Stefania Montemezzi MD, Presenter: Nothing to Disclose

Francesca Caumo MD, Abstract Co-Author: Nothing to Disclose

Ilaria Baglio, Abstract Co-Author: Nothing to Disclose

Lucia Camera, Abstract Co-Author: Nothing to Disclose

Gabriele Meliado, Abstract Co-Author: Nothing to Disclose

Carlo Cavedon DPhil, Abstract Co-Author: Nothing to Disclose

The study was aimed at improving the feasibility of total choline (tCho) detection in breast lesions and at estimating sensitivity and specificity of breast 3T-MR spectroscopy (MRS) to aid MR-based diagnosis of malignancy.

141 patients (157 lesions, range 0.05-108.86 cm3, mean 6.62 cm3) were enrolled (21-84 yrs, mean 58.5 yrs). All patients had breast abnormalities on mammography or sonography, confirmed by cytology and/or micro-biopsy. Single-voxel MRS was performed by means of a Philips Achieva STx 3.0T scanner. First-order pencil-beam shimming was used on a 15.6 cm3 volume centred on the region of interest (ROI), which ranged 0.34-8.0 cm3 (mean 1.33 cm3). MRS used TE=135ms, TR=3000ms, 128 samples, water (window 140Hz) and fat (SPAIR, offset 80Hz) suppression. When possible, MRS was performed before contrast agent injection and repeated thereafter. Pre-saturation was used to suppress signal from nearby regions. Local field homogeneity was evaluated by means of the FWHM of the unsuppressed water peak. A threshold was placed at 45Hz, above which MRS was not performed due to insufficient field homogeneity. tCho was estimated by means of the signal-to-noise ratio (SNR) of the peak at 3.2 ppm.

MRS was feasible in 89.5% of the lesions using pencil-beam shimming (mean FWHM of water peak 34Hz), compared to 54.2% (29Hz) when standard iterative shimming was used (first 80 patients). 59 lesions (52.2% of reliable spectra) showed detectable tCho (SNR 1.4-53.7, mean 8.5). Comparison with available histopathological examination of surgical specimens (or micro-biopsy for benign lesions) showed 87.5% sensitivity and 86.0% specificity. No correlation between lesion volume and SNR of the tCho peak was observed. Malignant lesions that showed no tCho had a volume of 0.7cc or less.

High-field MR spectroscopy is expected to improve SNR of the investigated metabolites, however field homogeneity is more difficult to achieve compared to 1.5T. The adjustment of the shimming process improved the fraction of cases for which high-field MRS resulted feasible. Further research is warranted to improve choline detectability and to confirm the observed sensitivity and specificity of the method.

MRS at 3T could improve the specificity of breast MR. Improving its feasibility is a key factor, however the possible correlation between tCho concentration and malignancy needs further investigation.

Montemezzi, S, Caumo, F, Baglio, I, Camera, L, Meliado, G, Cavedon, C, MR Spectroscopy of the Breast at 3 Tesla: A Clinical Experience.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13021676.html