
 • The main targets of IT applications for research in
   clinical ophthalmology are images and signals
   (optical and ultrasound data) - the most informative
   sources for differential diagnosis.
The main activities:

•Registration and parameterization of diagnostic signals and
images, selection of most informative parameters, creation of
databases and interfaces;

•Data mining and clinical decision support using learning
algorithms, remote access to services and expertise of physicians.

•Development of prognostic, diagnostic and treatment
recommendations using the knowledge portals.



                                                                5
                           18-          Luxexpo,
                     April 18-20, 2007, Luxexpo, Luxembourg
Materials and Methods
We have collected 7000 eye fundus images: healthy eyes, glaucoma and
 myopia.
 The automatic and interactive image processing algorithms developed for
 analysis of digital images of the eye fundus.

Framework of the eye fundus image analysis consist of:
• Estimation of the eye fundus image quality
IMG_1011.jpg   Optic nerve swelling    Optic nerve      Optic nerve head is swollen, pale
               (papilledema) of the    changes.         (beginning of secondary optic nerve
               right eye. Regression                    atrophy), with blurred elevated margins,
               stadium.                                 deflection of vessels over edge of the
                                                        disc.


                                                                       • Description table
                                                                           and grids for
                                                                           localization of the
                                                                           eye fundus changes
                                                                           and parametrization
                                                                           of retinal vessels.



                                                                                            6
                                    18-          Luxexpo,
                              April 18-20, 2007, Luxexpo, Luxembourg
Materials and Methods
• Localization of the main landmarks in the eye fundus
                     Optic nerve head




                                                            7
                         18-          Luxexpo,
                   April 18-20, 2007, Luxexpo, Luxembourg
Materials and Methods
• Segmentation of the optic nerve head and vessels
  Optic nerve head




                                                              8
                           18-          Luxexpo,
                     April 18-20, 2007, Luxexpo, Luxembourg
Materials and Methods
 • Calculation of quantitative parameters of segmented
 objects.
Ophthalmological image parametrisation, selection of diagnostic
                        parameters

                                             The example:
                                        33 images of 4 different
                                     pathologies, described by the
                                      complex of 34 parameters
                                       (total: 1122 parameters)



                                                               a1,1, k      a 2 ,1, k   ...   a n ,1, k                                                 a1,1, k         a 2 ,1, k     ...     a n ,1, k
                                                                                              a n , 2 ,k                                                                                      a n , 2 ,k
                                                                                                ...                                                                                             ...
                                     a1,1, 2       a 2 ,1, 2      ... a n ,1, 2               a n ,m ,k                    a1,1, 2          a 2 ,1, 2      ... a n ,1, 2                      a n ,m ,k
                                                                           an,2,2                                                                                   an,2,2
                          a1,1,1      a 2 ,1,1       ...        a n ,1,1 ...                                    a1,1,1       a 2 ,1,1         ...        a n ,1,1 ...
                          a1, 2 ,1    a 2 , 2 ,1     ...        a n , 2 ,1 a                                    a1, 2 ,1     a 2 , 2 ,1       ...        a n , 2 ,1 a
                                                                             n ,m , 2                                                                                 n ,m , 2
                           ...          ...          ...          ...                                            ...           ...            ...          ...
                          a1, m ,1    a 2 , m ,1     ...        a n , m ,1                                      a1, m ,1     a 2 , m ,1       ...        a n , m ,1




                                                               a1,1, k      a 2 ,1, k   ...   a n ,1, k                                         a1,1, k           a 2 ,1, k     ...     a n ,1, k
                                                                                              a n , 2 ,k                                                                                a n , 2 ,k
                                                                                                ...                                                                                       ...
                                     a1,1, 2       a 2 ,1, 2      ... a n ,1, 2               a n ,m ,k               a1,1, 2       a 2 ,1, 2      ... a n ,1, 2                        a n ,m ,k
                                                                           an,2,2                                                                           an,2,2
                          a1,1,1      a 2 ,1,1       ...        a n ,1,1 ...                               a1,1,1      a 2 ,1,1       ...        a n ,1,1 ...
                          a1, 2 ,1    a 2 , 2 ,1     ...        a n , 2 ,1 a                               a1, 2 ,1    a 2 , 2 ,1     ...        a n , 2 ,1 a
                                                                             n ,m , 2                                                                         n ,m , 2
                           ...          ...          ...          ...                                       ...          ...          ...          ...
                          a1, m ,1    a 2 , m ,1     ...        a n , m ,1                                 a1, m ,1    a 2 , m ,1     ...        a n , m ,1




                                                                                                                                                                                                           9
                             18-          Luxexpo,
                       April 18-20, 2007, Luxexpo, Luxembourg
Materials and Methods
    For interactive parametrization of blood
    vessels created program based on Matlab




                                                          10
                       18-          Luxexpo,
                 April 18-20, 2007, Luxexpo, Luxembourg
Interactive image processing
             Selection of blood vessels

The created program based on Matlab




                                          11
Automatic image processing
          Excluding of the network of blood vessels from fundus image

   A grey level morphology scheme proposed by Mendel’s was used.




                                                                        12
                                18-          Luxexpo,
                          April 18-20, 2007, Luxexpo, Luxembourg
                       Results
• Methods and software for eye fundus images
parametrization – including automatic recognition of optical
nerve region, automatic outlining of this region and calculation
of geometric parameters of the optical nerve, excavation and
retinal vessels created.
• The set of parameters reflecting the diagnostically important
eye fundus features calculated.
• The method of excluding of the network of blood vessels
from fundus image was developed enabling to improve
diagnosis of retinal damages.



                                                             13
                          18-          Luxexpo,
                    April 18-20, 2007, Luxexpo, Luxembourg
Results
Created portal for:
• clinical decision support for ophthalmolgists including
    automatic and interactive eye fundus image
    parameterization and recognition;

•   clinical decision support for family doctors including
    automatic and interactive data parameterization and
    recognition;

•   CME – continuous medical education for
    ophthalmologists and family doctors;



                                                             14
                         18-          Luxexpo,
                   April 18-20, 2007, Luxexpo, Luxembourg
The architecture of the created system

Opthalmologist
                     WEB
                   interface

                                                                                       MATLAB
                                                                                     COMPUTING
                  MATLAB
                                                                                       ENGINE
                 Application



                                                                                      MATLAB
                                                                                     Application
                                     Internet

                                                                                     MATLAB
                                                                                    WEB SERVER
                                                                        Database:
                                                                            text,
Cardiologist                                                             images,
                                                                          signals

                                                                                        PHP
                   WEB
                 interface



                                                                                                   15
                                     18-          Luxexpo,
                               April 18-20, 2007, Luxexpo, Luxembourg
• Complete framework for eye fundus image processing was
implemented in eHealth diagnostic system using Web
technologies.
technologies



Web site:




                                                           16
Three CME courses developed in colaboration
with Distance Learning Center:
•Diagnostic and Treatment of Eye Diseases
•New Diagnostic Methods in Ophthalmology
•Eye Diseases for General Practitioners




                                              17
 Conclusions

• Tools and services for eye fundus image analysis developed.
• Combination of IT and signal processing techniques into Web
  based services for health system can help with better
  diagnosis and even acquire new clinical knowledge.

• Web based tools give more flexibility for developers and
  medical practitioners: compatibility for different computer
  platforms, better service and response to needs, better
  access using available internet connection.



                                                                18
                           18-          Luxexpo,
                     April 18-20, 2007, Luxexpo, Luxembourg
Thank You for Your attention!

         http://tmc.kmu.lt




                                                 19
              18-          Luxexpo,
        April 18-20, 2007, Luxexpo, Luxembourg

				
DOCUMENT INFO
Shared By:
Categories:
Stats:
views:5
posted:9/9/2010
language:English
pages:19
Description: Outcomes and trends of Lithuanian Swedish telemedicine collaboration fundus