Using patient-derived xenograft models of colorectal liver metastases to predict chemosensitivity

Background: Few in vivo models for colorectal cancer have been demonstrated to show external validity by accurately predicting clinical patient outcomes. Patient-derived xenograft (PDX) models of cancer have characteristics that might provide a form of translational research leading to personalized cancer care. The aim of this pilot study was to assess the feasibility of using PDXs as a platform for predicting patient colorectal liver metastases responses, in this case by correlating PDX and patient tumor responses to either folinic acid, fluorouracil plus oxaliplatin or folinic acid, fluorouracil plus irinotecan-based regimens. Methods: Sixteen patients underwent potentially curative resection of colorectal liver metastases, and tumors were grafted into NOD.CB17-Prkdc(scid)/Arc mice. Mice were divided into groups to determine relative tumor growth in response to treatment. Tumors were analyzed by immunohistochemistry for Ki67 and Excision repair cross-complementation group 1. Results: An engraftment rate of 81% was achieved. Overall, there was a 67% positive match rate between eligible patient and PDX chemosensitivity profiles. There was a significant difference in relative decrease in Ki67 expression between sensitive/stable versus resistant PDXs for both treatment regimens. There was no statistically significant correlation between baseline ERCC1 expression and response to Oxaliplatin thorn 5-Fluorouracil in the PDXs. Conclusions: This pilot study supports the feasibility of using PDX models of advanced colorectal cancer in larger studies to potentially predict patient chemosensitivity profiles. (C) 2018 Elsevier Inc. All rights reserved.