MA McLean, AN Priest, I Joubert, DJ Lomas, MY Kataoka, H Earl, R Crawford, JD Brenton, JR Griffiths, E Sala
Magn Reson Med
(1)H-MRS was performed on 12 women (age range 45-72) with ovarian cancer of FIGO stage 3 or above using a 3T MRI system with an 8-channel cardiac receive coil. Respiratory-triggered PRESS-localized spectra (TE = 144 ms) were obtained separately from an ovarian mass and from metastatic disease. Peak areas were quantified relative to unsuppressed water using LCModel and spectra were discarded if LCModel reported signal-to-noise ratio (SNR) < 3 or if no metabolites were reported with standard deviation (SD) < 30%. The cystic fraction of each voxel was estimated by thresholding T(2)-weighted images, and this was used both to correct the reported metabolite concentrations and to calculate an expected SNR of choline using the measured SNR of water. Choline was detected in 10/12 primary tumors and 5/11 metastatic lesions (range 2.0-16.6 mM). Of the 8/23 failures, 7 had a predicted choline SNR < 2, confirming that the failure to detect choline could be explained by technical problems. Glycine was observed in one benign lesion. (1)H-MRS can be used to quantify choline in primary and metastatic masses in ovarian cancer, but the moderately high rate of failure to detect choline necessitates careful recording of data quality parameters to discriminate true from false negatives.