Abstract A24: Early-stage pancreatic cancer patient derived xenograft (PDX) models established from endoscopic ultrasound (EUS) fine needle aspiration (FNA) biopsies

Abstract

Abstract Background: We developed and characterized PDX models generated from FNA endoscopic biopsies at the time of diagnosis of pancreatic ductal adenocarcinoma (PDAC). Our goal was to determine if an FNA-PDX model could be established within a clinically relevant time frame to inform treatment decisions prior to disease recurrence in patients with early stage, localized disease. Methods: Patients presenting with a localized mass in the pancreas and no metastatic disease were considered for EUS/FNA biopsy. Under sonographic guidance FNA biopsies were obtained. On-site cytopathology confirmed a histologic diagnosis of PDAC. The diagnostic FNA specimen was sent to pathology for formalin fixed paraffin embedded tissue block analysis per standard institutional pathologic protocols. Additional endoscopic FNA passes from the primary tumor were obtained during the index procedure by EUS guidance. The cell specimen was washed in cold DMEM and centrifuged. The cell pellet was rapidly transplanted into a subcutaneous pocket within the flank of a single NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NOD scid gamma, NSG) mouse. Tumor growth was monitored with calipers and engrafted tumors were passaged to subsequent mice when they reached a diameter of 10mm2. Growth kinetics and immunohistochemical staining for human cells with antibody against the human HLA Class 1 ABC antigen were obtained. Treatment groups were established from the F3 passage and randomized for in vivo chemosensitivity testing using FDA-approved chemotherapeutics: gemcitabine, cisplatin, docetaxel, irinotecan and fluorouracil. Results: The warm ischemia time was negligible. The average cold ischemia time, defined as the time from the endoscope to transplantation in the mouse, was 39 minutes. Engraftment was successful in 9 of 24 (37.5%) individual FNA patient specimens. Engraftment however increased with the number of FNA biopsy passes: 3/12 (25%) engraftment from one pass and 6/12 (50%) from two passes. Tumor size and surgical resectability status by CT imaging did not correlate with engraftment success. Procedural variables such as needle gauge and cytopathologic features such as degree of differentiation and percent tumor cellularity also did not correlate with rate of engraftment. PDX tumor morphology was maintained between F0 and F4 mice by H&E and human HLA immunohistochemical staining. The initial engrafted PDX tumors took 12 weeks (average) to grow to a diameter of 5 mm. The rate of engraftment for subsequent passaged mice was faster, averaging 6 weeks for the F1 development and stable for passaged tumors to the F5 generation. As expected, in vivo chemosensitivity was variable between FNA-PDX tumors from different patients as depicted in patient model #048 which was resistant to oxaliplatin, intermediate to docetaxel and fluorouracil but sensitive to cisplatin, irinotecan and gemcitabine. Conclusions: Pancreatic cancer PDX models can be generated from the diagnostic FNA biopsy specimen obtained during a patient's EUS procedure. The rate of engraftment doubles with the number of endoscopic passes suggesting that the amount of biopsy specimen tissue is important for engraftment while cytopathologic features associated with aggressive tumor growth such as tumor cellularity and tumor cell differentiation did not. Chemosensitivity was variable between FNA-PDX models. In patients with early stage disease the FNA-PDX model is stable and results from chemosensitivity assays can be obtained in a clinically meaningful window. Citation Format: Kerrington Smith, Dawn Fischer, Timothy Gardner, Stuart Gordon, Vijayalakshmi Padmanabhan, Thomas Colacchio, Richard Barth, Arief Suriawinata. Early-stage pancreatic cancer patient derived xenograft (PDX) models established from endoscopic ultrasound (EUS) fine needle aspiration (FNA) biopsies. [abstract]. In: Proceedings of the AACR Special Conference: Patient-Derived Cancer Models: Present and Future Applications from Basic Science to the Clinic; Feb 11-14, 2016; New Orleans, LA. Philadelphia (PA): AACR; Clin Cancer Res 2016;22(16_Suppl):Abstract nr A24.