Background and Aims: Pancreatic cancer (PC) patients have a dismal 11% 5-year survival rate. Therapy personalisation has significant potential to improve this. Our team developed a world-first model to maintain and treat 3D pieces of human PC tissue in a dish (Kokkinos et al, Scientific Reports, 2021). To examine the potential translation to precision medicine, this project aimed to: (1) assess explant response to PC chemotherapies; (2) develop a secretions-based approach for rapid measurement of explant response; (3) demonstrate the utility of our model to identify chemoresistance pathways.
Methods: 27 patient-derived pancreatic tumour samples were obtained from surgical resections (Prof Haghighi; Prince of Wales hospital), then processed into 1-8mm3 explants. Explants were treated every 72h over 12 days, with the following drug combinations (used in neoadjuvant/adjuvant clinical setting): (i) Gemcitabine+5-fluorouracil (5-FU), (ii) Gemcitabine+Abraxane, (iii) 5-FU+Oxaliplatin+Irinotecan. Secretions were collected at endpoint for analysis of cancer and CAF-secreted markers by multiplex ELISA. Explants were fixed at endpoint for immunohistochemistry analysis of tumour cell/cancer-associated fibroblast (CAF) populations, cell survival and proliferation, fibrosis and spatial transcriptomics.
Results and Conclusions: (1) Patient/regimen-dependent responses in tumour cell and CAF populations were observed. FOLFIRINOX was the most effective treatment (based on >50% reduction in tumour cells), followed by Gemcitabine+Abraxane and Gemcitabine+5FU, consistent with clinical trends. Patient-dependent responses in fibrosis, cell proliferation and cell death markers were also observed. Follow up of patient overall survival is ongoing. (2) CA19-9 and IL6 were the most highly secreted markers and could follow changes in tumour and CAF populations. (3) Spatial transcriptomics identified differentially expressed genes in treatment-resistant tumour cells.