Room: AAPM ePoster Library
Purpose:
Liver iron concentration is often used as an indication for total body iron load. This iron concentration is correlated with T2* value of the liver from an MRI study. For this purpose patients with transfusion-dependent disease often undergo MR T2* mapping during treatment. The purpose of this study is to evaluate the T2* measurement accuracy from MR scanners of different vendors and with different processing software.
Methods:
We scanned a FerriScan phantom consisting of 30 tubes with solutions of different T2* values on a Siemens 1.5T and a GE 1.5T system. The phantom experiment was performed twice to check consistency. The T2* mapping were performed by using a multiecho gradient echo sequence with a multichannel phase array body coil. Siemens generates T2* map from in-line processing. The acquired T2* images were sent to a GE advantage workstation (AW) to calculate T2* and R2* maps with T2* curves for selected ROI. In addition, T2* curve is fitted and T2* map is generated by a custom-made Matlab software. Both GE AW and matlab perform three-parameter fitting to a single exponential decay model. We also evaluated patient images and the resulted T2* values.
Results:
GE workstation and the matlab code generate consistent results, with good correlation coefficient in the T2* curves fitting and with good consistency between the repeated phantom studies. T2* values measured from Siemens in-line processing was significantly different from other two methods.
Conclusion:
This study demonstrated inconsistent T2* values obtained between scanners from different vendor and software. Although the original signal acquisition could play a role in the measurement, algorithms for the model fitting are unknown and significantly contribute to the inconsistency. Reporting T2* values along with measured signals and fitted curve is recommended to provide confidence to the clinicians.
Not Applicable / None Entered.
Not Applicable / None Entered.