Perioperative Stromal Depletion Strategies in Pancreatic Ductal Adenocarcinoma

For Specific Aim 1, we will be monitoring the post-operative complication of clinically relevant pancreatic fistula within one week of surgery per the standard clinical guidelines as noted above. Common clinical presentations of pancreatic fistula include abdominal pain, leukocytosis and fever (temperature >100.4 degrees). Diagnostic work-up of pancreatic fistula will be with CT abdomen with contrast, which has a sensitivity of 63% and specificity of 83% for detecting pancreatic fistula. The pancreatic fistulas will be categorized into grades A, B or C as previously reported. Our study will be investigating only clinically relevant pancreatic fistulas, i.e. grades B or C. Depending on the grade of pancreatic fistula, patients will be treated as indicated with conservative treatment options including bowel rest, antibiotics, somatostatin analogues and percutaneous drainage or surgical re-exploration. We will also track other relevant post-operative complications such as delayed wound healing, development of wound dehiscence or wound infection.

The tissue analyses will include review of the immunohistochemical (IHC) stains for actin, hyaluronic acid staining with binding protein probe, proliferation as measured by ki-67% and Phospho-histone H3 (Ph-H3); cell apoptosis as evaluated by IHC stain for cleaved caspase 3 (CC3) and Tunel. If limited tissue sample is obtained via the core biopsies, the priority of tissue analysis will be as follows: (1) fixed in formalin for H&E and IHC (Ph-H3; CC3/Tunel; Hyaluronic Acid (HA) binding); (2) fixed in Optimal Cutting Temperature (OTC) such that IHC with difficult antibodies can be done (to potentially obtain mRNA or DNA). The IHC studies will be done at the UCSF Helen Diller Family Comprehensive Cancer Center Immunohistochemistry and Molecular Pathology Core. The tissue analyses of the biopsy and surgical specimens will be done by study pathologist.

The CT and magnetic resonance (MR) imaging analyses will be performed at the Abdominal Imaging at UCSF. To decrease the impact that metal stents may have on the functional MRI results, we will include only patients who have plastic stents in our study. In addition, to reduce variability in results, the DCE-MRI and DWI-MRI will be obtained on the same MR machine at a similar time of day as the baseline scan. The DCE-MRI images will be analyzed by calculating Ktrans and DWI-MRI images will be analyzed by calculating apparent diffusion coefficient (ADC) as described elsewhere. We will scan patients in a torso coil on a 3T clinical MR scanner. Imaging will include MR diffusion with b-values of 0, 125, and 500 for estimates of perfusion and tumor, and dynamic contrast-enhanced MR imaging (DCE-MRI) for measurement of Ktrans, blood volume, and blood flow. The region of tumor will be determined by MR imaging in reference to the baseline CT scans. The native T1 of the pancreas and liver will be calculated from a series of four, 3D, spoiled gradient recalled echo (SPGR) sequences with different flip angles. The conventional DCE-MRI will be acquired as a 3D, fast spoiled gradient echo sequence, covering the targeted areas of the pancreas or liver, at a temporal resolution of 5 sec over 6 minutes after the administration of 0.1 mmol/kg gadobenate dimeglumine. DCE-MRI images will be post-processed using MIStar software (Apollo Medical Imaging, Melbourne, Australia), which allows for motion correction to account for any shifts in data. Datasets with artifacts will be eliminated before further post processing. After contrast delivery, the new T1 is calculated and is presumed to change with the Gd concentration such that [Gd] = (1/T1-1/T10)/R1 where R1 is assumed to be 4.5 sec-1 mmol/L-1 at 3T.

Run-in Period with PEGPH20

• Days 1, 4