Breast cancer and non-malignant breast data: Quantification by FPT - Signal Processing in MRS with Biomedical Applications - page 306

in vivo MRS study [383] employed a very long TE of 288 ms, as well as an
inversion recovery sequence to diminish the fat signal, and water suppression
by chemical selective saturation together with improved shimming of the static
magnetic field. The reported results from this work [383] reveal that the
presence or absence of choline accurately detected all nineteen malignancies
and all sixteen noncancerous breast lesions. The choline SNR varied from
2.4 to 12.7, with a mean value of 5.4 for the malignant lesions.
Case reports have been published describing the use of in vivo MRS to
help rule out malignancies in unusual and suspicious appearing breast lesions.
These have included a desmoid tumor, found to have absent choline [384]
and a tubercular breast abscess also found to have absent choline, but with a
prominent peak at 0.9 - 1.3 ppm considered to represent lipidlactate [385].
Tse et al. [381] recently reviewed the various benign pathologies that have
been evaluated using in vivo MRS. Of a total of 46 fibroadenomas, three were
misdiagnosed as malignant. Of 20 fibrocystic lesions, two were false positives
on in vivo MRS.
There have been some studies assessing the ratio between unsuppressed
water to lipid in breast cancer and normal mammary tissue with in vivo MRS.
The ratio of water at 4.7 ppm to the lipid peak at 1.3 ppm was significantly
higher in breast voxels containing cancer compared to sites free of malignancy
among seventeen patients. This water to lipid ratio was also higher at sites of
malignancy compared to fourteen healthy referents [386]. However, there was
a very large standard deviation among the malignant ratios (6.0±6.9), which
indicated overlap with the normal values (0.36±0.25). When overlapping
metabolites cannot be resolved in 1D, then 2D MRS has been applied in
attempts to separate the resonances using crosscorrelation plots. Using in
vivo 2D MRS a significantly higher diagonal peak volume of unsuppressed
water to fat (1.4 ppm) ratio was reported among twentyone patients with
invasive breast cancer compared to fourteen healthy women with fatty breasts,
however no ranges were given [387]. The diagonal peak volume ratios of water
to methyl fat at 0.9 ppm and to olefinic fat at 5.4 ppm were also significantly
higher in the cancerous tissue using 2D MRS, as were the water to cross peak
volumes of unsaturated fatty acid right (2.1 ppm, 5.4 ppm), unsaturated fatty
acid left (2.9 ppm, 5.4 ppm) and triglyceride fat (4.3 ppm, 5.3 ppm) [387].
The elevated water to fat ratios observed in breast cancer may be related to
edema, due to lymphatic blockage and involvement of subdermal lymphatics
by the malignancy.
In the first reported investigation [388] of breast cancer via MRSI 1 , using
the choline SNR, a sensitivity of 87% and specificity of 85% were attained
among eight patients with breast cancer and 7 with benign breast pathology.
There were three other patients for whom MRSI was a technical failure. The
1 This MRSI investigation was carried out at 1.5T, shimming was performed to optimize
field homogeneity, 3 sequential Chemical Shift Selective (CHESS) pulses were used for water
suppression, lipid signals were attenuated with inversion pulse and TE=272 ms was selected.