OBJECTIVE: Reproducibility is an essential strength of any diagnostic technique for cross-sectional and longitudinal works. To determine in vivo short-term comparatively, the test-retest reliability of magnetic resonance spectroscopy (MRS) of the brain was compared using the manufacturer's software package and the widely used linear combination of model (LCModel) technique. METHODS: Single-voxel H-MRS was performed in a series of patients with different pathologies on a 1.5 T clinical scanner. Four areas of the brain were explored with the point resolved spectroscopy technique acquisition mode; the echo time was 35 milliseconds and the repetition time was 2000 milliseconds. We enrolled 15 patients for every area, and the intra-individual variations of metabolites were studied in two consecutive scans without removing the patient from the scanner. Curve fitting and analysis of metabolites were made with the software of GE and the LCModel. Spectra non-fulfilling the minimum criteria of quality in relation to linewidths and signal/noise ratio were rejected. RESULTS: The intraclass correlation coefficients for the N-acetylaspartate/creatine (NAA/Cr) ratios were 0.93, 0.89, 0.9 and 0.8 for the posterior cingulate gyrus, occipital, prefrontal and temporal regions, respectively, with the GE software. For the LCModel, the coefficients were 0.9, 0.89, 0.87 and 0.84, respectively. For the absolute value of NAA, the GE software was also slightly more reproducible than LCModel. However, for the choline/Cr and myo-inositol/Cr ratios, the LCModel was more reliable than the GE software. The variability we have seen hovers around the percentages observed in previous reports (around 10% for the NAA/Cr ratios). CONCLUSION: We did not find that the LCModel software is superior to the software of the manufacturer. Reproducibility of metabolite values relies more on the observance of the quality parameters than on the software used.
OBJECTIVE: Reproducibility is an essential strength of any diagnostic technique for cross-sectional and longitudinal works. To determine in vivo short-term comparatively, the test-retest reliability of magnetic resonance spectroscopy (MRS) of the brain was compared using the manufacturer's software package and the widely used linear combination of model (LCModel) technique. METHODS: Single-voxel H-MRS was performed in a series of patients with different pathologies on a 1.5 T clinical scanner. Four areas of the brain were explored with the point resolved spectroscopy technique acquisition mode; the echo time was 35 milliseconds and the repetition time was 2000 milliseconds. We enrolled 15 patients for every area, and the intra-individual variations of metabolites were studied in two consecutive scans without removing the patient from the scanner. Curve fitting and analysis of metabolites were made with the software of GE and the LCModel. Spectra non-fulfilling the minimum criteria of quality in relation to linewidths and signal/noise ratio were rejected. RESULTS: The intraclass correlation coefficients for the N-acetylaspartate/creatine (NAA/Cr) ratios were 0.93, 0.89, 0.9 and 0.8 for the posterior cingulate gyrus, occipital, prefrontal and temporal regions, respectively, with the GE software. For the LCModel, the coefficients were 0.9, 0.89, 0.87 and 0.84, respectively. For the absolute value of NAA, the GE software was also slightly more reproducible than LCModel. However, for the choline/Cr and myo-inositol/Cr ratios, the LCModel was more reliable than the GE software. The variability we have seen hovers around the percentages observed in previous reports (around 10% for the NAA/Cr ratios). CONCLUSION: We did not find that the LCModel software is superior to the software of the manufacturer. Reproducibility of metabolite values relies more on the observance of the quality parameters than on the software used.
Authors: Melissa E Murray; Scott A Przybelski; Timothy G Lesnick; Amanda M Liesinger; Anthony Spychalla; Bing Zhang; Jeffrey L Gunter; Joseph E Parisi; Bradley F Boeve; David S Knopman; Ronald C Petersen; Clifford R Jack; Dennis W Dickson; Kejal Kantarci Journal: J Neurosci Date: 2014-12-03 Impact factor: 6.167
Authors: Nicolas Fayed; Javier Garcia-Campayo; Rosa Magallón; Helena Andrés-Bergareche; Juan V Luciano; Eva Andres; Julián Beltrán Journal: Arthritis Res Ther Date: 2010-07-07 Impact factor: 5.156
Authors: Daniel Ta; Abdullah Ishaque; Ojas Srivastava; Chris Hanstock; Peter Seres; Dean T Eurich; Collin Luk; Hannah Briemberg; Richard Frayne; Angela L Genge; Simon J Graham; Lawrence Korngut; Lorne Zinman; Sanjay Kalra Journal: Neurology Date: 2021-06-14 Impact factor: 11.800
Authors: Alice M S Durieux; Jamie Horder; M Andreina Mendez; Alice Egerton; Steven C R Williams; C Ellie Wilson; Debbie Spain; Clodagh Murphy; Dene Robertson; Gareth J Barker; Declan G Murphy; Grainne M McAlonan Journal: Autism Res Date: 2015-08-20 Impact factor: 5.216
Authors: Nicolás Fayed; Yolanda Lopez Del Hoyo; Eva Andres; Antoni Serrano-Blanco; Juan Bellón; Keyla Aguilar; Ausias Cebolla; Javier Garcia-Campayo Journal: PLoS One Date: 2013-03-25 Impact factor: 3.240