Background: Patients with microsatellite stable/microsatellite instability-low (MSS/MSI-low) mCRC account for ~95% of mCRC cases, and there is a significant unmet need in patients with treatment-refractory MSS/MSI-low mCRC. Although immune checkpoint inhibitors (ICIs) have shown limited activity in this patient population, the addition of tyrosine kinase inhibitors (TKIs) may increase sensitivity to ICIs by promoting an immune-permissive tumor microenvironment. Furthermore, TKI-ICI combinations have demonstrated encouraging clinical activity in patients with mCRC, particularly in subgroups of patients without liver metastases (LM). Most recently, in the phase 3 LEAP-017 study of patients with non–MSI-high/mismatch repair deficient (dMMR) mCRC, although pembrolizumab + lenvatinib did not improve overall survival (OS) vs standard of care, subgroup analysis suggested the potential for clinical benefit in patients without LM (Kawazoe et al, ESMO-WCGI 2023). Zanzalintinib is a novel multi-kinase inhibitor targeting VEGFR, MET, and the TAM kinases (TYRO3, AXL, MER). In a phase 1 study, zanzalintinib alone and in combination with atezolizumab (an anti–PD-L1 ICI) showed preliminary anti-tumor activity and a manageable safety profile across multiple tumor types (Sharma et al, ESMO 2022). Methods: STELLAR-303 (NCT05425940) is a global, randomized, open-label phase 3 study to assess the efficacy and safety of zanzalintinib + atezolizumab versus regorafenib in adult patients with MSS/MSI-low mCRC. Patients must have documented MSI/MMR status (non–MSI-high/dMMR) and RAS status by tissue-based analysis, measurable disease per RECIST v1.1, ECOG PS 0 or 1, and radiographic progression during/after or intolerance to standard of care treatments for mCRC. Prior regorafenib, trifluridine/tipiracil, or PD-L1/PD-1 ICIs are not allowed. Patients with or without active LM at baseline are eligible; those with definitively treated LM (includes surgical resection, microwave/radiofrequency ablation, or stereotactic body radiation therapy, but not yttrium-90 or chemoembolization alone) are considered to have non-active LM if treated ≥6 months before enrollment with no evidence of radiologic progression on subsequent imaging. Patients are randomized 1:1 to receive zanzalintinib + atezolizumab or regorafenib. Planned enrollment is 874 patients; number of patients with LM will be capped to ensure results of the study are relevant to the real-world mCRC population. Primary endpoint is OS in patients without LM. The key secondary efficacy endpoint is OS in all randomized patients. A hierarchical testing strategy will be employed for primary and key secondary endpoints. Safety will also be assessed. Enrollment is ongoing in the US, Europe, and Asia-Pacific region. Clinical trial information: NCT05425940.