Related Experiment Videos
Reproducibility of metabolite concentration evaluated by intraclass correlation coefficient using clinical MR
Hitoshi Kubo1, Masafumi Harada, Minoru Sakama
1Department of Radiologic Techology, School of Health Sciences, University of Tokushima, Tokishima, Japan. kubo@medsci.tokushima-u.ac.jp
Objective:
The purpose of this study was to evaluate the reproducibility of in vivo quantitative proton MR spectroscopy (MRS) by repeated-measure analysis of variance and intraclass correlation coefficient (ICC) including assessment of both inter- and intrasubject variation.
Methods:
The concentration of metabolites was quantified using long-TR and short-TE stimulated-echo acquisition mode sequences with a quadrature head coil. All preparation procedures for MRS measurement except volume of interest determination were done by an automatic preparation scanning system equivalent to a clinical setting. Analyses were conducted by an LCModel running based on our original basis set by two different reference methods.
Results:
The ICC showed a good correlation (r = 0.61-0.99) in any metabolite and almost the same value between the two different reference methods, but the concentration of myoinositol had a large intersubject variation.
Conclusion:
Our results suggest that the reproducibility of quantified metabolite concentration is acceptable in the clinical setting regardless of the reference method, although quantified values may have different intersubject variations.