Pancreatic Cancer Patient-Derived Organoids Accurately Predict Response to Neoadjuvant Chemotherapy
Lyudmyla Demyan2, Daniel King1, Amber Habowski3, Astrid Deschênes3, Dennis Plenker3, Hardik Patel3, Luce St. Surin3, Oliver Standring2, Shamsher Pasha2, Arvind Rishi4, James M. Crawford4, Jeff Boyd1, Christine Metz5, Peter Gregersen5, Cristina Valente4, Sonia Abadali4, Steffi Matadial-Ragoo4, Danielle DePeralta2, Gary Deutsch2, Joseph Herman6, Mark Talamini7, David Tuveson3, Matthew Weiss2
1Medical Oncology, Northwell Health, New Hyde Park, New York, United States, 2Surgical Oncology, Northwell Health, New Hyde Park, New York, United States, 3Cold Spring Harbor Lab, Cold spring Harbor, New York, United States, 4Pathology, Northwell Health, New Hyde Park, New York, United States, 5Feinstein Institutes for Medical Research, New Hyde Park, New York, United States, 6Radiation Oncology, Northwell Health, New Hyde Park, New York, United States, 7Surgery, Northwell Health, Nyew Hyde Park, New York, United States
Objective: Patient-derived organoids (PDOs) have been explored as a biomarker of therapy response and personalized therapeutics for patients with pancreatic cancer. We hypothesized that PDOs may predict response to neoadjuvant (NAT) chemotherapy in patients with pancreatic adenocarcinoma (PDAC). Methods: During 2017-2021, PDO cultures were established from surgical specimens and fine needle aspiration or biopsies (FNA/FNB) obtained from consented patients with pancreatic cancer. Patient recruitment for the generation of PDOs was accomplished under IRB-approved protocols. Organoids of interest were analyzed through a translational pipeline incorporating DNA sequencing, RNA sequencing, and high-throughput drug sensitivity testing utilizing 123 compounds. A retrospective chart review was performed to obtain clinicopathological information, pathological chemotherapy response, somatic and germline DNA analysis, surgical, and oncological outcomes. Results: 136 samples from 117 patients with pancreatic cancer were collected. Among these, 80 samples were from surgical resections and 56 samples were from FNA/FNBs. 51% of patients were male. 37 (32%) of PDOs were derived from minority populations, consisting of 16% Black, 9% Asian, 7% Hispanic/Latino. 94 PDAC samples were subjected to molecular profiling and drug screening. 56% (n=53) of patients were clinical AJCC stage I and II, and 43% (n=41) AJCC stage III and IV. Organoids were established in an average of 21-56 days. Among surgical specimens, PDO generation was successful in 71% (15 out of 21) of patients who had received NAT prior to sample collection and in 76% (39 out of 51) in patients who were chemotherapy and radiation naïve at the time of collection. 54% of PDO were successfully established from FNA/FNBs pre-treatment. PDO transcriptomic subtypes were classified by the Moffitt system (Figure 1) and correlated with clinical patients’ outcomes. Pathological response to NAT correlated with PDO chemotherapy response; a representative example of poor NAT response (grade 3) based on pathology correlated with PDO pan-resistance to the standard of care chemotherapy agents (Figure 2). Conclusion: Herein we report the largest single-institution pancreatic cancer PDO library, including its recruitment of ethnic minorities. PDOs were successfully established from surgical specimens and FNA/FNBs, irrespective of tumor stage. The ability to establish PDOs from chemotherapy-naïve and post-NAT tissue enables longitudinal PDO generation to maintain dynamic chemotherapy sensitivity profiling. The proposed pipeline for PDO generation and drug testing is feasible to predict chemotherapy sensitivity and potentially enable patient-specific therapy, inclusive of patients with diverse backgrounds.
Figure 1. Classical and Basal-like subtypes of PDOs based on Moffitt classification.
Figure 2. PDO drug screening utilizing the standard of care chemotherapy agents.
The orange dot represents the case of interest; black dots represent the other sampled PDOs.
Back to 2022 Abstracts