0561

Viscoelastic properties of pancreatic ductal adenocarcinoma: ex vivo biomechanical validation of in vivo MR elastography

Anne-Sophie van Schelt^1,2, Nienke P.M. Wassenaar^1,2, Eric Schrauben¹, Jules L Nelissen¹, Jurgen H Runge^1,3, Marc C.V. van Turnhout⁴, Rob C.H. Driessen⁴, Jaap Stoker^1,5, and Aart J Nederveen¹
¹Radiology and Nuclear Medicine, Amsterdam UMC, location AMC, Amsterdam, Netherlands, ²Imaging and Biomarkers, Cancer Center Amsterdam, Amsterdam, Netherlands, ³Radiology, Netherlands Cancer Institute, Amsterdam, Netherlands, ⁴Soft Tissue Engineering and Mechanobiology, Eindhoven University of Technology, Eindhoven, Netherlands, ⁵Endocrinology, Amsterdam Gastroenterology, Amsterdam, Netherlands

Synopsis

Keywords: Elastography, Pancreas, Elastography

Motivation: There are discrepancies in pancreatic MRE outcomes. Biomechanical validation of MRE in pancreatic ductal adenocarcinoma (PDAC) poses substantial challenges.

Goal(s): This study aims to conduct biomechanical analysis on the surgically removed specimen and correlate this to preoperative MRE outcomes.

Approach: Six PDAC patients underwent MRE prior to surgical intervention, resulting in shear wave speed (SWS) and phase angle (ф) for whole-tumor and tumor-center. Compression testing was performed on specimen at 5%-strain resulting in Young’s moduli (E). Correlation analysis was done with MRE outcomes and E.

Results: Pearson’s correlation revealed a significant correlation between SWS_tumor-wholeand E (r(6)=.857, p=.029) and ф_tumor-center (r(6)=.844, p=.035).

Impact: Validation using ex vivo compression mechanical testing of pancreatic MRE outcomes gives more insight in the accuracy of the measured elastic properties in complex and inhomogeneous tumorous tissue, which is imperative for clinical application of MRE as a predictive biomarker.

Introduction

Magnetic Resonance Elastography (MRE) is a non-invasive technique to assess soft tissue viscoelastic properties.¹ Pancreatic ductal adenocarcinoma (PDAC) has a poor prognosis, with 20% becoming eligible for curative surgical resection after chemotherapy.² Recent work states that tumor stromal deposition could hinder drug delivery.³ Viscoelastic properties are hypothesized to be related to stromal deposition, therefore MRE could serve as a predictive measure for chemotherapy efficacy. A variety of MRE methodology and their reported values require biomechanical validation (through e.g. biopsies).^4,5 However, this poses substantial challenges, primarily due to inaccessibility of the pancreas and the heterogeneous nature of tumors, rendering analysis on biopsies inadequate. Rheology is limited by frequency disparities with MRE, while atomic force microscopy is not sensitive to tissue macrostructure.^6,7 Studies have utilized bioindentation or compression to characterize anisotropic elastic properties of tissue.^8,9 This study aims to conduct biomechanical analysis on the surgically removed specimen and correlate this to preoperative MRE outcomes.

Methods

Six patients (♂=4, age=68±8 years) clinically diagnosed with PDAC and planned surgical intervention were included. MRE was scheduled <2 weeks before resection. MRE data was acquired on a 3.0T MRI scanner (Ingenia,Philips,Best,Netherlands). Patients fasted 4 hours prior to examination. Axial multi-slice multi-frequency SE-EPI MRE (Figure 1) data was acquired in free-breathing with four compressed-air transducers.¹⁰ Inversions were done using the k-MDEV algorithm, resulting in shear-wave-speed (SWS) and phase-angle (φ) maps.¹¹ Regions-of-interest (ROI) were drawn over the pancreatic tumor on MRE magnitude images for whole-tumor and tumor-center (5-pixel kernel).

After surgical intervention, two specimens were taken from the middle of the tumor, using the pancreatic duct as reference for alignment with MRE. Specimens were cryopreserved in liquid nitrogen and fixated using optimal-cutting-temperature compound. Specimens were planarized using a cryotome (Shandon Cryotome,GMI,MN,USA). Compression testing was performed with a Microtester G2 (CellScale Biomaterials,Waterloo,Canada) using parallel-plate compression and a cantilever (2mm diameter steel-rod and 5x5mm plate). Compression tests consisted of five cycles at 5% nominal-strain ($$$\epsilon$$$), as deformations at 5%-$$$\epsilon$$$ correspond with expected deformations in MRE.¹ During the load phase deformation is applied with a 1% step-size. In the recovery phase the cantilever returns to the initial height. The force (F) exerted was measured through deflection of the cantilever. The area of compression (A) was approximated through the contact radius with the parallel plate.

To calculate the apparent Young’s modulus (E), Hooke’s law is fitted to the force-strain curve for the final 2.5%-$$$\epsilon$$$ in the last loading cycle ($$$\frac{\Delta\mathbf{F}/\Delta\epsilon}{A}$$$). Linear regression and Pearson's correlation was performed for whole-tumor and tumor-center data. Significance level was set at .05 and statistical analysis was done in SPSS (version28,IBM,Armonk,NY,USA).

Results

MRE parameters were SWS_center=1.34±0.25 and SWS_whole=1.41±0.35 m/s (p=.371) and ф_center=1.10±0.29 and ф_whole=0.96±0.18 rad (p=.101), with example SWS- and ф-maps shown in Figure 2. An compression test is shown in Figure 3. Simple linear regression of E and SWS resulted in: slope_center=0.53 with 95%-confidence-intervals of [-0.08-1.14] and R²=0.591, and slope_whole=0.43 [0.08-0.79] with R²=0.735. For ф the results were: slope_center=0.51 with 95%-confidence-intervals of [0.06-0.96] and R²=0.712, and slope_whole=0.63 [-0.37-1.63] with R²=0.432, both are shown in Figure 4.

Pearson’s correlation revealed a significant correlation between SWS_whole and E, but not for SWS_center (Tumor center: r(6)=.869, p=.074 and whole tumor: r(6)=.857, p=.029). Contrarily, ф_center showed significant correlation but not ф_whole (Tumor center: r(6)=.844, p=.035 and whole tumor: r(6)=.657, p=.156).

Discussion & Conclusion

Pearson’s correlation showed significant correlations for ф_center and SWS_whole with E at 5%-$$$\epsilon$$$. The deviations between whole-tumor and tumor-center could be due to inhomogeneity of PDAC microenvironment.

Accuracy of MRE in bovine livers and resected prostate tumors have been verified in previous studies.^12,13 A linear rise of SWS with fibrosis has been shown in bovine livers and mesenteric adipose tissue, while the shear attenuation has a mild inverse correlation with inflammation.^14,15
Cryopreservation has been shown to preserve biomechanical properties.¹⁶ Biological soft tissue is highly sensitive to test conditions. To overcome this, we implemented repeated measures of multiple specimens. However, there are other factors influencing complex viscoelastic behavior (e.g. lack of perfusion). Moreover, scattering caused by the vascular network in MRE influences outcomes. The small sample size and discrepancies in alignment hinder accurate comparison.

Future analysis could investigate the linear behavior at low-strain sweep as this could be potentially associated with multi-frequency MRE. Future work will involve histopathological analysis to give more insight into the underlying microenvironment.

To conclude, in vivo determined viscoelastic MRE parameters are correlated with ex vivo biomechanical properties measured by compression tests of PDAC tumors. MRE in PDAC shows the underlying elastic properties accurately, which gives impact and makes the application of MRE as a clinical tool viable.

Acknowledgements

No acknowledgement found.

References

1. Manduca A, Bayly PJ, Ehman RL, et al. MR elastography: Principles, guidelines, and terminology. Magnetic Resonance in Medicine 2020;85:2377–2390.
2. Siegel RL, Miller KD, Fuchs HE, Jemal A. Cancer statistics, 2022. CA Cancer J Clin 2022;72(1):7-33.
3. Wang D, Li Y, Ge H, Ghadban T, Reeh M. The Extracellular Matrix : A Key Accomplice of Cancer Stem Cell Migration , Metastasis Formation , and Drug Resistance in PDAC. Cancers 2022(14):3998.
4. Kolipaka A, Schroeder S, Mo X, Shah Z, Hart PA, Conwell DL. Magnetic resonance elastography of the pancreas: Measurement reproducibility and relationship with age. Magnetic Resonance Imaging 2017;42:1-7.
5. Shi Y, Glaser KJ, Venkatesh SK, Ben-Abraham EI, Ehman RL. Feasibility of using 3D MR elastography to determine pancreatic stiffness in healthy volunteers. Journal of Magnetic Resonance Imaging 2015;41:369-375.
6. Jugé L, Foley P, Hatt A, Yeung J, Bilston LE. Ex vivo bovine liver nonlinear viscoelastic properties: MR elastography and rheological measurements. Journal of the Mechanical Behavior of Biomedical Materials 2023;138:105638.
7. Wood CD, Vijayvergia M, Miller FH, et al. Multi-modal magnetic resonance elastography for noninvasive assessment of ovarian tissue rigidity in vivo. Acta Biomater 2015;13:295-300.
8. Lake SP, Barocas VH. Mechanics and kinematics of soft tissue under indentation are determined by the degree of initial collagen fiber alignment. Journal of the Mechanical Behavior of Biomedical Materials 2012;13:25-35.
9. Mulk Mu. Optimization of Indentation for the Material Characterization. In: University W, Electronic Thesis and Dissertation Repository; 2020.
10. Dittmann F, Tzschätzsch H, Hirsch S, Barnhill E, Braun J, Sack I, Guo J. Tomoelastography of the abdomen: Tissue mechanical properties of the liver, spleen, kidney, and pancreas from single MR elastography scans at different hydration states. Magnetic Resonance in Medicine 2017;78:976-983.
11. Tzschätzsch H, Guo J, Dittmann F, et al. Tomoelastography by multifrequency wave number recovery from time-harmonic propagating shear waves. Medical Image Analysis 2016;30:1-10.
12. Bilston LE. Soft tissue rheology and its implications for elastography: Challenges and opportunities. NMR in Biomedicine 2018;31:1-10.
13. Sahebjavaher RS, Nir G, Gagnon LO, et al. MR elastography and diffusion-weighted imaging of ex vivo prostate cancer: quantitative comparison to histopathology. NMR Biomed 2015;28(1):89-100.
14. Sinkus R, Lambert S, Abd-Elmoniem KZ, et al. Rheological determinants for simultaneous staging of hepatic fibrosis and inflammation in patients with chronic liver disease. NMR Biomed 2018;31(10):e3956.
15. van Schelt AS, Beek KJ, Wassenaar NPM, et al. Viscoelastic properties of small bowel mesentery at MR elastography in Crohn's disease: a prospective cross-sectional exploratory study. Eur Radiol Exp 2023;7(1):53.
16. Lam SK, Chan SC, Leung VY, Lu WW, Cheung KM, Luk KD. The role of cryopreservation in the biomechanical properties of the intervertebral disc. Eur Cell Mater 2011;22:393-402.

Figures

Figure 1 (top) MRE set-up and settings with an illustrative elastogram and the corresponding region-of-interest (ROI) of the whole tumor and tumor center. (center) Patient undergo planned surgical resection after which a specimen is taken from the middle of the cross section through the tumor. (bottom) The specimen is planarized and a punch creates a cylindrical shape. A visual representation of the compression set-up is shown.

Figure 2 MRE data of a 71-year-old male with PDAC with the anatomical MRE magnitude image (left), corresponding shear-wave-speed (SWS) map (middle) and phase angle (ф) map (right). The whole tumor is delineated in red; ROI determined MRE values: SWS_whole = 1.22±0.18 m/s and ф_whole = 0.87±0.18 rad .

Figure 3 Compression test of a resected specimen. (top-left) The force-strain curve is shown for five repetitions. Fitting is done on the last curve at the final 2.5% strain and shown in red. (top-right) Displacement as a function of time. (bottom-left) The force delivered by the cantilever as a function of time is shown. (bottom-right) The cantilever compression throughout the experiment.

Table 1 Young’s Modulus (E) measured with compression testing and Shear wave speed (SWS) and Phase angle (Ф) for the whole tumor and tumor center measured with MRE given for each patient.

Figure 4 (left) Young’s Modulus E (compression test) plotted against SWS and ф (MRE) in the tumor using simple linear regression (R²=0.735 and 0.432, respectively). (right) E plotted against SWS and ф in the tumor center using simple linear regression (R²=0.591 and 0.712, respectively).

Proc. Intl. Soc. Mag. Reson. Med. 32 (2024)

0561

DOI: https://doi.org/10.58530/2024/0561