0282

Magnetization Transfer and Chemical Exchange Saturation Transfer in Glioblastoma at 1.5T: Comparison of Early and Late Tumor Progression
Rachel W Chan1, Sten Myrehaug2, Greg J Stanisz1,3,4, James Stewart2, Mark Ruschin2, Arjun Sahgal2,3, and Angus Z Lau1,3
1Physical Sciences, Sunnybrook Research Institute, Toronto, ON, Canada, 2Department of Radiation Oncology, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, ON, Canada, 3Medical Biophysics, University of Toronto, Toronto, ON, Canada, 4Department of Neurosurgery and Pediatric Neurosurgery, Medical University, Lublin, Poland

Synopsis

This study aims to quantify the MT and CEST parameters in glioblastoma at 1.5T over four time points of chemoradiation. Previous approaches to quantify the MT/CEST signal in terms of response to treatment were at higher field strengths. Additionally, we analyzed different subregions within the clinical target volume and gross tumor volume, generated by thresholding based on the T1-weighted, FLAIR and DWI scans, for comparing the MT/CEST parameters between the early and late tumor progression. Results indicated that MT/CEST at 1.5T can be used for monitoring therapy and that the results depend on the specific tumor region analyzed.

Introduction

Glioblastoma multiforme (GBM) is the most aggressive and the most common form of brain tumor. The patient outcomes are poor despite standard treatment involving surgical resection, radiation therapy and chemotherapy1. Saturation transfer MRI, including quantitative magnetization transfer (MT)2-3 and amide proton transfer (APT) chemical exchange saturation transfer (CEST)4, have shown promise for monitoring response to treatment of brain tumors5-12. These CEST studies have been conducted at 3T or above. Recently, the feasibility of APT CEST at 1.5T was shown in a healthy human brain using a pulsed saturation method13. Using this approach13, we quantified the MT/CEST signal in GBM patients at 1.5T for distinguishing between early and late tumor progression.

In addition, since GBM tumors are spatially heterogeneous, we quantified the parameters over different tumor subregions defined based on thresholding the standard T1-weighted, FLAIR, and DW images. Subregions with signal characteristics related to those of a previous study that were found to be correlated with overall survival14 were examined. The median parameter values within each region were compared between early and late progression groups.

Methods

GBM study: Approval from the institutional research ethics board and informed consent from patients were obtained. Data from 34 patients (treated with intensity modulated radiation therapy with dose of 2 Gy per session with concurrent temozolomide) were analyzed. MRI was performed at four time points – at day 0 (“D0”) before radiation treatment, at fraction 10 and 20 of treatment (corresponding to days “D14” and “D28”, respectively) and 1 month following the final treatment fraction (“D70”).

MR protocol: Data were acquired on a 1.5T Philips Ingenia system. Standard clinical sequences included pre-contrast T1-weighted (“T1w”), post-contrast T1w (“T1w+C”), FLAIR and DWI scans. For saturation, MT and CEST sequences used short, block pulses designed to overcome the RF amplifier limitations on this scanner13. T1/T2/B0/B1 maps were acquired. MR parameters are shown in Figure 1.

Image pre-processing: The T1w+C and FLAIR volumes were registered to the T1w volume using the “flirt” function from FSL15, with 2D slices extracted to match the scanned MT/CEST slice. For CEST, motion correction across the saturation frequencies was performed using the FSL “mcflirt” function. The clinical target volumes (CTV) and gross tumor volumes (GTV), delineated based on T1w+C, were obtained for each time point.

Tumor subregions: In addition to the GTV and CTV, four subregions were analyzed. The subregions were generated by classifying pixels into high or low signal regions based on thresholding. For regions R1-R3, the thresholds were defined as the mean D0 signal within the CTV averaged over the entire patient cohort, from each of the T1w, FLAIR and ADC images, respectively. For subregion R4, the enhancing area within the GTV was determined based on increased signal on T1w+C compared to T1w.

Parameter fitting and analysis: Pulsed saturation data were fitted to the Bloch-McConnell model as in previous work13. The MT semisolid fraction, CEST parameters (including asymmetry, MTRAmide and MTRNOE, between 2-4ppm), ADC, T1 and T2 maps were quantified. Tumor progression was assessed at 6.9 months (=209 days)1, with early or late progression categories for tumors that progressed before or after 209 days, respectively. Differences between early and late progression groups were computed using the Wilcoxon Rank-Sum test for each time point, region and parameter.

Results and Discussion

Clinical images and parameter maps for a selected GBM tumor are shown across time points in Figure 2. The central tumor region had lower signal for the semisolid fraction (M0B) and CEST MTRAmide maps, and higher signal on the CEST asymmetry maps, compared to surrounding regions. In Figure 3, the CTV and GTV are shown of the thresholded subregions at D0.

In Figure 4, results comparing early and late progression over the CTV region are shown. At D0, the median MT semisolid fraction, CEST asymmetry and ADC were significantly different between the early/late groups (p<0.05). The values for early and late progression, respectively, were 6.0% (CI=[5.4, 7.4]%) and 7.4% (CI=[6.9,8.0]%) for the MT semisolid fraction, 0.58% (CI=[0.34, 0.85]%) and 0.33% (CI=[0.29,0.50]%) for CEST asymmetry, and 1.0×10-3mm2/s (CI=[0.91,1.1]×10-3mm2/s) and 0.9×10-3mm2/s (CI=[0.87,0.93]×10-3mm2/s) for ADC.

When thresholding was applied, there were increased differences in the MT fraction for the subregion R1 with low T1w signal compared to when the entire CTV was analyzed, as shown in Figure 5a,c. Additional differences were seen in the T1 value at D0 (Figure 5a). Difference in the CEST asymmetry was only apparent in the CTV and not in the thresholded subregions. Figures 5b,d show signal changes relative to D0, with differences in MT fraction and T2 value seen at D70 in the enhancing region R4. Differences were also present for the GTV at D70.

Conclusion

MT/CEST at 1.5T can be used for tracking the signal changes in GBM during chemoradiation. Significant differences were seen between early and late progression groups at certain time points and thresholded regions. Our results were dependent on the specific tumor region analyzed, suggesting that for robustness especially in heterogeneous tumors, subregions should be included for quantifying parameter maps to monitor response to therapy.

Acknowledgements

We gratefully acknowledge funding from NSERC (RGPIN-2017-06596, CRD 507521-16), Terry Fox (New Frontiers Program Project Grant) and CIHR.

References

1. Stupp R, Mason WP, Van Den Bent MJ, Weller M, Fisher B, Taphoorn MJ, Belanger K, Brandes AA, Marosi C, Bogdahn U, Curschmann J. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. New England Journal of Medicine. 2005 Mar 10;352(10):987-96.
2. Wolff SD, Balaban RS. Magnetization transfer contrast (MTC) and tissue water proton relaxation in vivo. Magn Reson Med. 1989;10(1):135–144.
3. Henkelman RM, Huang X, Xiang QS, Stanisz GJ, Swanson SD, Bronskill MJ. Quantitative interpretation of magnetization transfer. Magnetic resonance in medicine. 1993 Jun;29(6):759-66.
4. Wolff SD, Balaban RS. NMR imaging of labile proton exchange. Journal of magnetic resonance. 1990;86:164-9.
5. Mehrabian H, Myrehaug S, Soliman H, Sahgal A, Stanisz GJ. Evaluation of Glioblastoma Response to Therapy With Chemical Exchange Saturation Transfer. Int J Radiat Oncol Biol Phys. 2018 Jul 1;101(3):713-723. doi: 10.1016/j.ijrobp.2018.03.057.
6. Mehrabian, H., Desmond, K. L., Soliman, H., Sahgal, A. & Stanisz, G. J. Differentiation between Radiation Necrosis and Tumor Progression Using Chemical Exchange Saturation Transfer. Clin. Cancer Res. clincanres.2265.2016 (2017). doi:10.1158/1078-0432.CCR-16-2265
7. Jones CK, Schlosser MJ, Van Zijl PC, Pomper MG, Golay X, Zhou J. Amide proton transfer imaging of human brain tumors at 3T. Magnetic Resonance in Medicine: An Official Journal of the International Society for Magnetic Resonance in Medicine. 2006 Sep;56(3):585-92.
8. Zhou J, Zhu H, Lim M, Blair L, Quinones‐Hinojosa A, Messina SA, Eberhart CG, Pomper MG, Laterra J, Barker PB, Van Zijl PC. Three‐dimensional amide proton transfer MR imaging of gliomas: initial experience and comparison with gadolinium enhancement. Journal of Magnetic Resonance Imaging. 2013 Nov;38(5):1119-28.
9. Scheidegger R, Wong ET, Alsop DC. Contributors to contrast between glioma and brain tissue in chemical exchange saturation transfer sensitive imaging at 3 Tesla. Neuroimage. 2014 Oct 1;99:256-68. 10. Regnery S, Adeberg S, Dreher C, Oberhollenzer J, Meissner JE, Goerke S, Windschuh J, Deike-Hofmann K, Bickelhaupt S, Zaiss M, Radbruch A. Chemical exchange saturation transfer MRI serves as predictor of early progression in glioblastoma patients. Oncotarget. 2018 Jun 19;9(47):28772.
11. Jiang S, Eberhart CG, Lim M, Heo HY, Zhang Y, Blair L, Wen Z, Holdhoff M, Lin D, Huang P, Qin H. Identifying Recurrent Malignant Glioma after Treatment Using Amide Proton Transfer-Weighted MR Imaging: A Validation Study with Image-Guided Stereotactic Biopsy. Clinical Cancer Research. 2018 Jan 1:clincanres-1233.
12. Ma B, Blakeley JO, Hong X, et al. Applying amide proton transfer-weighted MRI to distinguish pseudoprogression from true progression in malignant gliomas. J Magn Reson Imaging 2016; 44:456–462.
13. Chan RW, Myrehaug S, Stanisz GJ, Sahgal A, Lau AZ. Quantification of pulsed saturation transfer at 1.5 T and 3T. Magnetic resonance in medicine. 2019 Jun 22.
14. McGarry SD, Hurrell SL, Kaczmarowski AL, Cochran EJ, Connelly J, Rand SD, Schmainda KM, LaViolette PS. Magnetic resonance imaging-based radiomic profiles predict patient prognosis in newly diagnosed glioblastoma before therapy. Tomography. 2016 Sep;2(3):223.
15. Jenkinson, M., Bannister, P., Brady, J. M. and Smith, S. M. Improved Optimisation for the Robust and Accurate Linear Registration and Motion Correction of Brain Images. NeuroImage, 17(2), 825-841, 2002.

Figures

Figure 1 – MR imaging protocol for GBM at 1.5T: The pulse sequence parameters for the GBM protocol are shown for the MT, CEST, WAter Shift And B1 (WASABI) sequence (used for B0 and B1 mapping), T1-weighted gradient echo sequence (for T1 mapping) and T2-weighted sequence (for T2 mapping). The reference frequency offset is represented by “ref”, at 100×103 Hz. The colon separates the start, step and end frequency offsets.

Figure 2 – Quantitative parameter maps: Standard clinical images from an example case are shown including the a) post-contrast T1-weighted (T1w) b) FLAIR images and c) apparent diffusion coefficient (ADC) maps. Quantitative maps are shown for the d) T1 relaxation time, e) T2 relaxation time, f) MT semi-solid fraction, g) CEST asymmetry and h) the CEST amide magnetization transfer ratio. Images are shown for all four imaging time points.

Figure 3 – Tumor regions: a) The gross tumor volume (GTV) and clinical target volume (CTV) are shown on the T1-weighted image of the same case as Figure 2, before treatment. b) Four tumor subregions are shown, with contours thresholded based the signal on each of the standard clinical images (T1w, FLAIR, ADC and T1w+C). The enhancing region (R4) is within the GTV, while the other regions (R1-R3) are thresholded within in the CTV.

Figure 4 – Early and late progression based on CTV: Median values in the CTV are shown between early (red) and late (blue) progression groups for a) T1, b) T2, c) ADC, d) MT semi-solid fraction, e) CEST asymmetry and f) MTRAmide. Significant differences between early and late progression are shown by asterisks (P<0.05). The solid lines represent the median values and shaded areas represent the interquartile ranges, across the four time points (D0, D14, D28 and D70).

Figure 5 – Early/late progression in tumor subregions: Maps show which parameters and regions have significance differences between the early/late progression groups for a) D0 (median) and b) D70 (normalized by D0). The matrix elements have intensity of (0.05 – p-value), where the p-value is obtained from the Wilcoxon Rank-Sum test. Zero intensity elements were not significant. Plots (as in those of Figure 4) of the MT semisolid fraction are shown for two subregions having c) low signal intensity on T1w (at D0) and d) enhancing signal (at D70, relative to D0).

Proc. Intl. Soc. Mag. Reson. Med. 28 (2020)
0282