Patient-Derived Xenografts Can Be Reliably Generated from Patient Clinical Biopsy Specimens.

BACKGROUND: Patient-derived xenografts (PDX) are clinically relevant human cancer models that can be used to guide individualized medicine. We aimed to generate PDX models from clinically obtained biopsy specimens (surgical or image-guided) hypothesizing that low volume biopsy specimens could provide sufficient viable tissue to successfully engraft PDX models from patients with unresectable or metastatic disease.

MATERIALS AND METHODS: We maintain a prospective high volume gastrointestinal malignancy PDX program. With informed consent and institutional approval, biopsy specimens (surgical or image-guided) were obtained from patients with unresectable or metastatic tumors: pancreatic adenocarcinoma (PDAC), cholangiocarcinoma, gastric and gallbladder carcinoma. Biopsies were implanted into immunodeficient mice. Tumor growth was monitored, viable tumor was passed into subsequent generations, and histopathology was confirmed.

RESULTS: In this study, biopsy specimens from 29 patients were used for PDX engraftment. Successful PDX engraftment was variable with highest engraftment rates in gastric and gallbladder carcinoma specimens (100%) compared to engraftment rates of 33% and 29% in PDAC and cholangiocarcinoma respectively. PDX models created from metastasis biopsies compared to unresectable primary tumor tissue demonstrated higher engraftment rates (69% versus 15.4%, p = 0.001). PDX models demonstrated higher engraftment rates when biopsies were obtained during surgical operations (n = 15) compared to image-guided (n = 14) (73% versus 14%, p = 0.003). Patient age, pretreatment status, or ischemic time was not different between biopsy methods.

CONCLUSIONS: PDX models can be successfully created from clinical biopsy specimens in patients with metastatic or unresectable GI cancers. The use of clinical biopsy specimens for PDX engraftment can expand the repertoire of stage-specific PDX models for downstream basic/translational research.