Horizon Information Portal

MSOE Library Home

MSOE Library Home

Home

Software development and automation of the American College of Radiology MR Phantom Testing Accreditation Criteria for the General Electric Magnetic Resonance Imaging Whole Body Scanner / by Steven J. Huff.

by Huff, Steven J.

Magnetic resonance imaging

Phantoms (Radiology)

Automatic data collection systems

2 v., 428 leaves : ill. ; 29 cm.

Advisor: Dr. Henry Welch.

Committee members: Dr. Christopher Taylor, Dr. Lisa Milkowski.

Introduction and background -- Review of literature -- Design methods and implementation -- Data collection results and discussion -- Forced failure modes testing -- Conclusions and recommendations -- References -- Bibliography.

This MSE Graduate Capstone project describes the design, development, and validation of a software tool-set for the automated data collection and analysis of the American College of Radiology (ACR) Magnetic Resonance Imaging (MRI) Phantom Accreditation testing criteria. ACR accreditation is becoming increasingly more important in the MRI clinical community as well as in the health care insurance industry. ACR accreditation assures that clinical MRI scanners meet a set of minimum quality standards and that those standards are maintained over the operating life of the MRI scanner.

The software tool designed and developed for this project is called the American College of Radiology - Phantom Analysis (ACR-PA) tool. The ACR-PA tool was designed to allow any knowledgeable user of the General Electric (GE) MRI scanner to automatically collect and analyze ACR phantom test data in an automated reliable manner in a reasonable amount of time. The data generated from this software can be used to "pre-qualify" a GE MRI scanner for ACR accreditation or could be used for Daily Quality Assurance (DQA) testing. The ACR-PA tool was designed to replace a very manual, tedious, and operator dependent / subjective testing procedure for acquiring and analyzing ACR phantom test data. The ACR-PA tool automates the data acquisition (scanning) and image analysis portions of the ACR phantom test procedure.

Literature research was done to search for any similar software tool set and to get historical information on the ACR MRI Accreditation Program. The ACR-PA tool was designed using OOD-UML techniques and used the rapid prototyping software development approach. The milestone accomplishments for this project were as follows:

1. Requirements Gathering and Critical to Quality parameter identifications. 2. Top-level software design using UML practices and methods. 3. Design and assembly of a prototype ACR phantom holder. 4. Design and implementation of the GUI interface and workflow. 5. Design and implementation of the data acquisition software. 6. Design and implementation of the image analysis software modules: a) Geometric Distortion Test b) High Contrast Spatial Resolution Test c) Slice Thickness Accuracy Test d) Slice Position Accuracy Test e) Image Intensity Uniformity Test f) Percent Image Ghosting Test g) Low Contrast Object Detectability Test. 7. Verification testing and data collection.

Data were collected and analyzed on a number of GE MRI systems at GE Medical Systems headquarters in Waukesha, Wisconsin and at hospitals and clinics in the local Milwaukee metro area. The data show that the ACR-PA tool is a fast and reliable method for collecting and analyzing ACR phantom image quality data. Total execution time of the ACR-PA tool is approximately 20 minutes as compared to several hours for the currently existing manual measurement method. Data generated by the ACR-PA tool were verified against independent manual testing via a formal Verification Test Plan and showed good correlation between the automated image analysis results and manual measurement techniques.

Horizon Information Portal 3.25_9885

Powered by SirsiDynix

© 2001-2013 SirsiDynix All rights reserved.

Horizon Information Portal