The number of biomarker-stratified therapeutic options for patients with metastatic colorectal cancer (mCRC) has increased as the molecular understanding of colorectal cancer (CRC) has progressed. Human epidermal growth factor receptor 2 (HER2/ERBB2) amplification has emerged as a biomarker of mCRC and occurs in 1–4% of patients with mCRC ^{[1][2][3][4][5]}[1,2,3,4,5]. To identify HER2 amplification, in addition to conventional tests, such as immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH), tissue- or plasma-based next-generation sequencing (NGS) is used ^[6][7][6,7]. Resistance to anti-epidermal growth factor receptor (EGFR) therapy and the efficacy of anti-HER2 treatment suggest that HER2 amplification is an actionable alteration in CRC ^[8][9][10][8,9,10]. With ongoing research into HER2-amplified CRC, new therapeutic strategies have been developed. As there is no preclinical biological signal for HER2 targeting with single-agent trastuzumab or tyrosine kinase inhibitors (TKIs), combination therapies or antibody-drug conjugates (ADCs) have been clinically developed for patients with HER2-amplified mCRC ^[11][12][11,12]. Currently, combination of trastuzumab with pertuzumab, or lapatinib, or trastuzumab conjugated to deruxtecan are recommended in the guidelines [13]. In Japan, the combination of pertuzumab and trastuzumab was approved for treating patients with HER2-amplified mCRC in 2022 ^[14][15][14,15].

2. Molecular Characteristics of HER2-Amplified CRC

CRC was one of the first solid tumors to be molecularly characterized. Several genes and pathways have been shown to be involved in tumor initiation and growth. A series of recurrent mutations in APC, KRAS, SMAD4, and TP53 are crucial recurrent driver events that accumulate during adenoma formation and progression to sporadic CRC, often correlating with specific stages of the cancer development process ^[16][17][16,17]. Molecular studies have also shown that alterations in WNT–β-catenin, membrane receptor tyrosine kinases (RTKs), and downstream MAPK and PI3K signaling pathways are nearly ubiquitous events in CRC ^[17][18][17,18]. In 2012, unsupervised clustering analysis of the Cancer Genome Atlas (TCGA) data of on 276 primary CRC cases for somatic copy number (CN), whole-exome sequencing, DNA methylation, messenger RNA and microRNA sequencing, and protein array yielded two subtypes: hypermutated and non-hypermutated tumors [17]. Hypermutated tumors were associated with right-sided tumors and hypermethylation, whereas somatic CN alterations (CNAs) were enriched in non-hypermutated tumors, suggesting chromosomal instability. One of the regions of focal amplification, identified in 4% of CRC cases, involved chromosomal region 17q21.1, which contains HER2. An international effort coordinating analytics compared six independent transcriptome-based subtyping systems, resulting in a consensus molecular subtype (CMS) classification that enabled the categorization of most CRCs into one of four CMSs: CMS1 (microsatellite instability immune subtype), CMS2 (canonical subtype), CMS3 (metabolic subtype), and CMS4 (mesenchymal subtype) [19]. The relationship between CMS and biological and clinical features and prognosis has been reproduced in multiple studies ^[20][21][22][20,21,22]. However, the relationship between HER2 amplification and CMS remains unclear. In the original manuscript on the development of CMS, the prevalence of HER2 amplification was 0% for CMS1, 1% for CMS2, 3% for CMS3, and 5% for CMS4, whereas CNAs were frequently observed in CMS2 and CMS4 ^[19][23][19,23]. The relationship between HER2-amplified CRC and molecular subtypes, such as CMS, requires further investigation. HER2 is a member of the epidermal growth factor receptor (HER/EGFR/ERBB), an RTK protein, also including EGFR, HER3, and HER4. Binding of a ligand to the extracellular region of EGFR, HER3, and HER4 results in a three-dimensional conformational change, allowing for dimer formation with other HER family members. However, no endogenous ligands for HER2 are known. It is thought to form a heterodimer by binding to homodimers or activated EGFR, HER3, and HER4 to form a signal transduction molecule. Activation of HER2 signal is triggered by either heterodimerization with another HER protein or homodimerization of HER2 [24]. HER2 amplification leads to aberrant signaling in downstream pathways, particularly those that result in the activation of extracellular signal-regulated kinase 1/2 (ERK1/2) signaling, similar to RAS or BRAF V600E mutations in CRC. HER2 amplification is mainly observed in RAS/BRAF wild-type mCRC, although approximately 20% of HER2 amplification co-occurs with RAS mutations ^[5][25][26][5,25,26]. This mechanism of ERK1/2 signaling cascade activation strongly suggests resistance to anti-EGFR therapy for patients with HER2-amplified mCRC. Yonesaka et al. found a large region of CN gain on chromosome 17 encompassing ERBB2 in a cetuximab-resistant CRC cell line [12]. In addition, they showed that, in 233 cetuximab-treated patients with mCRC, the median progression-free survival (PFS) and overall survival (OS) were significantly shorter in patients with HER2-amplified mCRC. Resistance to second- or later-line anti-EGFR therapy in HER2-amplified mCRC has been replicated in several subsequent studies ^[8][12][8,12]. In contrast, Sartore-Bianchi et al. showed the clinical impact of HER2 amplification on first-line chemotherapy used in combination with an anti-EGFR monoclonal antibody [10]. It remains unclear whether HER2 amplification occurs frequently after anti-EGFR therapy as an acquired resistance mechanism. Bertotti et al. showed that the frequency of HER2-positive mCRC was 13.6% in 44 patients with KRAS wild-type tumors, who had progressed to cetuximab or panitumumab treatment [11]. This finding suggests that a HER2 test after anti-EGFR therapy could potentially identify acquired HER2 amplification, which can be targeted by anti-HER2 treatment.

3. Diagnostic Procedures of HER2-Amplified CRC

IHC and FISH are typically used to identify HER2 amplification in CRC tumors, using modified and customized criteria for assessing HER2 positivity in breast and gastric cancers. In the HERACLES Diagnostic Criteria, with archival test (n = 256) and clinical validation (n = 830) cohorts, an IHC staining scale of 0–3 was retained [7]. However, because the cellularity of HER2 amplification is quite homogeneous, with all positive cases displaying amplification in >50% of cells, a 3+ HER2 score in more than 50% of tumor cells by IHC or a 2+ HER2 score and a HER2/CEP17 ratio >2 in more than 50% of tumor cells by FISH are required to be considered a HER2-positive diagnosis. An international collaboration between GI-SCREEN (Japan), NCTN-SWOG (USA), and Korea harmonized the diagnostic criteria, integrating data based on IHC and FISH for HER2-positive mCRC ^[27][29]. Their assessment of 475 CRC tumor samples showed heterogeneity in HER2 expression, resulting in a 10% cutoff for HER2-positive cells. In addition, HER2-positive CRC cells are usually gland-forming types that show strong lateral membrane staining, and basal membrane staining is not always observed. Thus, complete lateral or circumferential membrane staining is required for HER2 positivity. NGS allows for the sequencing of a large number of nucleotides in a short time frame, resulting in the simultaneous detection of multiple biomarker alterations. The internal collaborative study among GI-SCREEN, NCTN-SWOG, and Korea demonstrated a strong correlation between CN by FISH and NGS ^[27][29]. A study of 102 patients with CRC also showed 92% concordance between IHC and NGS in identifying HER2-amplified tumors, which increased to 99% concordance when cases with equivocal result in IHC were considered positive [6]. The increasing number of biomarkers for mCRC, such as NTRK fusions, high tumor mutational burden, wild-type RAS, BRAF V600E mutation, and high microsatellite instability may justify the use of NGS in patients with mCRC, but its cost-effectiveness requires consideration. Liquid biopsy analysis of ctDNA is another promising method to identify HER2 amplification in mCRC. In sequencing of ctDNA from 1107 patients with mCRC in the GOZILA study, a large-scale ctDNA genomic profiling program in Japan, HER2 had the highest median plasma CN (pCN) among all CNAs, suggesting the role of HER2 amplification as a driver and targetable alteration in mCRC [14]. In addition, in 75 patients tested using both ctDNA and tissue HER2 testing, the sensitivity and specificity of HER2 amplification of ctDNA versus tissue were 82% and 83%, respectively. Patients with HER2 amplification in tissue, but not in ctDNA, had a significantly lower ctDNA fraction, indicating that low tumor shedding is associated with false-negative ctDNA results. A study analyzing ctDNA in 47 evaluable plasma samples enrolled in the HERACLES trial found HER2 amplification in 46 samples, yielding a sensitivity of 97.9% for identifying HER2-positive CRC ^[28][30]. Since the pCN of HER2 amplification is generally affected by ctDNA fraction, the pCN adjusted by ctDNA fraction (adjusted pCN: ApCN) may be more useful for assessing the CN. Indeed, ApCN was more strongly correlated with the ISH HER2/CEP17 ratio and HER2 CN determined by quantitative reverse transcriptase-polymerase chain reaction than was pCN. Taken together, the combination of IHC, FISH, tissue NGS, and ctDNA analysis are methods that can support clinical trial enrollment and treatment decisions.

4. Efficacy of HER2-Targeted Treatment for Patients with HER2-Amplified mCRC in Clinical Trials

In HER2-positive tumors, available therapeutic agents targeting HER2 include anti-HER2 antibodies, TKIs, and ADCs. Monoclonal antibodies targeting HER2, such as trastuzumab and pertuzumab, bind to the extracellular domain of HER2 and inhibit dimerization, resulting in antibody-dependent cellular cytotoxic (ADCC) effects. TKIs including lapatinib, pyrotinib, neratinib, and tucatinib inhibit cell proliferation by blocking the tyrosine kinase activity of HER2. Lapatinib, pyrotinib, and neratinib are pan-HER TKIs, but tucatinib is a HER2-selective TKI. Trastuzumab emtansine (T-DM1) and trastuzumab deruxtecan (T-DXd) are HER2-targeted ADCs, which are covalently attached to a microtubule inhibitor and a topoisomerase inhibitor, respectively. When trastuzumab binds to HER2 on the tumor surface, its ADC is internalized and causes cytotoxicity by releasing the cytotoxic agents. In preclinical studies, anti-HER2 drug monotherapy using anti-HER2 antibody or a pan-HER TKI did not suppress the tumor growth of HER2-amplified CRC due to insufficient suppression of HER2/EGFR activation or induction of HER3 phosphorylation ^[11][29][11,31]. However, the combination of trastuzumab and lapatinib potently impaired growth by preventing HER2/EGFR/HER3 reactivation. These preclinical findings suggested the potential of dual targets for HER2 and EGFR/HER3 to overcome the resistance against anti-HER2 antibody or a pan-HER TKI alone. The HERACLES trial, conducted in Italy, was the first clinical trial of dual HER2-blocakde, which evaluated the efficacy of combined trastuzumab and lapatinib in patients with KRAS exon 2 wild-type and HER2-amplified mCRC, who were heavily pretreated with standard-of-care therapies, including prior EGFR-targeted antibodies ^[30][32]. Of the 914 patients with KRAS exon 2 (codons 12 and 13), 48 (5%) had HER2-positive tumors, defined as HER2 IHC3+ or HER2 IHC2+ with a HER2/CEP17 ratio ≥ 2 by FISH. The objective response rate (ORR), or the primary endpoint, was 30% (8/27 patients). In the HERACLES-B trial, which evaluated the efficacy of pertuzumab and T-DM1 in patients with RAS/BRAF wild-type and HER2-amplified mCRC refractory to standard treatments, the ORR was 9.7% and the median PFS was 4.1 months (95% CI, 3.6–5.9 months) ^[31][33]. Tucatinib is an oral TKI that is highly selective for HER2. The efficacy of trastuzumab and tucatinib in patients with RAS wild-type and HER2-amplified mCRC was evaluated in the MOUNTAINEER trial. At a median follow-up of 20.7 months, the confirmed ORR among 84 patients who received the combination treatment was 38.1%, with a median PFS of 8.2 months (95% CI, 20.3–36.7 months) ^[34][36]. In addition to the dual HER2-blockade, the activity of T-DXd in HER2-amplified mCRC was also explored in the DESTINY-CRC01 trial ^[35][37]. Of the 53 patients with RAS/BRAF wild-type and HER2-positive mCRC enrolled in the trial, 24 (45.3%) patients had a confirmed objective response.