Juha Peltonen MSc, Presenter: Nothing to Disclose

Teemu Makela, Abstract Co-Author: Nothing to Disclose

Alexey Sofiev, Abstract Co-Author: Nothing to Disclose

Eero Salli, Abstract Co-Author: Nothing to Disclose

The automatic daily quality control system provides an effective tool for simultaneous tracking of multiple quality control parameters of a large number of MRI scanners. With accurate up-to-date and long-term data certain hardware problems can be sorted out effectively.

A state-of-the-art system was developed to perform effective daily MRI quality control in a large hospital district. In total, 13 geographically scattered MRI scanners from 3 vendors were included in the quality control system. Due to the large number of scanners, the image analysis had to be automated to obtain up-to-date as well as long-term statistical data in easily interpretable format.

A homogeneous test phantom with a circular cross section is imaged every morning at each MRI site by the local radiographer. This image is sent to an image processing server using DICOM transfer protocol. The analysis is performed by in-house developed software based on Insight Segmentation and Registration Toolkit (National Library of Medicine, US). Monitored parameters include signal-to-noise ratio, image uniformity, ghosting ratio and the scanner centre frequency. The automatically generated results are presented as online graphs on an internal web page showing the temporal development of the quality parameters. Additionally, graphical MATLAB (MathWorks, Natick, MA) program is used for further data analysis and baseline determination. The time series of the four quality control parameters were successfully produced for all scanners. Considerable deviations from the expected long-term evolution were observed in several cases. In one case, simultaneous anomalies in signal-to-noise and ghosting ratios were verifiably caused by a faulty gradient amplifier.  

A subtle drift in scanner hardware performance can remain unnoticed until the defects are clearly visible in clinical images. With continuous analysis and clear representation of the daily quality control results these tendencies can be detected effectively. Resolving of error sources can begin before significant impact on patient studies. The described system could allow shorter response times to hardware problems in a large cohort of MRI scanners.

Peltonen, J, Makela, T, Sofiev, A, Salli, E, Automatic Daily MRI Quality Control Analysis System Applied in a Large Hospital District.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14005524.html