In Vivo MRS for Cancer Diagnosis and Treatment Monitoring

Institution: University of California, Irvine
Investigator(s): Hyeon-Man Baek, Ph.D. -
Award Cycle: 2006 (Cycle 12) Grant #: 12FB-0031 Award: $90,000
Award Type: Postdoctoral Fellowship
Research Priorities
Detection, Prognosis and Treatment>Imaging, Biomarkers, and Molecular Pathology: improving detection and diagnosis



Initial Award Abstract (2006)
Compared to conventional breast examinations (clinical, mammography, and ultrasound), MRI has been proven to be the most sensitive imaging modality for detecting invasive cancer in dense breast or breast with implants. It is fast becoming the optimal choice for screening young women at a high risk of developing breast cancer. However, despite its high sensitivity, MRI also detects many incidental lesions. The false positive finding may lead to a great anxiety to patients, and in many cases unnecessary biopsies or over-treatment. Other adjunct imaging modalities that can achieve a higher diagnostic accuracy, especially in women with dense breasts, are greatly needed.

Magnetic resonance spectroscopy (MRS) is a non-invasive technique that provides information about tumor metabolism. We will investigate in vivo quantification of choline [choline and its metabolites are needed for structural integrity and signaling roles for cell membranes, neurotransmission (acetylcholine synthesis), and as a major source for methyl-groups] levels in benign, in-situ and invasive cancers. Two hypotheses will be tested: (1) whether the additional information measured by MRS may improve the specificity of MRI (morphological and kinetic enhancement characteristics) in differentiating benign from malignant tumors, and (2) whether the addition of metabolic changes to MRI morphological and kinetic parameters will enhance the prediction of complete response in patients receiving neoadjuvant chemotherapy. To evaluate the potential of our MRS approach, we will include patients: (1) with inoperable locally advanced breast cancer, (2) with clinically documented lymph node involvement, and (3) those with down-staged tumors to facilitate breast conservation surgery.

The innovative component is the quantification of in vivo MRS and morphologic and enhancement kinetic parameters in MRI for diagnosing breast cancer and monitoring neoadjuvant chemotherapy efficacy Ultimately this quantitative method can be developed to provide an MRI Computer-Assisted Detection (CAD) system. The success of this study may have beneficial effect by potentially reducing negative biopsies and provide a better treatment for each individual patient and reduce side effects caused by ineffective therapies.


Final Report (2008)
Magnetic resonance spectroscopy (MRS) is a non-invasive technique that has great potential to provide information about tumor metabolism, which may be used in tumor diagnosis and evaluating the therapeutic response of the tumor. The purpose of the present research was to investigate the role of MRS to improve the specificity of dynamic contrast enhanced MRI (DCE-MRI) in breast cancer diagnosis and also to evaluate the early response to neoadjuvant chemotherapy treatment. In this study, we hypothesized that the additional information measured by MRS (Choline levels = Cho) may improve the specificity of DCE-MRI (morphologicaland kinetic enhancement characteristics) in differentiating benign from malignant tumors, and that the addition of metabolic changes (Cho) to morpholological and enhancement kinetic parameters will improve the prediction of complete response in patients receiving neoadjuvant chemotherapy. Choline is a micronutrient found in lipids that make up cell membranes. Increases in Cho level is thought to associate preferentially with tumors as a result of vascularity changes and altered metabolism in tumor cells.

Of the 27 malignant lesions, the analysis of DCE enhancement time course showed 20 (74%) washout pattern, and 4 (15%) plateau pattern. The plateau pattern is considered equivocal, but is suspicious for recommending biopsy. In MRS, three of these four patients showed Cho levels > 3.2 that were subsequently diagnosed as cancer, the other patient had Cho levels of 3.2, and would be mis-diagnosed as benign on MRS. Three malignant lesions that showed persistent enhancing kinetics were mis-diagnosed as benign on DCE. Two of them had Cho levels <3.2, thus would also be mis-diagnosed as benign based on MRS. Therefore, MRS would not improve the sensitivity of DCE-MRI. Of the 9 benign lesions, there were two false positive findings based on DCE kinetics, which resulted in 78% specificity. One patient had the washout pattern and the other had the plateau pattern. Their Cho levels were 2.9 and 2.7, respectively, thus they would be correctly diagnosed as benign based on the MRS criterion of levels < 3.2. Therefore, consistent with previously published studies [i.e., Jacobs et al, JIARI (2005) and Meisamy et al, Radiology (2005)], the value of MRS may lie on improving specificity of DCE-MRI.

35 patients with histopathology-confirmed breast cancer who underwent neoadjuvant chemotherapy (NAC) were included in the MRI and MRS study. They had at least 2 follow-up (F/U) studies, F/U-1 after 1-2 cycles AC, and F/U-2 after 4 cycles AC or 2 cycles AC followed by first cycle of taxane regimen. The ROC analysis was performed to differentiate between pathological complete response (pCR) and non-pCR patients based on the changes measured at F/U-I and F/U-2 time points. The percentage change could only be calculated from patients who had measured Cho pre-NAC. At F/U-I, the area under the ROC curve for tumor size, Cho level and combined (i.e., tumor size + Cho) were 0.66, 0.67, and 0.72, not significantly different between pCR and non-pCR groups. At F/U-2, the areas under the ROC curves was 0.90 for tumor size, 0.73 for Cho level, and 0.92 for both combined, all showing significant differences (P < 0.005). Our findings suggests that the tumor size change at F/U-2 was the most accurate predictor of pCR. Overall, our results indicate that MRS may be able to detect tumor responses to chemotherapy with more sensitivity that physical tumor shrinkage as a marker of response to therapy.

This study demonstrates that the MRS technique can be incorporated into the clinical breast MR imaging protocol with an acceptable scan time. Since interpretation of DCE-MRI is set to reach a high sensitivity with compromised specificity, the metabolic information measured by MRS may be used for improving the specificity in diagnosis of breast tumors. Future studies will need to examine many more patients with benign tumors. In addition, this study demonstrates that the metabolic Cho changes were higher than the size changes in the pCR group but not in non-pCR group. Our results show that the combination of MRI and MRS produce more accurate results than either modality alone. Therefore, we believe the metabolic information provided by MRS may further enhance the impact of MRI on breast cancer therapy.


Symposium Abstract (2007)
Background: Magnetic Resonance Spectroscopy (MRS) is a non-invasive technique that has great potential to improve breast cancer diagnosis and monitor therapeutic response to treatment. It has been shown that malignant tumor contains elevated levels of choline-containing compounds (Cho) compared to benign tumor and normal breast tissue, and the change of Cho level can be used as a marker to assess the response of breast cancer patients with neoadjuvant chemotherapy. An early decrease in the concentration of Cho may serve as an early response indicator to predict final outcome after completing all chemotherapy.

Materials and Methods: Nineteen cancer patients receiving neoadjuvant chemotherapy were studied. These patients had lesions greater than 2.4 cm. MR imaging and MRS were performed at the baseline before therapy, then at 2 follow-up times, F/U-1 after 1-2 cycle of AC (Adrimycin and Cyclophosphamide), and F/U-2 after 4 cycles AC or 2 cycles AC followed by first taxane regimen. Patients were categorized into responders or non-responders based on the 1-demensional RECIST criteria. Responders were defined as subjects with >30% 1-dimentional tumor size reduction at FU-2 compared to the baseline. For each patient the changes in the level of Cho and the tumor size at the F/U-1 study relative to their respective baseline were measured, and compared between the R and NR groups.

Results: 14 of 16 patients who were responders had positive Cho before treatment. The mean percentage change in Cho level after 1-2 cycle AC was -50%, while the median percentage change in lesion size was -18%. After completing F/U-2, all patients did not show any detectable Cho. The median lesion size change in F/U-2 study was -76%. Three patients were classified as non-responders. All 3 had a positive Cho at the baseline. The mean percentage change in Cho level after 1-2 cycle AC was -14%, while the median percentage change in lesion size was -10%. Of them, 2 patients showed increased Cho in F/U-2, and the remaining 1 patient decreased but still detectable. The median percentage changes in Cho level and lesion size in F/U-2 study relative the baseline were 3.3% and -27%, respectively. In the R group, significant reductions in Cho level and tumor size were observed at the early F/U-1 with respect to the baseline (p < 0.002, p < 0.008), whereas no significant change was observed for either measure in the NR group.

Discussion: In the R group, 14 of 16 patients had baseline positive Cho, and all 14 showed reduced Cho level in F/U-1, and no detectable Cho in F/U-2. In the NR group, all 3 patients had baseline positive Cho. Although they also showed Cho reduction in F/U-1, but at a smaller percentage compared to those in the R group, and 2 had transient reduction which showed increased Cho in F/U-2. The Cho level showed different response patterns between R and NR group, suggesting that it may serve as an early response predictor. Early Cho reduction may be a sensitive diagnostic tool for predicting final outcome to chemotherapy treatment in breast cancer.

Detection of choline signal in human breast lesions with chemical-shift imaging.
Periodical:Journal of Magnetic Resonance Imaging
Index Medicus: J Magn Reson Imaging
Authors: Baek HM, Chen JH, Yu HJ, Mehta R, Nalcioglu O, Su MY
Yr: 2008 Vol: 27 Nbr: 5 Abs: Pg:1114-21

Proton MR spectroscopy for monitoring early treatment response of breast cancer to neo-adjuvant chemotherapy.
Periodical:Annals of Oncology
Index Medicus: Ann Oncol
Authors: Baek HM, Chen JH, Nalcioglu O, Su MY
Yr: 2008 Vol: 19 Nbr: 5 Abs: Pg:1022-4

Choline as a biomarker for cell proliferation: do the results from proton MR spectroscopy show difference between HER2/neu positive and negative breast cancers?
Periodical:International Journal of Cancer
Index Medicus: Int J Cancer
Authors: Baek HM, Chen JH, Nalcioglu O, Su MY
Yr: 2008 Vol: 123 Nbr: 5 Abs: Pg:1219-21