We have continued our research efforts at improving methods of detecting breast disease, especially cancer, using proton Magnetic Resonance Spectroscopy. Our hypothesis remains that breast carcinomas have higher levels of total choline concentration than those of benign breast lesions. Our studies were performed on a 1.5T clinical MR scanner using a custom-built four channel multicoil assembly which compresses the breast sagittally. We made use of contrast-enhanced MRI and single voxel proton MRS using STEAM at echo times of 31 and 270 msec. Over the past year, we have studied breast lesions in twenty-six women, and made pathologic correlations using needle biopsy or post-surgical specimens. This population had 17 women with malignant breast lesions and 9 women with benign processes. The MRS examination was performed by sampling nominal voxel sizes less than 4cc which were prescribed from the MR images. Spectra showed resonances from water and mobile fatty acids; in a majority of carcinomas, the N-trimethyl resonance of compounds containing choline at 3.2 ppm was present as well. Using the presence of Cho as a test for malignancy, the sensitivity and specificity of our entire patient population (N=55) were calculated at 79% and 95%, respectively. The positive and negative predictive values of this test were 96% and 75%. Spatially localized proton MRS can provide sufficient sensitivity and spectral resolution at 1.5 T to observe choline in human breast lesions. Testing for the presence of choline serves as a test for malignancy that is useful in characterizing breast lesion either on its own or in combination with MRI criteria.
Showing the most recent 10 out of 414 publications
