5008

Axillary Chemical Exchange Saturation Transfer (CEST) MRI Contrast is Consistent with Secondary Breast Cancer Treatment-Related Lymphedema
Rachelle Crescenzi1, Paula Donahue2, and Manus Donahue1

1Radiology and Radiological Sciences, Vanderbilt University Medical Center, Nashville, TN, United States, 2Physical Medicine and Rehabilitation, Vanderbilt University Medical Center, Nashville, TN, United States

Synopsis

Breast cancer treatment-related lymphedema (BCRL) is a chronic condition with 30% two-year incidence in cancer survivors treated with lymph node dissection. Changes in the tissue microenvironment indicate edema rich in macromolecular proteins. We hypothesize that chemical exchange saturation transfer (CEST) MRI, after accounting for transmit field (B1) heterogeneity and longitudinal (T1) relaxation time variation, will be sensitive to affected tissues in patients with BCRL. We report that after performing appropriate correction procedures in the upper extremities, it is possible to detect disease-specific CEST contrast in the affected and contralateral arms of BCRL patients.

Introduction

Breast cancer treatment-related lymphedema (BCRL) is a chronic lifelong condition with 30% two-year incidence in cancer survivors treated with lymph node dissection1. BCRL is clinically detected after macroscopic swelling of the limb is apparent, which is essentially late to intervene with prophylactic therapies. Changes in the tissue microenvironment from this mechanical lymphatic dysfunction comprise interstitial edema rich in macromolecular proteins, fatty acids, and fibrosis. We hypothesize that chemical exchange saturation transfer (CEST) MRI, after accounting for transmit field (B1) heterogeneity and longitudinal (T1) relaxation time variation, will be sensitive to affected tissues in patients with BCRL.

Methods

Demographics. The study cohort (n=33; gender=female; handedness=right) consisted of patients with unilateral BCRL (n=13; mean±standard deviation; age=48±7 years; body-mass-index BMI=30.2±5 kg/m2; mean BCRL stage=1.46, stage range=0-2) and healthy age- and BMI-matched controls (n=20; age=42±15 years; BMI=26.9±6 kg/m2). Subjects provided consent in accordance with the IRB and were scanned using 3T MRI.

Experiment. Imaging was performed over a bilateral FOV (slices=9, spatial resolution=1.8x1.47x5.5 mm3, Figure 1a). B1 efficiency maps were acquired using a dual-TR approach (TR1=30 ms, TR2=130 ms, FA=60 degrees). T1 mapping was achieved using the multi-flip angle method (FA=20, 40, 60 degrees, TR/TE = 100/4.6 ms). For CEST acquisition, a frequency-selective gaussian saturation pulse (nominal B1 amplitude=2 μT, duration=75 ms, Δω=±5.5ppm, stepsize=0.25ppm, and 6 reference acquisitions at Δω=8000ppm) was applied with a multi-slice EPI readout (EPI factor=7). To inform correction procedures in this region, CEST imaging was also repeated in 3 volunteers for varying B1 amplitudes (1, 1.5, 2, 2.5, and 3 μT) representing 50-150% efficiency of the nominal B1.

Analysis. Arm muscle was segmented from the left and right sides of each subject (Figure 1b). T1-relaxation time and B1efficiency maps were calculated voxel-wise (R2015b, Mathworks, Natick, MA) using standard methods2,3 (Figure 1c-d).

Next, we explored appropriate procedures for calculating CEST contrast in this region with high B0 and B1 variability. CEST metrics were calculated from the z-spectrum Z(Δω)=Ssat/S0 corrected for ∆B0 and ∆B1 (Figure 2a). The z-spectrum dependence on ∆B1 in the deep arm muscle is shown (Figure 2b). Z-values were corrected for B1efficiency based on a quadratic model (Table 1, eqn. 1) calibrated in the deep arm muscle for ZAPT and ZNOE following methods previously reported4,5.

The B1-corrected proton transfer ratio (PTR'APT, Table 2, eqn. 2), magnetization transfer ratio asymmetry (MTR'asym, Table 2, eqn. 3), and T1-compensated AREX metric6 (Table 2, eqn. 4) were quantified voxel-wise and averaged in identical ROIs as the T1 maps. The Spearman’s correlation coefficient was calculated between imaging metrics. The Wilcoxon rank sum test was applied to test differences in study parameters between controls and patients. In all cases, p-value<0.05 was required for significance.

Results

Significant correlations were found between MTR’asym and AREX, and MTR’asym and T1-relaxation time, but not between AREX and T1-relaxation time (Figure 3) as expected. Comparing matched-controls and patients with BCRL, the mean T1-relaxation time and PTR’APT in the arm muscle follow similar trends (Figure 4a-b). Mean MTR’asym in the arms of healthy female controls was -0.008±0.016, and was significantly lower in affected arms of patients with BCRL (-0.023±0.015, p=0.008) and contralateral arms (-0.024±0.011, p<0.001, Figure 4c). The mean AREX in the arms of controls was -0.007±0.024, and was significantly lower in affected arms of patients with BCRL (-0.029±0.027, p=0.01) and contralateral arms (-0.041±0.028, p<0.001, Figure 4d).

Discussion

The CEST effect measured by PTR’APT is largely reflective of T1 differences among BCRL patients compared to controls. The systemic effects of chemotherapy may contribute to increased adipose tissue deposition7, consistent with decreased T1-relaxation observed in both arms of patients with BCRL, while additional lymphatic fluid congestion may increase T1 in the affected arm. MTR’asym and AREX were mostly negative in the arm muscle indicating a greater NOE-CEST effect, which was more pronounced in BCRL patients indicating a more structured tissue microenvironment8. This trend is consistent with palpated fibrotic adipose tissue and increased lymphatic macromolecular uptake on the contralateral side in primarily stage 2 BCRL patients. While AREX demonstrates potential discriminatory capacity for lymphedema, independent of changes in T1, MTR’asym demonstrates reduced noise and similar significant trends.

Conclusion

We report that after performing appropriate B1 calibration procedures for CEST imaging in the arm muscle, as well as correction procedures for variation in T1 relaxivity, it is possible to detect unique CEST contrast in BCRL patients that is consistent with subclinical disease characteristics. These findings could have relevance for using CEST as a non-invasive tool to identify subclincial BCRL or to evaluate lymphatic response to emerging lymphedema therapies.

Acknowledgements

We would like to thank our research coordinators, Helen Mahany and Katie Lants as well as our MRI technicians, Chris Thompson, Leslie Mcintosh and Claire Jones for assisting with recruitment and care of the study participants. We would like to thank the NIH/NINR for funding this study (NIH/NINR 1R01NR01507901). We would like all our study participants in the giving of their time and dedicated support.

References

1. DiSipio T, Rye S, Newman B, Hayes S. Incidence of unilateral arm lymphoedema after breast cancer: a systematic review and meta-analysis. Lancet Oncol 2013; 14(6): 500-15.

2. Yarnykh VL. Actual flip-angle imaging in the pulsed steady state: a method for rapid three-dimensional mapping of the transmitted radiofrequency field. Magn Reson Med 2007; 57(1): 192-200.

3. Wang J, Qiu M, Kim H, Constable RT. T1 measurements incorporating flip angle calibration and correction in vivo. J Magn Reson 2006; 182(2): 283-92.

4. Singh A, Cai K, Haris M, Hariharan H, Reddy R. On B1 inhomogeneity correction of in vivo human brain glutamate chemical exchange saturation transfer contrast at 7T. Magnetic Resonance in Medicine 2013; 69(3): 818-824.

5. Windschuh J, Zaiss M, Meissner JE, Paech D, Radbruch A, Ladd ME et al. Correction of B1-inhomogeneities for relaxation-compensated CEST imaging at 7 T. NMR Biomed 2015; 28(5): 529-37.

6. Li H, Li K, Zhang XY, Jiang X, Zu Z, Zaiss M et al. R1 correction in amide proton transfer imaging: indication of the influence of transcytolemmal water exchange on CEST measurements. NMR Biomed 2015; 28(12): 1655-62.

7. Catalan V, Gomez-Ambrosi J, Rodriguez A, Fruhbeck G. Adipose tissue immunity and cancer. Front Physiol 2013; 4: 275.

8. Xu J, Zaiss M, Zu Z, Li H, Xie J, Gochberg DF et al. On the origins of chemical exchange saturation transfer (CEST) contrast in tumors at 9.4 T. NMR Biomed 2014; 27(4): 406-16.

Figures

Table 1. A quadratic model of the dependence of Z-values on B1 amplitude Z(B1)=p2B12+p1B1+p0 (eqn. 1) was calibrated in the arm muscle at 3.0T MRI. The polynomial function coefficients (second order coefficient p2, first order coefficient p1) are given below for Z in the range of APT and NOE around ±3.6ppm where proton transfer ratio in the muscle was determined maximal. Corrected Z'=Z-ΔZ was calculated with the correction factor ΔZ=Z(B1nom)-Z(B1rel) for B1nom=2μT and B1rel=B1nom·B1efficiency.

Table 2. Equations of CEST metrics evaluated in the arm muscle at 3.0T MRI.

Figure 1. a) A coronal view of the upper extremities of a healthy female volunteer shows the location of bilateral transverse imaging. b) Regions of interest (red overlay) were segmented in the deep arm muscle. In an identical field of view, quantitative mapping of c) T1-relaxation time (ms), d) B1efficiency (ratio), e) MTRasym (ratio) and f) MTR'asym (ratio) corrected for B1efficiency was performed. MTR'asym maps demonstrate less heterogeneity in regions of B1efficiency >1 (arrows), and greater symmetry between left and right arms after correction in this healthy volunteer.

Figure 2. B1 dispersion experiments. a) The z-spectrum is plotted as a function of saturation frequency offset (Δω, in units of ppm) from water resonance (Δω=0ppm) for varying nominal B1 amplitudes (1μT to 3μT) acquired from the arm muscle of a healthy female volunteer. b-c) Z-values (defined in Table 1) were measured as a function of B1 amplitude in the arm muscle, and these relationships were used to correct z-values for B1 efficiency. The dependence of Z on B1 is tissue-specific and necessitates in vivo evaluation for each application of CEST imaging, as previously demonstrated4,5.

Figure 3. Significant correlations were found between a) MTR'asym and AREX (Spearman’s rho=0.89, p<0.01), and b) MTR'asym and T1-relaxation time (Spearman’s rho=0.29, p<0.05). c) A weak correlation was observed between AREX and T1-relaxation time (Spearman’s rho=0.07, p=0.56).

Figure 4. T1-relaxation time and CEST metrics in the deep arm muscle of healthy female controls (N=40 including left and right arms), the affected arms of patients with BCRL (N=13), and the contralateral arms of patients with BCRL (N=13). Group values were compared for statistical significance using a nonparametric Wilcoxon rank-sum test, with *p<0.05 required for significance. Error bars represent one standard deviation from the group mean.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)
5008