Pancreatic cancer (PC) is a devastating disease due to the lack of early detection resulting in late diagnosis and poor prognosis. Thus, there is a need for a technology that can visualize morphological and cellular changes in the pancreas during onset and progression of PC. Nearly 90% of human PC have a mutation in Kras and mice that express mutant Kras develop lesions resembling those in human PC. Utilizing MR microscopy, we examined pancreata from EL-KRAS (EK) and p48-Cre/LSL-Kras (KC; p48 is DNA binding subunit of Ptf1a) mice which target mutant KRAS to the pancreas via unique approaches. EL targeting employs human mutant Kras where all acinar cells express this oncogene. p48 targeting employs a mutation in an endogenous mouse Kras allele expressed to the pancreatic exocrine compartment. The source (human vs mouse), targeting (EL vs p48) and delivery (artificial vs endogenous) of mutant Kras contribute to altered phenotypes including differences in normal appearing parenchyma, frequency of acinar to ductal metaplasia, precancerous lesion formation, and progression to advanced disease. We examined pancreas from EK and KC using multiparametric high resolution MR imaging to compare their MR signatures.Formalin-perfused and fixed pancreata of EK and KC mice at different ages (3 or 4M, 8M, 12M; n=4) were examined using MR microscopy after rehydration in PBS. High spatial resolution images were acquired on a Bruker Avance 600MHz microimager using a 3D RARE pulse sequence with fat suppression. Imaging parameters were TR 2500ms, RARE factor 8, effective TE 40ms, FOV 20mm x 16mm x 3.55mm, matrix size 320x256x100 for pixel size 62μm x 62μm x 35μm. Images were analyzed using VolView to visualize the pancreatic architecture three-dimensionally. Mean diffusivity was measured from diffusion weighted images (TR/TE 2500ms/16.5ms; b = 1000s/mm², duration of diffusion gradients 3ms, delay between diffusion gradients 7ms, slice thickness 0.5mm, in-plane resolution 62µm x 62µm, 12-16 slices). T₂ maps were acquired from the same slices used in diffusion imaging with a Hahn spin-echo sequence and 16 echoes (TR/TEmin 6000ms/10ms).

MR microimages of EK mouse pancreas at different ages (Figure 1) revealed progressive changes to the tissue with advancing age. At 3 months, the architecture of the EK pancreas resembled that of a normal mouse pancreas [1]. The ductal network was seen as hyperintense lines; the parenchyma which consists of protein-rich acinar cells appeared with much lower signal intensity in the background. In contrast to normal pancreas, tiny cysts appeared as hyperintense blobs in the splenic end of the pancreas even at this age. Pancreas from 8-month old EK mouse retained the basic architecture of ducts and acini, but harbored more cysts. Volume rendering revealed that cysts were located in proximity to ducts. Pancreas from the12-month old EK mouse showed many cysts as well as loss of normal architecture.

High resolution images of pancreas from 4-month old KC mice (Figure 2) appeared similar to those of younger EK mice. At 8 months, however, several large cysts appeared over the entire pancreas in KC mice, and at 12 months, deterioration of the normal tissue and cysts with irregular boundaries were readily seen.

Diffusion weighted imaging differentiated the cysts in the two mutant Kras mice. Application of diffusion gradients partially attenuated the signal in EK mouse cysts, indicating the presence of both fluid and cellular components in those cysts (Figure 3) which is consistent with EK mice harboring cystic papillary neoplasms that resemble IPMN [2]. In contrast, diffusion gradients completely attenuated the signal in KC mouse cysts which rules out classifying KC mouse cysts as cystic papillary neoplasms (Figure 4). Further, average mean diffusivity of KC mouse cystic fluid was 2.05x10^-3 mm²/sec whereas it was 1.38x10^-3 mm²/sec in EK mice, which indicates that cystic fluid in KC mouse pancreas is less viscous than that in EK mouse.

MR phenotyping further revealed that differences existed between EK and KC mice in their pancreatic parenchyma representing acinar cells. In 8M old KC mice, mean diffusivity was lower (0.82x10^-3 mm²/sec) and T₂ relaxation time was higher (46.64msec) relative to 4M old mice (1.21x10^-3 mm²/sec and 41.10msec, respectively). Lower diffusivity signifies increased cellularity and increased T₂ likely results from lower vascular density in older KC mice. In contrast, no such age related alteration was detected in EK mice parenchyma (1.01x10^-3 mm²/sec and 40.50msec at both ages).

Multiparametric MR microscopy of pancreas from EK and KC mice reveal different cystic and parenchymal characteristics which might be a reflection of different trajectories of disease progression in these two mutant Kras mouse models of PC.