Vous êtes sur la page 1sur 17

IMPACT BUSINESS INFORMATION SOLUTIONS, INC.

IBIS

Medical Image Data Management: Clinical Research

October, 2011
Agenda  

  Introduc)on  

  Medical  Image  Data  Overview  

  Clinical  Research,  Drug  Development  Overview  

  Tradi)onal  Technology  Overview  

  Current  State  of  Play  


  Evolu)ons  and  Milestones  
  Use  Cases  

  Road  Map  for  the  Future  


Introduction  
  Impact  Business  Informa)on  Solu)ons  (IBIS)  
  Image  Data  Management  Solu)ons  Provider  
  IBISimg  
  IBISecg  

  Tony  O’Sullivan  
  VP  Opera)ons,  IBIS  

  Contact  Informa)on  
  12  Roszel  Road,  Suite  A207,  Princeton,  NJ  
  hQp://www.ibisworks.com  
  609.520.8502  
Medical    Image  Data:  Basics  
  DICOM  Standard  
  MRI  (Magne)c  Resonance  Imaging)  
  CT  (X-­‐ray  computed  tomography)  
  PET  (Positron  emission  tomography)  
  Ultrasound  
  etc.  

  File  structure  
  Mul)ple  files  (slices)  per  scan  
  Each  file  contains  image  plus  DICOM  
“header”  (metadata)  
  Not  fully  self-­‐describing  -­‐  addi)onal  
metadata  may  be  required  
  Wide  variety  in  number  of  files  per  scan,  
and  in  file  size  

  Tradi)onal  file  storage  


  Picture  Archive  and  Communica)on  
(PACS)  system  
  File  based  storage  
Drug  Development  Overview  

DRUG  DISCOVERY   PRE-­‐CLINICAL   CLINICAL  TRIALS   FDA  REVIEW  

10,000   5   1  
250  
Compounds  
Traditional  Clinical  Trial  Process  
Pharma  Company/Sponsor   CDMS  

1.  Ini@ate  Clinical  Trial  (Study)  


2.  Define  Study  Specifica@on  
3.  Set  up  Sites/Subjects  
11.  Load  to  CDMS  

10.  Receive  results  (HD,  DVD,    


STUDY  SPEC  
FILE  SERVER                Yp,  email,  etc)  
Clinical  Research  Organiza@on  (CRO)  
   9.  Transmit  results  (ASCII)  to    
4.  Agree  Study  Specifica@on              sponsor  
5.  Conduct  Study  at  Sites  
   8.  Receive  and  process  
STUDY  SPEC  

Clinical  Trial  Sites  


7.  Transmit  scans  to  CRO  
6.  Generate  scans  
Issues  with  Traditional  Clinical  Trial  Process  

  Business  Process  Issues  


  Manual  process  with  mul@ple  poten@al  points  of  failure  
  No  systema@zed/standardized  data  management  process  

  Data  Access/Review  Issues  


  Image  data  are  stored  at  Core  Labs,  not  readily  accessible  
  Access  can  be  expensive  
  Data  re-­‐purposing  can  be  difficult/impossible  

  Data  Security  Issues  


  Non-­‐secured  data  transfer    
  No  audit  trail  

  Data  Quality  Issues  


  Dependent  on  the  Core  Lab  
  Not  possible  to  verify    
  No  mechanism  for  enforcement  of  standards  
Evolution:  1  
  Up  to  2006  
  Tradi)onal  process  
  No  ISV  clinical  research  solu)ons  for  medical  image  data  
management  available;  only  hospital  PACS  systems  

  2006  –  07  
  Web-­‐based  solu)on  
  Ownership  of  the  data    
  Secure  Image  Data  Transfer  over  HTTP(S)  
  Applet-­‐based  
  Standards-­‐based  manifest  file  (e.g.  CDISC  ODM)  
  Data  Quality:  rules-­‐based,  repeatable  
  Integra@on  with  CDMS  
  Integra@on  with  DICOM  worksta@ons  for  image  review  
  Image  data  storage  on  the  file  system  
Evolution:  2  
  2008  
  Oracle  11g:  storage  of  the  image  data  in  the  database  using  Oracle  
Mul)media  DICOM    
  No  more  file  system.  Storage  of  the  image  data  in  DICOM  objects  on  the  
database  
  Enhanced  quality  conformance  checking  
  Streamlined  data  backup/archive  process  

  2009-­‐10  
  SOA-­‐based  Web  Services  
  Comprehensive  REST  API  for  data  upload,  normaliza@on,  valida@on,  
download,  and  review  
  SOA-­‐based  Adapters  with  JMS  Messaging  
  Integra@on  with  3rd  party  systems,  DICOM  Worksta@ons  
  Automated  analysis  of  image  data  with  plug  and  play  algorithms  
  Automated  alerts  for  outliers  or  specified  criteria;  mobile  review  
 Validation/Performance  Test  
  2011  Performance  Test  Summary  
  Measured  five  major  processes  using  REST  Services:  data  upload,  extract/
normalize,  validate,  download,  view  
  Representa@ve  sample  of  image  data  from  five  modali@es  –  MR,  PET,  CT,  
Ultrasound  and  Mammography  
  Increased  user  connec@ons  from  25  to  50  to  100  to  800  as  we  increased  
database  nodes  from  1  to  2  to  4  

  Results  Summary  
  Scalability  
  From  25  to  800  concurrent  users  –  almost  zero  performance  degrada@on  
  From  one  to  four  database  nodes,  performance  gain  averaged  250%  
  Data  Throughput  
  Over  800%  faster  than  equivalent  tests  on  commodity  hardware  (1  TB  of  image  
data  uploaded  in  under  1  hour)  
  Performance  
  DICOM  Viewer:  Sub-­‐second  @me  to  first  image  over  DICOM  protocol  for  PET,  CT,  
MR  
  As  user  connec@ons  are  increased,  user  response  @mes  improve  consistently  
Use  Cases  

  Automated  Analysis  
  Automated  algorithmic  image  processing  
  Plug  and  play  algorithms  

  Alerts  Use  Case  


  Configurable  alert  criteria  
  Alerts  via  email,  text  message  (SMS)  

  Physician  mobile  review  (iPad,  droid)  


  Remote  pa@ent  review/browser  
  Remote  image  review,  e.g.  Osirix  on  iPad  
Use  Case  #1:  Physician  Mobile  Review    
Research  Facility,   Medical Image Data Management System
Hospital,  etc  
Software Component:
Oracle 11g
DICOM Listener Service

3   -­‐  Storage  
Scan   4.1   -­‐   QC  
Medical Image
4.2  Scan  Status   Repository
Software Component:
Scan Services 7.2.3.1  retrieve  
1  
2   5  
Scan  Message   Data/logs
Vendor  PACS   Repository
System  (GE,   JMS  Bus  
Siemens,  Agfa,   7.2.3  
etc.)  
6.1   6.2  

7.2.1  Alert      

Adapter Service 7.2.2  “Retrieve  scan”        


Alerts/Study Adapter
Software Component: 7.2.4  View  Scan        
Tracking Service

7.1  
Use  Case  #2:  Automated  Analysis  
Research  Facility,   Medical Image Data Management System
Hospital,  etc  
Software Component:
Central Repository
DICOM Listener Service

3  
-­‐  Storage  
Scan   4   -­‐   QC  
4  Scan  Status   Medical
Software Component: Image Repository
7.2.2  retrieve  
Scan Services
8.3  store  
1  
5   Data/logs
2   Scan  Message   7.2.1/7.2.3  Retrieve/Send  Scan  for  Analysis  
Vendor  PACS   Repository
System  (GE,   JMS  Bus  
Siemens,  
Agfa,  etc.)  
6.1   6.2  
8.2  Return  result  scan(s)        

7.2.4  Send  for  Analysis   3rd party


Adapter Services Image
Auto Analysis Adapter Analysis
Software Component:
8.1  Receive  result  scan(s)         Algorithm
Tracking Service

7.1  
Evolution:  3  -­‐  Current/Future  
  2011  
  Grid:  Distributed  compu)ng  for  Analy)cs  
  DICOM  Proxy  
  Seamless  image  transfer  from  source  with  mul@ple  rou@ng  op@ons  
  Automated  distribu@on  of  data  quality  specifica@ons  and  checking  at  the  source  

  2012  and  beyond  


  Cross-­‐over  to  Healthcare:  Integra)on  with  large  enterprise  solu)ons  such  as  
Oracle  HIE    
  Web  Service  Orchestra@on  /  BPEL  enablement  
  Master  Person  Index  integra@on  
  etc.  

  Web  Services  related  capability  


  Mashups  
  Cloud  
  SaaS  
Platform  Overview:  Current    
Image Data Providers (Upload)
Web CDISC SFTP/ DICOM Remote Folder
  Highlights  
Web UI
Services ODM FTP Protocol Share Scan   Enterprise  plamorm  for  medical  
image  innova@on  
  Vendor  Neutral  Archive  with  
SOA: Web Services Layer integra@on  of  remote  
Security: Authorization, Data Access Control repositories  
Core Data: Study, Patient, Scan, etc
  Robust  SOA  Web  Services  layer  
for    secure  image  data  
         JMS  Bus  
opera@ons  
•  DICOM Conformance   Suite  of  SOA-­‐based  Adapters  for  
•  Workflow 3rd  party  integra@on,  
VNA   •  Event-based messaging automated  analysis,  DICOM  
worksta@on  integra@on,  etc.  
SOA: Adapter (Services) Layer   Enterprise  Architecture  using  
DICOM Protocol (Push/Pull), Remote Review, Analysis Oracle  Weblogic  Server  11g,  
Oracle  Database  11g    

Image Data Consumers


DICOM Analysis Mobile Image Data
Web UI
Workstation Algorithms Review Archive
Platform  Overview:  Road  Map  
Image Data Providers (Upload) Enterprise Integration
REST CDISC SFTP/ DICOM Remote Folder
Web UI
Services ODM FTP Protocol Share Scan
Workflow  Orchestra@on  /  BPEL  

Security  Profiles  
SOA: Web Services Layer
Master  Person  Index  
Security: Authorization, Data Access Control
Core Data: Study, Patient, Scan, etc

         JMS  Bus   Exalogic  

•  DICOM Conformance
•  Workflow
Data/DICOM   •  Event-based messaging
Repository  
Exadata  

SOA: Adapter (Services) Layer


DICOM Protocol (Push/Pull), Remote Review, Analysis Grid  compu@ng/Distributed  Analy@cs  
Site  Integra@on  

Image Data Consumers


DICOM Mobile Analysis Image Data
Web UI
Workstation Review Algorithms Archive
Wrap  up  

  Q  &A  

Vous aimerez peut-être aussi