Transcript Graffigna

Medical Image Processing
Server using DICOM.
Barberis, Sebastián (1), Graffigna, Juan Pablo
(2), Isoardi, Roberto (1)
1 – Escuela de Medicina Nuclear. Mendoza. Argentina.
2 – Universidad Nacional de San Juan. San Juan. Argentina.
[email protected]
Problematic Situation



Algorithms developed by research groups
present inconvenients that make difficult their
validation and clinical implementation.
Highly trained medical and technical staff
usually work in very demanding
environments and cannot employ the
developed tools.
For the validation, a selection of studies and
its later processing is required. Actually,
validations are performed manually.
Objetive
–
–
–
The implementation of a centralized
processing server that allows solving
different needs in diagnosis and research.
The development of a DICOM system
interact within a PACS.
Quickly associate research processing
algorithms to clinical implementation,
allowing their validation.
Connecting to a PACS system
Diagnostic
Workstation
Storage
Server
PACS
Manager
HIS / RIS
Network Service
Workstation
Modalities
DICOM Printers
Processing
Server
Features







Communication of studies in clinical environments (DICOM).
Different algorithms of processing:
•
Intermodality/intrapatient registration for radiotherapy.
•
Intramodality/intrapatient registration for treatment
monitoring.
•
Processing of single studies for diagnosis.
•
Others.
Easy administration and set-up
Automatic Function.
Multi-thread.
Stability against failures.
Development with GPL tools.
Tools used






Support tools for design:
• Notation diagrams in Unified Modeling Language UML.
• Tool box BOUML® release 2.32
Develop Enviroment and compiler:
• Code::Blocks 8.02
• Compiler MinGW 3.4.5
DICOM facilities:
• CTN of Mallincrodt Institute of Radiology.
Database facilities.
•
MySQL 5.0 API for C.
Processing Algorithms
Research project.
Implementation model - Application Data Flow
Implementation of the DICOM standard

The whole communication and transfer of images with remote
application entities is carried out using the DICOM network
protocol using the TCP/IP protocol. The functions supported
by the entities of the system are:
SCU
(Application Entitys ‘Send’)
Storage.

SCP
(Application Entitys 1 to N)
Storage.
Verification.
UL services supported are: A-ASSOCIATE, A-RELEASE, AABORT, A-P-ABORT and P-DATA
Implementation of the DICOM standard


The reception Applications Entity in this system
support a great number of storage service class
(listed in its Conformance Statement) and the
verification service class.
As the communication module only supports the
storage and verification, the response to other
request (as posts C-FIND or C-GET) is reject whit
the corresponding response messages.
Features database

1) Contains the configuration parameters (available
processing algorithms, conditions for them,
assigned port, etc).
2) Served for recording studies information
(general and specific data).
Both sets of data are used to making decisions.

Decision-making system: SQL Query


Thread use


Allows opening a listening port for each different
processing algorithm (including their capacity of
solve one o more association request).
Allows the separation of tasks and avoid problems
that can halt its overall performance. The latter is
important because the imaging techniques can be
used in a development phase, causing instability in
the process
Test and application

This tool has been implemented in a Medical
Imaging and Radiotherapy center, “Escuela de
Medicina Nuclear, Mendoza, Argentina” including
three image processing algorithms.
• Quality control in computed tomography.
• Flip process to prepare PET studies for
Radiotherapy Planning.
• Cropping and segmentation of MRI brain
studies.
Test and application
Quality control in computed tomography
Cropping and segmentation
of MRI brain studies.
Flip process to prepare PET
studies for Radiotherapy
Planning.
Conclusions




This work has allowed the development of a
Processing Server in accordance with DICOM
standard.
It is possible to integrate new processing
algorithms to clinical environments.
The application allows an automatic management
of studies. The decision made determines which
operation must be executed, which studies take
part, and to whom the results must be sent.
The system can receive several studies, execute
algorithms, or send results simultaneously using
different processing threads. Operations, ports,
conditions, etc., can be quickly modified by means
of a SQL database.
Conclusions



For the design and the development of the server,
GPL software tools have exclusively been used.
A weakness of the system is the necessity of
employing automatic algorithms that require no
user intervention. Even though there is a tendency
to this kind of techniques, there is no possibility to
interact with the user during processing tasks.
Regarding the implementation of the system in a
clinical environment, it is necessary to execute it
for a prolonged time in order to evaluate the system
and algorithms. It will allow to analyze the
organizational changes.