Clinicians and scientist interested in cancer research, pancreas imaging, and MR Elastography.Outcomes for major cancer types (e.g. colorectal cancer) have improved steadily in the last decade(s), but pancreatic cancer has retained a poor prognosis. It induces a fibrous network of connective tissue, deposited in and around the cancer (“desmoplastic stroma”). New chemotherapeutic agents targetting stromal structures may be able to improve prognosis, but require dedicated monitoring tools different from current morphology based criteria such as RECIST¹. The stromal deposition leads to a rise in local viscoelastic properties^2,3. MR Elastography (MRE) is uniquely able to assess viscoelastic properties by probing the 3D propagation of shear waves. The intravoxel-incoherent motion (IVIM)-derived diffusion parameter (D) is also related to tissue density and could be a relevant marker for stromal deposition. Dynamic contrast-enhanced (DCE)MRI derived K^trans values relate to tissue vascularity. Here we present the first results of a feasibility study in which we added MRE in ongoing studies of patients with pancreatic cancer and show side-by-side comparisons in four subjects of DCE MRI, IVIM modelled diffusion MRI and MRE. MRE was added to the scan protocol of four pancreatic cancer patients (1M/3F, mean age: 67y) who were included in an ongoing study in which IVIM and DCE were performed on a clinical 3T Ingenia MR Scanner (Philips Healthcare, The Netherlands)⁴. MRE was performed with a portable transducer placed either ventrally (pancreas, n=2), against the left side of the chest (pancreatic tail, n=1) or the right side of the chest (liver, n=1). In a single healthy male volunteer (27y), only ventral MRE acquisition of the pancreas was performed. MRE was performed according to Garteiser et al⁵ with in-phase TE=6.9ms, f_MECH=56Hz, f_MEG=160Hz, 4×4×4 mm³ voxels in four 20s breath holds. Given the limited coverage (32 mm) in the slice direction, we used T1TFE Dixon water-only images to guide placement of the MRE field-of-view (FOV). Raw phase images were processed in dedicated software to obtain final elastograms³. Stiffness (MRE), water diffusivity D (IVIM) and K^trans (DCE) were assessed in the tumour (n=3) and pancreatic liver metastasis (n=1) using regions of interest (ROIs) after co-localising to the high-resolution Dixon water-only, post-contrast anatomical scan.

Pancreatic cancer

Total measurement time for the MR protocol amounted to circa 45 minutes. All subjects tolerated the additional breath holds and mechanical waves for the MRE acquisition well. Data of patient 1, 2 and 3 are shown in Fig. 1–3, with anatomical image and zoomed images with stiffness, D and K^trans overlays. For patient 1 (M/72y), the stiffness of the tumour was 2.55 kPa, with D of 1.80∙10^-3 mm²/s and K^trans of 0.32 min^-1. For patient 2 (F/74y), this was 3.52 kPa, 1.16∙10^-3 mm²/s and 0.35 min^-1. Patient 3 (F/52y) had a pancreatic tail cancer with 1.8 kPa, 1.25∙10^-3 mm²/s and 0.42 min^-1.

Healthy volunteer

The healthy volunteer’s pancreatic stiffness (1.52 kPa) was in concordance with literature values (f_MECH=40 Hz: 1.15 ± 0.17 kPa, f_MECH=60 Hz: 2.09 ± 0.33 kPa)⁶. No IVIM or DCE was performed.

Liver metastases

Pancreatic metastases to the liver in patient 4 (F/69y) were stiff, with the largest metastasis at 13.9 kPa. The elastogram showed a relatively spared triangle centrally in the liver, also noted on the anatomical image (Fig. 4). MRE slices were outside the IVIM/DCE FOV, hence no D or K^trans data were obtained for the metastases.

Pancreatic cancers and pancreatic liver metastases were substantially stiffer than normal pancreatic tissue in three out of four patients observed. As can be observed in Fig. 1–3, D was relatively high for patient 1, but had normal values for patients 2 and 3 compared to literature values⁷. K^trans values showed lower values for the cancers in patient 1 and 2 compared to surrounding pancreas tissue, in patient 3 this was not as clear. In subsequent patients we aim to improve spatial coverage and resolution, allowing assessment of the whole pancreas and small pancreatic lesions and compare MRE results with histological analysis of surgical (cf. Klaassen et al⁴). This will enable the implementation of MRE as a treatment response monitor tool in pancreatic cancer. The latter is highly relevant in an era in which new stroma targeting agents prove difficult to monitor using morphology-based tools such as RECIST. MRE of pancreatic cancer patients is feasible with current MRE equipment and can be added to existing MR protocols incorporating diffusion weighted imaging and DCE-MR. MRE showed higher stiffness values in three out of four pancreatic cancer patients. Assessment of its performance in treatment response monitoring of the new stroma targeting agents should be prioritized.