Authored date:2006-06-07

Introduction

Over the years numerous studies have evaluated the use of MR in the characterization of breast lesions and in monitoring tumor response to treatment. The results have shown that although MR imaging and dynamic contrast-enhanced MRI have high sensitivity and can be helpful in differentiating benign from malignant tumors, they are not perfect predictors; and additional diagnostic criteria are needed to help clarify equivocal MRI results [1–2]. Combining in vivo MR spectroscopy (MRS) with contrast-enhanced MRI increases the specificity of breast MRI due to the additional biochemical information obtained with MRS. MR spectroscopy has been performed in the breast with 1H (proton) as well as with other nuclei including phosphorus (³¹P), carbon (¹³C) and fluorine (¹⁹F), which is used for studies of drug metabolism namely 5-fluorouracil (5-FU). ¹H MRS has practical advantages over MRS with other nuclei: It has the highest sensitivity and can be performed at the end of an imaging exam without patient or coil repositioning. ¹H MRS does not require additional hardware and can be readily incorporated into routine clinical breast MRI examinations.

¹H MRS of the breast

A typical ¹H spectrum of breast tissue consists of large lipid and water signals (F. 1). Historically the Choline peak seen at 3.2 ppm was originally attributed to malignant tumors (F. 2, 3, 5 and 6) in studies performed at 1.5T. However, small choline signals have been detected in benign lesions and in normal breast tissue at higher field strength. Choline has also been seen in lactating breast spectra along with the characteristic lactose signal [3] (Fig. 4).

Fig. 1  Single voxel spectra from healthy breast tissue acquired on a 1.5T MAGNETOM Avanto with TR = 1500 ms, TE= 135 ms, and a voxel size of 8 cm³. Spectrum A shows the non suppressed water and lipid signals acquired with 8 averages in 12 s. Spectrum B was acquired with 128 averages in 3:12 min, with suppression of the lipid and water signals. It shows a flat range between the residual water and lipid peaks. No choline peak is seen at 3.2 ppm. (Courtesy of Linda Moy, M.D. and Vivian Lee, M.D., Ph.D., New York University School of Medicine, New York, USA.)

Fig. 2  SVS Spectra from a biopsy proven adenocarcinoma acquired on a MAGNETOM Trio, A Tim System, with TR = 2000 ms, TE = 135 ms, 64 averages, 2:56 min and a voxel size of 6.8 cm³. They were measured with lipid suppression and weak water suppression, which leaves a residual water peak for reference. Spectrum A displays part of the residual water signal along with the choline peak. Spectrum B is an enlarged view of the choline peak seen in A. (Courtesy of Linda Moy, M.D. and Vivian Lee, M.D., Ph.D., New York University School of Medicine New York, USA.)

Fig. 3  SVS spectra showing the choline peak from a biopsy proven high grade malignant tumor acquired in 3:12 min on the MAGNETOM Symphony with lipid suppression and weak water suppression, which leaves a residual water peak for reference. (TR =1500 ms, TE =135 ms, 128 averages, voxel size = 8 cm³). Spectrum A shows a suppressed lipid signal, the residual water peak and the choline peak (arrow). Spectrum B displays the rescaled choline peak. (Courtesy of Linda Moy, M.D. and Vivian Lee, M.D., Ph.D., New York University School of Medicine, New York, USA.)

Fig. 4  Single voxel spectra from a healthy lactating breast. The spectrum in A was acquired on a 1.5T MAGNETOM Avanto from a 17 cm³ voxel with TE = 135 ms and TR = 1500 ms, in 3:12 min. The spectrum in B was acquired on a 3T MAGNETOM Trio, A Tim System from a 10 cm³ voxel with TR = 1500 ms and TE = 135 ms in 1:16 min. The multiple lactose peaks are better resolved in spectrum B. Both spectra were acquired with lipid and weak water suppression. Peaks between 0 and 2.8 ppm are residual lipid signals. (Courtesy of Linda Moy, M.D. and Vivian Lee, M.D., Ph.D., New York University School of Medicine, New York, USA.)

Detection of a small choline peak in the presence of large lipid signals presents a technical challenge because lipid sidebands may interfere with the choline peak at 3.2 ppm. Breast ¹H MRS on the MAGNETOM systems uses a special single voxel spectroscopy (SVS) spin echo sequence* with spectral lipid suppression, which reduces the effect of lipid on the choline signal. The weak water suppression option leaves a residual water peak that serves as a reference. The sequence also includes averaging of phase and frequency corrected signals, which minimizes signal loss from breathing motion.

Breast ¹H MRS: Monitoring response to chemotherapy

Figures 5 and 6 illustrate the use of MR spectroscopy in monitoring the effect of chemotherapy on a diagnosed invasive breast carcinoma†. Spectra were acquired following each of the 6 cycle treatment. They demonstrate a gradual reduction in the choline peak, which is hardly visible after the sixth cycle of chemotherapy. Using the Phoenix functionality of the syngo software the same measurement protocol was recalled for each follow-up examination. This insured that the same parameters were used for all measurements except for the voxel size, which was adjusted to the reduced size of the tumor.

† Courtesy of Professor T. J. Vogl, University of Frankfurt/Main, Germany.

Fig. 5  Spectra from follow up examinations of an invasive breast carcinoma treated with chemotherapy. The spectra were measured following each of the 6 cycle treatment. They were acquired on a 1.5 T MAGNETOM Sonata with TR = 1500 ms, TE = 135 ms, in 3 to 5 minutes depending on the voxel size, which was adjusted to the size of the tumor, and varied from 
(18 x 11 x 14) mm³ to (10 x 10 x 10) mm³. All spectra display a residual water peak at 4.7 ppm (green arrow) and a choline peak at 3.2 ppm (yellow arrow). Spectrum A (on page 96) shows the choline peak prior to chemotherapy. Spectra B, C, D, E, F and G were measured following cycles 1, 2, 3, 4, 5 and 6 respectively. A reduction in choline is clearly seen following the second, third and fourth, fifth and sixth cycles. The reduction in choline was accompanied by a reduction of the                                       tumor size. (Courtesy of Professor T. J. Vogl, University of Frankfurt / Main, Germany.)

Fig. 6  Selected spectra from the follow-up examinations described in Fig. 5, showing a reduced choline peak after the second, fourth and sixth cycle.

Conclusion

The results presented here show that the technical challenges that hindered applications of ¹H MRS to breast tumor characterization are now being overcome with improved hardware and software. Developments that improve detection of small signals, such as choline in ¹H MRS breast applications, include the availability of higher field strength scanners and of special software that considerably reduces the effect of large lipid signals on small neighboring peaks. With these new developments and with the addition of fast and reliable choline quantitation techniques, ¹H MRS will improve even further the specificity of MR in breast tumor classifications.

* This information about this product is preliminary. The product is under development and not commercially available in the US and its future availability cannot be ensured.

References

[ 1 ] David K.W. Yeung, PhD, et al. Breast cancer: In vivo proton MR spectroscopy in the characterization of histopathologic subtypes and preliminary observations in axillary nodes metastases. Radiology, 225, pp 190–197, 2002.
[ 2 ] Rachel Katz-Brull, Philip T. Lavin, Robert E. Lenkinski. Clinical utility of proton magnetic resonance spectroscopy in characterizing breast lesions. J. Natl Cancer Inst, 94(16), pp 1197–1203, 2002.
[ 3 ] Kvistad KA, et al. Characterization of neoplastic and normal human breast tissues with In vivo 1H MR spectroscopy. 
J. Magn. Reson. Imaging 10; pp 159–64, 1999.