1. CTTI PROJECT Emory University, Quality Assurance and Review Center (QARC) and Washington University in St. Louis

2. Background  Quality Assurance Review Center (QARC): A research program within the University of Massachusetts Medical School (UMMS) that provides radiotherapy (RT), quality assurance (QA), diagnostic imaging data management, and clinical research support.  MAX: A flexible and all-inclusive platform developed by QARC in order to address all of the specific needs of the clinical cooperative groups and industry partners  Advanced Technology Consortium (ATC): A cooperative group supported by the National Cancer Institute to create a robust QA process to collect and review the image-based planning and verification data for patients enrolled on the 3D Oncology Group (3DOG) prostate dose- escalation protocol (RTOG 9406)  CERR is a MATLAB based, open source radiotherapy research environment which includes an extensive set of tools commonly found in radiotherapy planning systems, in addition to several review and analysis features. It has been developed by Joseph O.Deasy and his team at the Division of Bioinformatics & Outcomes Research, Washington University, St. Louis.

3. CTTI Overview  Extend existing review clients to support the retrieval over the grid and the subsequent display of radiation therapy treatment objects  Begin processes to bring legacy systems such as MAX into compliance with caBIG objectives and grid technologies

4. Architecture of grid-Enabled CERR caGrid enabled data collection in cooperative groups Data service for DICOM RT Objects Data service for CERR Objects CERR client capable of interacting with caGrid services Access data and store image review results at ATC, QARC, CALGB, ACRIN, NCIA archives Capable of executing MATLAB codes locally or remotely

5. Core Components Computational Environment for Radiotherapy Research (CERR) caGrid data services to store CERR Objects and DICOM RT Objects caGrid and In Vivo Imaging Middleware for query; high performance data transfer between clients and data repositories; authentication and authorization

6. Extensible Review Tool for Treatment Planning Data Images (T/S/C planes), Structures, Doses, DVHs Data Format Conversion RTOG/DICOM Import DICOM Export Support for multiple images series CT, MRI, PET Open-source (MATLAB) code available from http://radium.wustl.edu/CERR Free-standing version (“compiled” MATLAB) Computational Environment for Radiotherapy Research (CERR) — Deasy

7. A caGrid data service for storing CERR objects Stores the metadata of the CERR MATLAB objects in a XML database Provides Query and high performance data transfer from/to the client CERR client is modified to support grid interactions Grid based Query/Retrieve interface Grid security interface Grid Q/R GUI Components of Grid Enabled CERR

8. A caGrid data service for storing CERR objects Capabilities Query, Retrieve, Submit High performance bulk data transport Components Berkeley DB XML database caBIG In Vivo Imaging Middleware v1.2 caGrid v1.1 Security Ongoing work to add authentication and data level and role-based authorization framework Will leverage the caGrid security infrastructure CERR Data Service

9. Missing pieces needed for a production grade development and deployment  Security  Deploy the GAARDS security infrastructure at ATC and provide a policy and mechanism for remote users to request and receive grid security accounts  Extend grid-enabled CERR so that it can allow a remote reviewer to login and obtain grid credentials  Extend CERR Data Service to support service level authorization and role based data access control  CERR Information Model  Extend the data model used by CERR to allow for richer querying of data (CERR lite)  Extend the AIM information model to support annotations of RT objects and provide the ability to create AIM annotations from CERR

10. MAX – Recent Development  This past year the MAX database at QARC has undergone three critical upgrades to support Remote Users.  1. Migrated all existing data from Jet to enterprise level SQL Server 2005.  2. MAX has been made “Remote User” aware. Users are presented with an interface which “morphs” to provide them with exactly the functionality they require.

11. MAX – Recent Development  3. Terminal Services have been developed using Terminal Services Gateways to allow the Remote Desktop Protocol (RDP) on Port 3389 to be encapsulated in the internet standard HTTPS Port 443. This provides all of the functionality of MAX in a secure, server side solution. Users regularly log into MAX from remote sites and have full functionality using only an Internet Explorer Web Browser. (see image below)

12. MAX – Current Terminal Server Gateway Configuration

13. MAX – Key Technologies  The foundation of the grid enabled MAX is already established in the enterprise level, fundamental design of the SQL server tables and relationships.  The Dicommunicator is a technology built into MAX that manages all DICOM imaging. It is based on the DICOM standard study, series and image universal identifiers. It provides tools for viewing, adjusting and annotating DICOM images. 

14. MAX – Key Technologies  The Query-By-Form technology built into MAX provides users with a simple interface capable of querying virtually any data. Users are able to create complex queries that return clear results and reports.

15. MAX - Objectives  For the grid enabled MAX project the SQL Server tables and relationships will be migrated to the open-source database system MySQL.  The system will be federated with MAX at QARC to provide agile development and version control. As MAX evolves at QARC, new versions will be published to the remote sites.

16. Vision of Grid enabled MAX  Coordinate radiation therapy patient care and research at local sites using MAX derived system  Why?  Convergence of research and patient care data management, decision support and QC requirements  Commitment by many academic sites to involve most patients in a clinical trial or other translational study (e.g. TCGA)  MAX already has most of the functionality needed for such a system  Key requirements are grid enablement and development of tools and data models to support interoperability with local EMR and PACS systems  Result: tool that could support data exchange and decision support that would cross boundaries between local medical institutions, QARC, RTOG, ATC, CALGB, ACRIN