4129 Background: Pancreatic ductal adenocarcinoma (PDAC) carries a poor prognosis with a 5-year survival rate of 10.0%. Previous studies in stool microbiome indicate that microbiome composition has been associated with therapy response and pathogenesis across multiple cancers, and PDAC patients (pts) with higher bacterial diversity have demonstrated greater long-term survival. The fecal microbiome has not previously been characterized for PDAC pts with cancer cachexia or associated interventions. The study addressed the changes in microbiome over the course of treatment and the association between baseline bacterial composition and outcome in PDAC pts with cachexia. Methods: Stool specimens were collected from the PNCX1 trial (NCT02400398), where all pts were given a semi-elemental diet—enzymatically hydrolyzed protein—with enteral tube feeding. Stool samples (n = 29) were collected at time points aligned with enteral feeding and chemotherapy cycles separated by 6 weeks (C1D1, C2D1, and C3D1) and analyzed using 16S v4 sequencing of the microbiome. Microbiome changes from C1D1 to C3D1, weight stability, and overall survival (OS) were measured alongside microbiome characterization. Results: Pts with a complete set of stool samples were analyzed (n = 6) for differences in microbiome composition across treatment cycles. C3D1 samples were significantly associated with both an increased population of Veillonella and Actinomyces and decreased Bacteroides and Butyricicoccus compared to C1D1. Baseline stool microbiome composition was also evaluated to predict weight stability throughout treatment. In patient stool samples (n = 8) at C1D1, greater abundance of Veillonella (p = 0.0006) and reduced Bifidobacterium (p = 2.62E-5) were linked to greater weight stability. Microbiome alpha-diversity was also characterized using Shannon and Chao1 indices, where stable weight was related to reduced species richness (Chao1, p = 0.0194) but not evenness (Shannon, p = 0.1716). C1D1 patient stool samples were then analyzed and compared to OS (n = 16). Although no significant differences in global microbiome composition were noted between OS < 180 days and OS > 180 days, Parasutterella, Tyzzerella, Phascolarctobacterium, and Lachnoclostridium were identified as more prevalent in OS > 180 days despite their relatively low abundance. Conclusions: We are among the first evaluate stool bacteria changes over treatment course in PDAC pts. While Veillonella was associated with weight stability in a cohort of advanced PDAC pts all receiving enteral feeding, several genera were found in abundance in pts with prolonged OS, though this needs further validation. The potential impact of the gut microbiome and enteral feeding on weight stability is provocative given that cachexia is a hallmark of PDAC and an effective strategy to mitigate this process would be transformative.