Andres Saucedo1, Manoj Kumar Sarma1, Sumit Kumar1, Kavya Umachandran1, Melissa Joines1, Stephanie Lee-Felker1, Maggie DiNome2, and Michael Albert Thomas1
1Radiological Sciences, University of California, Los Angeles, Los Angeles, CA, United States, 2Surgery, University of California, Los Angeles, Los Angeles, CA, United States
Synopsis
Multi-parametric
MR techniques have been used to diagnose and monitor the therapeutic outcome of
cancer in the breast and other tissues and organs. One-dimensional MRSI has
shown significantly elevated choline and higher water-to-fat ratio in malignant
tumors as compared to healthy controls. Two-dimensional MRS resolves peaks
along an additional spectral dimension which overcomes the overlap limitation
of 1D MRSI, thereby providing more discriminatory information for developing
non-invasive methods for cancer grade determination. This study presents the
first application of an accelerated, echo-planar based technique that acquires
correlated (2D) spectroscopy data for each voxel of 1.5ml resolution within a
3D volume (5D EP-COSI) in breast cancer. Our preliminary results in a pilot
cohort of malignant and breast cancer patients demonstrate changes in
unsaturated fatty acids and increased choline in malignant group compared to
benign and healthy women. These pilot results indicate the potential application
of the 5D EP-COSI technique which may be useful in improving the specificity of
breast cancer.
Introduction
Magnetic
resonance imaging and spectroscopy (MRI/MRS) have been applied to diagnose and
monitor the therapeutic outcome of cancer in the breast and other tissues and
organs. Dynamic contrast-enhanced (DCE)-MRI
has high sensitivity in detecting the presence of cancerous lesions, however
the addition of magnetic resonance spectroscopic imaging (MRSI) greatly improves
the specificity of diagnosis1-4. One-dimensional MRSI has shown
significantly elevated choline and higher water-to-fat ratio in malignant
tumors as compared to healthy controls. Due to peak overlap in 1D MRS methods,
however, the correlation between lipid and other metabolite levels and cancer
grade can be obscured. Two-dimensional MRS resolves peak information along an
additional spectral dimension which overcomes the overlap limitation of 1D MRSI,
thereby providing more discriminatory information for developing non-invasive
methods for cancer grade determination. This study presents the first
application of an accelerated, echo-planar based technique that acquires
correlated (2D) spectroscopy data for each voxel of 1.5ml resolution within a
3D volume (5D EP-COSI)5-6 in breast cancer.
Methods
Subjects
with malignant (n=20, mean age of 53.2 years) and benign (n=11, mean age of
36.7 years) breast cancer as well as 7 healthy volunteers (mean age of 38.7
years) were recruited as part of an IRB-approved study to record 5D EP-COSI
data on a 3T scanner with the following parameters: TR = 1.5 s, TE = 35 ms,
voxel size = 1.0 × 1.0 × 1.5 cm3, matrix size = 16 × 16 × 8,
spectral width SW2 = 1190 Hz, SW1 = 1250 Hz, 512 t2
points, 64 t1 increments. A non-water-suppressed EP-COSI data with t1=1
was also recorded. Both water and non-water-suppressed scans were acquired for
a total scan time of 28.8 minutes. The data was non-uniformly undersampled in
the ky-kz space and t1 dimensions with an
acceleration factor of 8, and reconstructed using Group Sparsity-based
compressed sensing6. EP-COSI data taken from 12 malignant, 6 benign
and 5 healthy cases were considered and compared qualitatively. Voxels
corresponding to the tumor region in both benign and malignant cancer patients
were determined based on the subtraction images from clinically-obtained
DCE-MRI.Results
The
5D EP-COSI technique records 2D MRS data within multiple slices, allowing for
broad coverage of the tumor volume (Figure 1(A)). Figure 1(B) shows the chemical
shift image of methylene fat (1.4 ppm) demonstrating localization accuracy. The
location of the tumor region from a malignant patient and the corresponding
COSY spectra is demonstrated in Figure 2, which shows the presence of elevated
choline and water, in contrast to the surrounding region. Malignant and benign
tumors both present high water content, though choline is elevated only in
malignant cases (Figure 3). Additional metabolites in the 3-4 ppm region, such
as myo-Inositol, glycine, and taurine are also measureable (Figure 4). Figure 5(A)
shows choline-to-fat ratio (Cho/Fat) as a function of cancer grade among the 12
malignant patients. Cho/Fat ratio was able to distinguish the cancer grades and
progressive increase was observed with the increasing grades. Figure 5(B) shows
the distribution of Cho/Fat as a function of Ki-67 metric.Discussion
Our
preliminary results in a pilot cohort of malignant and breast cancer patients
demonstrate changes in unsaturated fatty acids and increased choline in
malignant group compared to benign and healthy women. In contrast, earlier reports
using 1D MRS quantified the lipid peak at 1.4ppm only 1-3. Several ex vivo high
resolution magic angle spinning (HR-MAS) studies have reported detection of
choline, glycine, taurine, myo-inositol and more7-8. For the first time, our
preliminary multi-dimensional MR Spectroscopic Imaging data has shown the
feasibility of detecting many metabolites more than choline similar to reported
by ex vivo MRS studies. One of the drawbacks of this study was due to chemical
shift displacement errors (CSDE) by using conventional slice-selective 90
degree RF pulses and CSDE may have eliminated detection of uridine diphosphate
hexose and other metabolites in the low field region; this can be minimized
using adiabatic RF pulses.Conclusion
Our
pilot results indicate the potential application of the 5D EP-COSI technique which
may be useful in improving the specificity of breast cancer. Further investigation
using a larger cohort of breast cancer patients will be necessary in the
future.Acknowledgements
This research was
supported by a CDMRP Breakthrough Step I award #W81XWH-16-1-0524.
References
1) Sharma U, Agarwal
K, Sah RG,,, et al. Can Multi-Parametric MR Based Approach Improve the
Predictive Value of Pathological and Clinical Therapeutic Response in Breast Cancer Patients?. Front Oncol. 2018 Aug 15;8:319..
eCollection 2018. Epub Ahead.
2) Bolan PJ, Kim E,
Herman BA,, et al. MR spectroscopy of breast
cancer for assessing early treatment response: Results from the ACRIN
6657 MRS trial..
J Magn Reson Imaging. 2017
Jul;46(1):290-302. Epub 2016 Dec 16.
3) Nelson MT, Everson LI, Garwood M, Emory T, Bolan PJ.
MR Spectroscopy in the diagnosis and treatment
of breast cancer.. Semin Breast Dis. 2008 Jun
1;11(2):100-105.
4)Lipnick et
al. Combined DCE-MRI and single-voxel 2D MRS for differentiation between benign
and malignant breast lesions. NMR in Biomedicine, 2010; 23
(8): 922-930.
5)
Wilson NE, Burns BL, Iqbal Z, Thomas MA. Correlated spectroscopic imaging of
calf muscle in three spatial dimensions using group sparse reconstruction of
undersampled single and
multichannel data.. Magn
Reson Med2015;74:1199–208.
6) Burns et al.
Group sparse reconstruction of multi-dimensional spectroscopic imaging in human
brain in vivo. Algorithms, 2014; 7(3): 276-294.
7) Haukaas TH,
Euceda LR, Giskeodegard GF, Bathen TF. Metabolic portraits of breast cancer by
HR MAS MR Spectroscopy of intact tissue samples. Metabolites 2017;7, pii: E18.
8) Choi JS,
Baek HM, Kim S, et al. HR-MAS MR spectroscopy of breast cancer tissue obtained with core needle
biopsy: correlation with prognostic factors. PLoS One
2012;7:e51712.