e15052 Background: Gastrointestinal (GI) cancers, i.e., pancreatic (PC), colorectal (CC), and esophageal (EC), are known to be both highly aggressive and clinically underdiagnosed, resulting in high rates of relapse after standard neoadjuvant therapy with curative intent resection. While cancer biomarkers (i.e. CEA, CA19-9) can aid in identifying early relapse post-resection, they lack accuracy and do not aid in identifying underdiagnosed patients. Recently, a circulating phagocytic macrophage engorged to ≥50 µm, i.e. cancer-associated macrophage-like cells (CAMLs), were identified in GI cancers and associated with highly aggressive cancer subtypes. However, monitoring CAMLs pre- and post-surgery and their clinical value in minimal residual disease remains unexplored. In this pilot study, we prospectively evaluated the blood of n = 133 patients with GI cancer pre- and post-surgery to assess CAMLs in resectable disease, and their association with progression free survival (PFS), relapse free survival (RFS), and overall survival (OS). Methods: In a single-blind multi-institutional prospective pilot study, we recruited n = 133 GI patients with newly diagnosed clinically resectable disease (n = 118), or with metastatic disease scheduled for curative resection post systemic therapy (n = 15). Whole peripheral blood (7.5mL) was drawn prior to neoadjuvant chemotherapy and/or radiation therapy at T1, and a post-surgery sample (T2) was obtained. CAMLs were filtered, stained with cytokeratin & CD45, and enumerated for CAML engorgement of ≥50 µm, as previously described. Cox proportional regression hazard ratios for PFS, RFS & OS using univariate and multivariate analyses were run over 96 months. Results: Of the patients with clinically resectable EC (n = 36/133), PC (n = 75/133) or CC (n = 22/133), 32% of patients (n = 43/133) did not undergo curative resection. T1 samples drawn pre-surgery were available from n = 105 patients and T2 samples drawn post-surgery were available from = 52 patients. Engorged ≥50 µm CAMLs were found in 51% (n = 53/105) of T1 patients and were significantly associated with worse PFS (HR = 3.4 CI95% 2.0-5.8, p < 0.001) and worse OS (HR = 63.5 CI95% 2.0-6.1, p < 0.001), compared to patients with < 50 µm CAMLs. Additionally, engorged ≥50 µm CAMLs were found in 60% (n = 31/52) of T2 patient samples and were associated with significantly worse RFS (HR = 5.7, CI95% 2.4-13.2, p < 0.001) and OS (HR = 5.3, CI95% 2.2-13.3, p < 0.001) compared to patients with < 50 µm CAMLs. Conclusions: These data suggest that GI cancer patients with CAMLs ≥50µm prior to curative intent resection have high rates of progression, and their presence post surgery may identify patients with residual disease that are likely to relapse. Given these results, more refined studies should evaluate CAMLs ability to identify such at risk patients who may benefit from alternative therapeutic approaches.