Background: Refractory microsatellite stable colorectal cancer (MSS CRC) is immunologically cold and single-agent anti-PD-1/PD-L1 drugs are ineffective; novel immune-based approaches are needed. Evorpacept (E, ALX148) is an engineered protein (high-affinity CD47-blocker fused to an inactive IgG Fc region), which blocks the CD47/SIRPα innate immune inhibitory phagocytosis checkpoint expressed on CRC and phagocytes, respectively. The Fc region of E does not bind to Fcγ receptors, thereby limiting hematologic toxicity, and is intended to be given in combination. In CT26 CRC syngeneic models, E ± anti-PD-1 monoclonal antibody decreases tumor growth, reduces myeloid immunosuppression, increases dendritic cell activation, and increases T cell activation (Kauder, 2018); E enhances the antibody-dependent cellular phagocytosis activity of cetuximab (C) in vitro (Kauder, 2018); and E + pembrolizumab (P) was well-tolerated in the first-in-human trial (Lakhani, 2021). Methods: AGICC-ALX148 21CRC01 (NCT05167409) is a phase 2, single-arm, multicenter, investigator-initiated trial of E (15 mg/kg weekly), C (400 mg/m² then 250 mg/m² weekly), and P (200 mg every 3 weeks) in 21-day cycles for patients with unresectable MSS/proficient mismatch repair CRC refractory to oxaliplatin, irinotecan, and a fluoropyrimidine, regardless of tumor sidedness and RAS/BRAF status. Additional key eligibility criteria include ECOG performance status 0-1, evaluable disease per RECIST v1.1, adequate hematologic and end organ function, absence of prior checkpoint inhibitor use, and absence of significant autoimmune disease. Six patients will be enrolled in Stage 1 (safety run-in) and treated with ECP. The study will proceed to Stage 2 (dose expansion, N = 42, and all treated with ECP) if less than 33% of patients in Stage 1 experience a dose-limiting toxicity. Otherwise, additional patients will be enrolled in Stage 1 at lower dose level(s). The co-primary objectives are to determine 1) the recommended dose of E with CP, and 2) objective response rate by RECIST v1.1 (by one-sided exact test with α = 0.05, H₀ p ≤ 3% [historical controls], H_A p ≥ 15%; power is 87%). The study will close for futility if there are no responses (partial or complete) in the first 24 evaluable patients (by MinMax design with α = 0.025 [1-sided]; power is 87%). Secondary and exploratory aims include determination of progression-free survival, overall survival, safety, response assessment by iRECIST, and blood- and tumor-based immune modulation and baseline tumor expression (PD-L1, EGFR, and CD47) for association with tumor response. The study is open through the Academic GI Cancer Consortium and 5 patients have been enrolled at time of submission. Clinical trial information: NCT05167409.