General Overview
More than 20 years have passed since the first studies on clinical applications of magnetic resonance spectroscopy (MRS) in the breast were published (1, 2), but the role of MRS in breast diagnostics remains ancillary, if not at all absent, in the clinical routine (3).This is, most likely, due to the fact that MRS is a technically challenging sequence, which requires not only a specific expertise, but also a constant active and close cooperation between physicist, radiologists and technologists (4). In addition, the acquisition times of the sequence are relatively long, and the additional information acquired is often considered of limited value (5). Despite that, the research on MRS showed interesting advances in the last years.
Technique and Interpretation
Both 1.5T and 3T magnets can be used to perform MRS. The acquisition can be performed only on the target lesion (single voxel) or on a larger area in the breast (multivoxel), to evaluate also the surrounding tissue. In most of the analysis, single voxel sequences have been preferably used. MRS can be acquired using either point-resolved spectroscopy sequence (PRESS) or stimulated echo acquisition mode (STEAM) sequences. Both sequences provide similar results, with only subtle differences between the sequences (6, 7). The raw spectral data need to be processed in order to obtain the final spectra. Various filtering and other manipulations are often performed on the data, which may have quite profound effects on the quality of the final spectrum (6). Once the spectrum has been obtained, the various metabolites have to be identified and labelled. The MR spectrum can be analyzed in a qualitative (visual) or quantitative (peak area or signal to noise ratio) way (5). Initial studies focused mostly on the choline peak at 3.2 ppm, and some also measures the ratio between the water and fat peak (4.8 ppm and 1.3 ppm respectively) (7). Recently, following studies on the lipid metabolism in cancer, the interest of researchers started to focus also on the fatty acid peaks. Both ex-vivo and in-vivo studies suggested the possibility of analyzing lipid metabolism in normal tissue, benign and malignant lesions using proton MRS (8). Some studies also analyzed the feasibility of 31P MRS, with which it is possible to analyze nucleotide triphosphates (NTPs), phosphocreatine (PCr), and inorganic phosphate (Pi). Though, only few preliminary studies are available (9).
Clinical Applications
Proton MRS has been introduced as a method to improve lesion characterization with breast MRS. A meta-analysis published in 2013, which analyzed the performance of Choline peak in lesion characterization, found a pooled sensitivity of 73% and specificity of 88%, but also a significant publication bias (5). Only a few studies are available, which analyzed water to fat ratio and fatty acid, with very promising results (10). More recent studies suggested the use of choline measurements in breast MRI to improve the characterization of indeterminate enhancing lesion, suggesting that the implementation of MRS might improve the selection of lesions that need further investigation (11). Other interesting, but currently under investigated, topics in breast MRS include: the evaluation of tumor subtype and tumor aggressiveness; assessment of lymph nodes status; assessment of response to neoadjuvant chemotherapy (10, 12).

Acknowledgements

Thanks to Pascal A.T. Baltzer, M.D., for the help and the always valuable advices.