Background
2D L-COSY has been successfully applied in vivo to investigate metabolic profiles in brain1-5, muscle6, and breast7. To improve peak separation, most 2D L-COSY have been performed at 3T, however, there has been no test-retest repeatability study conducted at 3T to assess the precision of in vivo metabolic concentrations. The primary objective of this study was to evaluate the test-retest reliability of 2D L-COSY in detecting brain metabolites in a phantom and healthy controls (HC) in a 3T system.
Method
Ethics approval was sought prior to study. In-vitro evaluations were performed on brain phantom containing brain metabolites at physiological concentrations and pH. After informed consent, six HC had eight-weekly scans over a period of three months. To evaluate any inter-week effects on MRS, there was a one month interval between the first set and second set of four-weekly scans. 2D L-COSY was acquired from the posterior cingulate gyrus from a 27cm3 voxel. Raw spectral data was processed with Felix software1. Intra and inter-subject variability were measured. Test-retest repeatability was calculated using coefficient of variation (CV), standard error measurement (SEM) and intra-class correlation coefficient (ICC), using SPSS software.
Results
Twenty-two metabolites were studied. CV for in-vitro and in-vivo data ranged from 0.71% to 17.86% and 2.98% to 36.86%, respectively. N-acetyl aspartate, choline, creatine, lysine and phenylalanine were found to be consistently reliable in intra/inter subject measurements. The SEM for in vivo data ranged from 0.0016 to 0.08. The ICC was more than 0.95 for diagonal peaks while the ICC for cross peak were not less than 0.45, which reflects minimum variability between repeated measurements. The inter-week interval had minimal effect on metabolite ratios (p=0.08).
Conclusion
The low variability in metabolite concentration in this study shows a high level of reliability of 2D L-COSY in the human brain.
References