Lia Bartella MBCHB, Presenter: Nothing to Disclose

Elizabeth Anne Morris MD, Abstract Co-Author: Nothing to Disclose

Dale David Dershaw MD, Abstract Co-Author: Nothing to Disclose

Laura Liberman MD, Abstract Co-Author: Nothing to Disclose

Sunitha Thakkur PhD, Abstract Co-Author: Nothing to Disclose

Wei Huang PhD, Abstract Co-Author: Nothing to Disclose

The purpose of this study was to determine the sensitivity and specificity of in vivo proton MR spectroscopy in breast cancer diagnosis and correlate with breast MRI.

In accordance with our IRB approved protocol, to date, breast MRI spectroscopy was performed on 12 prospectively accrued women with suspicious breast MRI lesions (>1.2cm). Using MR images , single-voxel MRS data were collected from a rectangular box encompassing the lesion, with 135 ms TE and 2 s TR and a total scanning time of about 10 min. The voxel size ranged from 1.8 to 9.0 mL3. Findings were defined positive if the signal-to-noise ratio of the choline resonance peak at 3.23 ppm was greater than or equal to two, negative otherwise. A 1.5T magnet and a phased array breast coil were used. The patients had subsequent surgery or needle biopsy and pathology was correlated with the MRS and MRI data. Statistical analysis was performed with statistical software (Epi-Info, CDC).

The median MRI lesion size was 1.6 cm (range, 1.3- 6). Histologic analysis yielded cancer in 8 (66%) and benign findings in 4 (33%). Among 8 breast cancers, histology was infiltrating ductal carcinoma in 4 (50%), infiltrating lobular carcinoma in 3 (37.5%), and ductal carcinoma in situ in 1 (12.5%). A choline peak was present in 9/12 lesions, including 8/8 cancers and 1/4 benign lesions. MRS had a sensitivity of 100% (95% CI, 63-100%) and a specificity of 75% (95% CI, 19-95%). Correlation with the MRI BIRADS classification revealed all BIRADS 5 lesions and a BIRADS 6 lesion were malignant. 2/6 (33%) of the BIRADS 4 lesions were malignant and 4/6 (66%) were benign. If biopsy had only been performed on lesions with a choline peak at spectroscopy, the positive predictive value would have increased from 66% to 89% (p=0.33); biopsy would have been spared in 3/12 (25%) of lesions, missing none of the cancers .

Proton MRS was easily incorporated into breast MR studies for lesions as small as 1.3 cm and was useful in characterizing benign lesions, invasive and in situ carcinomas. In our limited experience, it has a high negative predictive value and may be useful in narrowing the number of MR detected lesions that require biopsy.

Bartella, L, Morris, E, Dershaw, D, Liberman, L, Thakkur, S, Huang, W, Proton MR Spectroscopy (MRS) Using Choline Signal as Malignancy Marker Improves Specificity Compared to Conventional MRI in Diagnosis of Breast Cancer: A Preliminary Study.  Radiological Society of North America 2004 Scientific Assembly and Annual Meeting, November 28 - December 3, 2004 ,Chicago IL. http://archive.rsna.org/2004/4406485.html