TITLE:
Optimization of Liver Iron Load Assessment by Pixel-Based T2* MRI in Thalassemic Patients
AUTHORS:
Neda Gholizadeh, Shahram Akhlaghpoor, Hamid Soltanian-Zadeh, Emad Fatemizadeh, Mohammad Reza Alinaghizadeh, Amin Shams Akhtari, Mohammad Nourizadeh
KEYWORDS:
Thalassemia; Liver; Iron Overload; MRI; T2*
JOURNAL NAME:
Open Journal of Radiology,
Vol.2 No.2,
June
28,
2012
ABSTRACT: Purpose: To improve liver iron load assessment by investigating the precision of different approaches of T2* Measurement. Background: Iron overload is a major problem in the treatment of thalassemic patients. Liver iron concentration (LIC) is an important index toward the management of body iron load. The accuracy of iron load estimation may suffer from the methodology of T2* measurement and there is no complete agreement upon the best approach of T2* calculation. Methods: 32 β-thallasemic patients (18 male) with the mean age of 20.0 ± 6.5 years were involved in this study. A multi-echo fast gradient-echo technique on a 1.5 T MRI system was used to measure liver iron overload and the T2* map of liver was reconstructed on a pixel-by-pixel basis. The T2* map and MRI images were utilized to deter- mine accurate location of ROI (region of interest). The mean of T2* were computed from the ROIs. The reproducibility of calculated T2* values in two methods were obtained. Moreover, the mean of the pixel’s T2* was calculated in the entire liver parenchyma of one slice. The T2* value of the entire slice was compared with the ROI approach. Results: In the ROI based method, the CoV for the intra-observer reproducibility was 8.5% and for the inter-observer was 9.78%. In the pixel based method, the CoVs for intra-observer and inter-observer reproducibility were 2.79% and 3.91%. There was an acceptable correlation (r = 0.96) between the T2* values calculated by the ROI and the entire slice. Conclusions: The pixel-based approach is more precise to determine the appropriate placement of the ROI. The assessment of T2* in the entire slice reduces the user-based errors significantly.