Background: Although regularly used in breast and gastric cancer, HER2 testing is not routinely performed in colorectal cancer (CRC). Immunohistochemistry (IHC) and in situ hybridization (ISH) scoring have not been optimized for assessment of HER2 overexpression and amplification in CRC. Varying rates of HER2 overexpression (2%-11%) have been reported in previously untreated CRC (Wang World J Gastrointest Oncol 2019) and patients with advanced/metastatic CRC with HER2 amplification or overexpression respond poorly to current standard of care therapies (Sartore-Bianchi Oncologist 2019). We characterized HER2 prevalence in a retrospective cohort analysis of clinical trial patients. Methods: The HORIZON III trial (NCT00384176) evaluated FOLFOX + bevacizumab or cediranib (AZD2171) as first-line (1L) treatment in patients with metastatic CRC (Schmoll J Clin Oncol 2012). The primary objective of this analysis of a subset of HORIZON III samples was to characterize HER2 prevalence in 1L metastatic CRC. Secondary and exploratory objectives included correlating HER2 status to other molecular findings (eg, RAS/RAF mutations), clinicopathologic characteristics, and patient outcomes. IHC was performed using a monoclonal anti-HER2 antibody PATHWAY HER2 [4B5] Ventana on primary CRC tumor sections and scored according to ASCO/CAP guidelines for gastric cancer (Bartley Arch Pathol Lab Med 2016). H&E staining was performed to determine the adequacy of tumor samples (ie, > 100 viable tumor cells per specimen). All tumors with an IHC score of 2+ were analyzed for amplification by ISH (HER2 IQFISH pharmDx dual probe kit, Agilent Technologies K573111-5). Targeted mutation panel testing was used to determine other molecular alterations. Descriptive statistics were used to summarize baseline and clinical outcome data by HER2 status. The current analysis was approved by the AstraZeneca bioethics review board. Results: Of the 1614 patients in HORIZON III, 396 met the inclusion criteria of appropriate consent, sufficient tumor sample for analysis, and a unique identifier. HER2-positive tumors (IHC3+ or IHC2+/ISH+) were identified in 2.1% of samples; 1.3% were IHC3+ and 0.8% were IHC2+/ISH+. Compared with IHC3+, IHC2+ tumors were more heterogeneous with mixed components of cells weakly expressing HER2 and lacking HER2 expression. Conclusions: This exploratory analysis of the HORIZON III study provides insights into HER2 prevalence in 1L, unselected, advanced/metastatic CRC, finding that 2.1% of tumors were HER2 positive. The demographic characteristics of patients with analyzable samples were representative of the entire HORIZON III study population; however, further assessment of KRAS status and other clinicopathological characteristics is needed. These data may inform future clinical development and support selection of patients with CRC who are likely to benefit from HER2 targeting therapies. Clinical trial information: NCT00384176.