Improving Outcomes With MRI/MRS
Research evaluates the roles of MRI and MRS in brain tumor therapy.
Research conducted collaboratively by ACRIN® and RTOG® to assess the potential role of perfusion MRI and magnetic resonance spectroscopy (MRS) as early indicators of response to antivascular endothelial growth factor (VEGF) therapy for recurrent glioblastoma (GBM) holds promise for improving historically poor clinical outcomes in patients with this malignant brain tumor.
Participants in the ACRIN 6677/RTOG 0625 protocol received the anti-VEGF drug bevacizumab in combination with one of two cytotoxic agents — temozolomide or irinotecan. Thirty-four participants were enrolled in the trial's advanced-imaging component at 10 sites, and a subset of 20 participants underwent additional MRS scanning at four sites — first at baseline before treatment and then at two weeks, eight weeks, and every two months after treatment, for a total of 96 weeks.
Using MRS in a Research Environment
Eva-Maria Ratai, Ph.D., assistant professor of radiology at Harvard Medical School in Cambridge, Mass., and assistant in neuroscience at Massachusetts General Hospital in Boston, performed the central review of the MRS data and presented results of the study's associated aims at the 2011 RSNA Annual Meeting, held Nov. 27-Dec. 2. "Unlike standard structural MRI, spectroscopy provides biochemical information that helps us better understand a drug's mechanism of action," states Ratai. "Within the first day of drug administration, changes can be seen in three peaks on the MR spectrum; these are derived from N-acetylaspartate (NAA), choline (Cho), and creatinine molecules. Typically, the higher the Cho — a marker of increased cellular turnover — and the lower the NAA peaks — a marker of neuronal density and viability — the more malignant the tumor. Our goal was to determine if early changes, e.g., at eight weeks in the Cho and NAA MRS markers, correlated with patient outcome."
For the study's spectroscopic analysis, the voxels (volumetric pixel) identified on the postcontrast T1-weighted images were classified as tumor-enhancing, nonenhancing tumor periphery, and normal tissue on the contralateral side. Data analysis involved quantification of the various biomarker ratios.
Results showed that while at two weeks, all measured markers showed improvement, at eight weeks, tumors of participants that showed a positive response to treatment could be distinguished from those not responding to the treatment. Of the 13 participants with analyzable MRS datasets, nine were progression-free at six months. Those nine participants had higher levels of NAA/Cho at eight weeks, suggesting an association of an increase in NAA/Cho with treatment response (shown in Figure 1).
"Despite the small sample size, at least three imaging and data time points (for a total of 75) were available per participant. With an area under the curve of 85 percent and a fairly broad confidence interval (0.53-1), the correlation between spectroscopic data and outcome is significant," summarizes Ratai. Data analysis performed on tissue around the periphery of the tumors also showed interesting results. "After eight weeks of anti-angiogenic therapy increases in periphery N-acetylaspartate (the marker for neuronal integrity) were associated with survival," says Ratai.
MRS Poised to Expand Clinical Care
"The study's findings suggest that changes in the NAA/Cho may be useful as an imaging biomarker in assessing response to antiangiogenic treatment for GBM (see Figure 2). Confirmation about whether a patient's tumor is responding to treatment would allow physicians early on in the course of treatment to choose different treatment options for those not responding," according to Ratai. In addition, by helping to identify molecular profiles associated with different patterns of brain tumors, these biomarker changes could also assist in tumor classification and staging and guiding biopsy.
Conducting the ACRIN 6677 and 6684 trials has significantly increased the number of sites competent to perform spectroscopy, many of which experienced a learning curve in performing MRS and submitting data for review. "Personnel at each site were required to submit imaging data to confirm their ability to complete the data acquisition correctly," comments Ratai, "and some sites had to repeat the process to ensure we received good data for the analysis and made good use of the study's participant's time."
Ratai points to the improved data she is receiving from sites conducting the ACRIN 6684 trial assessing tumor hypoxia in GBM using 18F-fluoromisonidazole (FMISO) with PET and advanced MRI sequences, including MRS. "So far I have MRS data for 15 cases from five different sites out of 20 patients that enrolled in the study. This is significantly improved performance as compared to the MRS data received for the ACRIN 6677 trial."
Some of the inherent challenges in performing MRS in a multicenter clinical trial include the importance of reproducibility as well as maintaining a homogenous field when acquiring the MRS data. "The method's sensitivity requires ongoing scanner shimming (calibration) to maintain magnetic field homogeneity for the marker peaks to be more easily quantified," says Ratai. In addition, lipid contamination can result when tumors close to the skull are examined.
Overall, Ratai is optimistic about the results of the trials. "We are hopeful about the prospect of combining MRS data with other quantitative imaging measurements to develop a very good predictor of whether a specific treatment is actually benefitting a patient."
By Nancy Fredericks, M.B.A., and Julie Catagnus, M.S.W., ELS
Julie Catagnus, M.S.W., ELS, is a freelance writer.