¹H-MR spectroscopy of suspicious breast mass lesions at 3T: a clinical experience

Objectives: To test 3T proton magnetic resonance spectroscopy (¹H-MRS) for breast mass lesions. Methods: Patients with BI-RADS 4–5 lesions at mammography/ultrasound were prospectively enrolled. After contrast-enhanced breast MRI, single-voxel MRS (point-resolved volume selection, PRESS); pencil-beam shimming; volume of interest 1 cm³; TR/TE = 3000/135 ms) was performed. Spectra were considered reliable if the full width at half maximum (FWHM) of the water peak was ≤45 Hz. A signal-to-noise ratio of the total choline (tCho) peak at 3.21 ppm ≥2 was used as cutoff for malignancy. All lesions underwent needle sampling. Final pathology was available for all malignant lesions; for benign lesions the reference standard was final pathology or at least 1-year negative follow-up. Results: Reliable spectra were obtained in 115/127 lesions (91%), with a mean FWHM of 32.4 Hz (range 8–45 Hz). A tCho peak SNR ≥2 was detected in 66 malignant lesions (62 invasive cancers; 4 ductal carcinoma in situ) and in 3 benign lesions. Excluding lesions located ≤1 cm from the skin (n = 3) or pectoral muscle (n = 11), sensitivity was 65/73 [89%, 95% confidence interval (CI): 80–95%], and specificity 25/28 (89%) (95% CI: 72–98%). Considering only invasive cancers, sensitivity reached 61/68 (90%, 95% CI: 81–96%). MRS additional time was 8 min. Conclusions: When lesions close to the skin or pectoral muscle are excluded, 3T ¹H-MRS of mass lesions ≥1 cm showed a high diagnostic performance, however, insufficient to avoid needle biopsy.