Four-and-one-half years' experience in monitoring of reproducibility of an MR spectroscopy system--application of in vitro results to interpretation of in vivo data

Abstract

The primary purpose of this work was to assess long-term in vitro reproducibility of metabolite levels measured using 1H MRS (proton magnetic resonance spectroscopy). The secondary purpose was to use the in vitro results for interpretation of 1H MRS in vivo spectra acquired from patients diagnosed with Canavan disease. 1H MRS measurements were performed in the period from April 2006 to September 2010. 118 short and 116 long echo spectra were acquired from a stable phantom during this period. Change-point analysis of the in vitro N-acetylaspartate levels was exploited in the computation of fT factor (ratio of the actual to the reference N-acetylaspartate level normalized by the reciprocity principle). This coefficient was utilized in the interpretation of in vivo spectra analyzed using absolute reference technique. The monitored time period was divided into six time intervals based on short echo in vitro data (seven time intervals based on long echo in vitro data) characterized by fT coefficient ranging from 0.97 to 1.09 (based on short echo data) and from 1.0 to 1.11 (based on long echo data). Application of this coefficient to interpretation of in vivo spectra confirmed increased N-acetylaspartate level in Canavan disease. Long-term monitoring of an MRS system reproducibility, allowing for absolute referencing of metabolite levels, facilitates interpretation of metabolic changes in white matter disorders.

Figure 1

Water‐scaled NAA levels obtained from short TE ${}^1\mathrm{H}$ MRS in vivo spectra (a), water‐scaled NAA levels obtained from long TE ${}^1\mathrm{H}$ MRS in vivo spectra (b), NAA levels determined based on absolute standard (before application of $f_T$ coefficient) from short TE ${}^1\mathrm{H}$ MRS in vivo spectra (c), and NAA levels determined based on absolute standard (before application off $f_T$ coefficient) from long TE ${}^1\mathrm{H}$ MRS in vivo spectra (d). Full triangles represent two cases diagnosed with Canavan disease, open triangles represent patients diagnosed with other neurological disorders, dashed lines indicate the threshold of ‘mean + 2 SD’ computed separately for patients aged 2‐12 months and 12‐24 months. The NAA concentrations were not corrected for relaxation and are expressed in institutional units.

Figure 2

Temporal variability of metabolite levels obtained from ${}^1\mathrm{H}$ MRS in vitro measurements: Cho (a), Cr (c), and NAA (e) in vitro levels corrected for TG, R1, and R2; and Cho (b), Cr (d), and NAA (f) in vitro levels obtained using a water‐scaling technique. Dots represent short TE data, while circles represent long TE data. The metabolite levels are expressed relative to their mean levels in the period from April 2006 to September 2010.

Figure 3

Relationship between CV and SNR for NAA, Cho, and Cr in vitro levels. Dots represent data corrected for TG, R1, and R2 and triangles represent data obtained using water‐scaling technique. Both short and long TE in vitro data are included (Table 1). Statistically significant relationship $(CV=\frac{106.9}{SNR}+1.2,p<0.01)$ was found between CV and SNR in water‐scaling technique.