03/19/2026

Molecular and Immune Landscape of HER2-Amplified Colorectal Cancer

AACR 2026 PRESENTATION
Authors Sunyoung Lee, Michelle Weitz, Adam Dugan, Saif Nirzhor, Kayla Layng, Scott Kopetz, James Yu, Asif Rashid, Funda Meric-Bernstam, Mohamed Nuh, Monica Hsiang, Kanwal Pratap Singh Raghav

Abstract

Background: HER2 amplification defines a biologically distinct subset of colorectal cancer (CRC), yet its transcriptomic and immune correlates remain incompletely characterized. We profiled HER2-amplified CRCs to delineate clinicogenomic, immune, and transcriptomic features associated with ERBB2-driven oncogenesis.

Methods: We used Tempus Lens (Tempus AI, Inc., Chicago, IL) to query the Tempus multimodal de-identified database and identify patients with stage III/IV CRC who underwent Tempus xT (DNA) and xR (RNA) testing with tumor purity ≥30%. HER2 amplification was defined as ERBB2 copy number (CN) ≥6. Demographic and clinical characteristics, somatic and germline alterations, quanTIseq-based immune cell infiltration estimates, and biomarkers (TMB, PD-L1, MSI) were compared. RNA-seq data were normalized to log₂(TPM+1), and differential expression and pathway enrichment analyses on the hallmark gene set using fgsea compared HER2-amplified and non-amplified tumors.

Results: Among 16,394 patients with diagnoses of colon or rectal adenocarcinoma, 445 (2.7%) were HER2-amplified. These tumors occurred at a younger median age (58 vs. 60 years, p=0.001) and were more often left-sided (70% vs. 57%, p=0.011). Where data were available, ERBB2 amplification showed 88% (36/41) concordance with external HER2 IHC/ISH and a correlation between ERBB2 copy number and mRNA expression (ρ=0.38, p<0.001), demonstrating concordance across DNA, RNA, and protein levels. ERBB2-activating mutations, most frequently V777L (3.15%), followed by G776V, S310F, and H878Y, co-occurred in 9.2% of amplified tumors, suggesting dual genomic and structural activation. HER2-amplified CRCs had fewer KRAS mutations (14% vs. 48%, p<0.001; G12D most common), lower mean TMB (5 vs. 8 mut/Mb, p<0.001), and were universally MSS (MSI-H 0%). Germline mutation frequencies were similar between groups, with CHEK2 and MUTYH most frequent. HER2-amplified tumors demonstrated reduced infiltration of M1 macrophages, NK cells, and Tregs (all p<0.01) and lacked significant enrichment of antigen-presentation, Th1, checkpoint, or co-stimulatory signatures (q=0.06-0.08), indicating a relatively immune-inactive microenvironment compared to non-amplified CRCs. Gene set enrichment analysis revealed a trend toward positive enrichment of cell-cycle and DNA replication programs (E2F, MCM, PLK1, AURKB, TOP2A) and suppression of KRAS signaling in HER2-amplified patients.

Conclusions: HER2-amplified CRCs exhibit a coordinated ERBB2 activation axis linking DNA amplification, mRNA overexpression, and protein upregulation. These tumors are characterized by a hyperproliferative yet immune-depleted transcriptional profile. The biological interplay between HER2 signaling, proliferation, and immune modulation warrants further study, and future work should include clinical outcome correlations.

VIEW THE PUBLICATION