TE = 32 ms vs TE = 100 ms echo‐time ¹H‐magnetic resonance spectroscopy in prostate cancer: Tumor metabolite depiction and absolute concentrations in tumors and adjacent tissues

Abstract

PurposeTo compare the depiction of metabolite signals in short and long echo time (TE) prostate cancer spectra at 3T, and to quantify their concentrations in tumors of different stage and grade, and tissues adjacent to tumor.Materials and MethodsFirst, single‐voxel magnetic resonance imaging (MRI) spectra were acquired from voxels consisting entirely of tumor, as defined on T 2‐weighted and diffusion‐weighted (DW)‐MRI and from a biopsy‐positive octant, at TEs of 32 msec and 100 msec in 26 prostate cancer patients. Then, in a separate cohort of 26 patients, single‐voxel TE = 32 msec MR spectroscopy (MRS) was performed over a partial‐tumor region and a matching, contralateral normal‐appearing region, defined similarly. Metabolite depiction was compared between TEs using Cramér‐Rao lower bounds (CRLB), and absolute metabolite concentrations were calculated from TE = 32 msec spectra referenced to unsuppressed water spectra.ResultsCitrate and spermine resonances in tumor were better depicted (had significantly lower CRLB) at TE = 32 msec, while the choline resonance was better depicted at TE = 100 msec. Citrate and spermine concentrations were significantly lower in patients of more advanced stage, significantly lower in Gleason grade 3+4 than 3+3 tumors, and significantly lower than expected from the tumor fraction in partial‐tumor voxels (by 14 mM and 4 mM, respectively, P < 0.05).ConclusionCitrate and spermine resonances are better depicted at short TE than long TE in tumors. Reduction in these concentrations is related to increasing tumor stage and grade in vivo, while reductions in the normal‐appearing tissues immediately adjacent to tumor likely reflect tumor field effects. J. Magn. Reson. Imaging 2015;42:1086–1093.

Diffusion MRI and MR spectroscopy reveal microstructural and metabolic brain alterations in chronic mild stress exposed rats: A CMS recovery study

“…The lower the value of CRLB, the higher the accuracy of the estimated concentration of metabolites. The present study considered only those metabolites having CRLB < 20%, for further statistical analysis, as this threshold is widely used in clinical (Basharat et al, 2015; Järnum et al, 2011) and preclinical studies (Kumar et al, 2012; Mlynárik et al, 2008; Tkáč et al, 2004).…”

Supervised risk predictor of central gland lesions in prostate cancer using 1H MR spectroscopic imaging with gradient offset‐independent adiabaticity pulses

“…Due to the adiabatic pulses and shorter TE used in our study (88 msec vs. 120–145 msec), a relatively more intense citrate signal is recorded. For the singlet signal ratio Cho/Cr in the normal central gland, we recorded an average value of 1.37 ± 0.76 compared with the value of 1.70 ± 0.89 reported for 1.5T, but in another study this ratio was about 0.6 at 3T, which may be due to the use of different signal quantification and TE value.…”

“…9 Due to the adiabatic pulses and shorter TE used in our study (88 msec vs. 120-145 msec), a relatively more intense citrate signal is recorded. For the singlet signal ratio Cho/Cr in the normal central gland, we recorded an average value of 1.37 AE 0.76 compared with the value of 1.70 AE 0.89 reported for 1.5T, 34,35 but in another study this ratio was about 0.6 at 3T, 31 which may be due to the use of different signal quantification and TE value. The significance to separate cancer voxels from normal voxels in the central gland using the (Cho + Cr)/Cit or Cho/Cr ratio was similar or better than what has been published elsewhere using an endorectal coil.…”

Uncertainties of calculated Cramér‐Rao lower bounds: implications for quantitative MRS

“…9,11 The rejection threshold, above which the estimate is considered too uncertain, is currently set at 50%, 11 but is often reduced to either 20% or 30% for quantitative MRS purposes. [12][13][14][15] Several researchers have raised issues over the use of rCRB for quantitative MRS purposes. For example, Tisell et al 16 highlighted that the rCRB is boosted for low-concentration metabolites and thus might not be a suitable quality indicator.…”