Abstract B24: Gastrointestinal microbiome changes in stage IV pancreatic cancer patients treated with pembrolizumab with or without paricalcitol on the Stand Up to Cancer (SU2C) Pancreas Catalyst Trial

Background: Pancreatic cancer remains a deadly disease, and immunotherapy is effective in only 1-2% of microsatellite instability-high pancreatic cancers. The response to immunotherapy is very heterogeneous across tumor types, but recently there is evidence that the gastrointestinal microbiome may influence the response to immunotherapy. The SU2C Catalyst Trial is a randomized phase II maintenance trial to compare outcomes using a combination regimen of the vitamin D agonist, paricalcitol, plus the PD-1 inhibitor pembrolizumab vs. pembrolizumab alone. As a part of this trial, we are examining the composition of the gut microbiome. Methods: Patients with metastatic pancreatic cancer receiving standard first-line chemotherapy are eligible after achieving best response of at least stable disease and are then randomized to pembrolizumab +/- paricalcitol groups. Patients remain on study treatment until disease progression or treatment-ending toxicities. Patients collect stool samples within 72 hours prior to day one of each cycle, 24-72 later, and ca. 30 days after treatment discontinuation. DNA is extracted from fecal samples and 16S rRNA gene libraries are generated targeting the V4 region of bacterial and archaeal rRNA gene. Libraries are sequenced on the Illumina MiSeq and sequence reads are analyzed using QIIME2. Results: To date, we have analyzed the gut microbiota of six patients who have completed between one to four cycles of study treatment. We identified 28 taxa that occurred at ≥5% relative abundance in any given patient. As is true for most microbiome studies, intersubject variation was much greater than intrasubject variation. However, in three patients, there was significant change in the composition of the microbial communities over time. One patient had a “healthy” microbiome across the one treatment cycle with a good balance of Bacteroidetes and Firmicutes and the presence of Faecalibacterium prausnitzii. This patient discontinued the study due to treatment-related hypophysitis Grade 2, which was resolved by prednisone treatment and remains in survival follow-up. One patient had a relatively healthy microbiome during cycle one and two, but the composition of the gut microbiome became severely dysbiotic at the time the patient discontinued study treatment due to disease progression; this patient died five months later. The remaining four patients were dysbiotic across all treatment times, with some having very low alpha diversity. Conclusions: