A patient-derived subrenal capsule xenograft model can predict response to adjuvant therapy for cancers in the head of the pancreas.

Background: Although gemcitabine is commonly used as adjuvant therapy for pancreatic adenocarcinoma and pancreaticobiliary-type periampullary cancers, not all patients appear to benefit. This translational study evaluates the potential of a patient-derived subrenal capsule pancreatic cancer xenograft (SRCPCX) model to identify within eight weeks after surgery those tumours which will respond to gemcitabine.

Methods: SRCPCXs from 32 pancreatectomy patients were established in six to ten NOD/SCID mice per patient. After four weeks the mice were randomly assigned to receive gemcitabine or saline for four more weeks. After eight weeks, gemcitabine response in the grafts was evaluated by the percentage of tumour growth inhibition (%TGI), histological morphology and immunohistochemical markers (Ki-67, CK7 and cleaved caspase-3). These were collated into an Overall Response. Survival was assessed by Kaplan-Meier and Cox multivariate analyses.

Results: 375 of 450 pieces of tissue from 27 of 31 patients were evaluable. In 90% of patients, histopathological and immunostaining features of saline-treated control grafts were concordant with their original tumours. At follow up, six of 15 patients whose tumours had an Overall Response to gemcitabine died, compared with ten of 12 whose tumours did not respond (P = 0.025, Fisher's exact test). This was associated with improved survival on Kaplan-Meier analysis (P = 0.013). Cox multivariate analysis indicated that Overall Response, stage and grade were independent predictors of survival.

Conclusion: This SRCPCX model retains major histopathological and immunohistochemical characteristics of the original tumour and when a combination of measures is used, enables early assessment of tumour sensitivity to gemcitabine in pancreatic cancers.