Background:An estimated 18% of patients (pts) with GC/GEJC have HER2-positive (HER2+; IHC 3+ or 2+/in situ hybridization–positive [ISH+]) tumors. Although IHC/ISH are the standard testing methods for determining HER2 positivity, NGS can identify HER2 (ERBB2) amplification alongside other biomarkers. This study assessed the concordance of detecting HER2 amplification by tissue/plasma NGS with IHC/ISH HER2 status in pts with GC/GEJC.

Methods:Pts were retrospectively identified in the Tempus database. The primary analysis assessed tissue and plasma NGS HER2 copy number (CN) concordance with a HER2+ (IHC 3+ or 2+/ISH+) or HER2− (IHC 0, 1+, or 2+/ISH−) status; NGS samples were collected in ≤30 days of the IHC sample collection date. HER2 amplification was defined as tissue CN ≥8 or plasma log2(CN) ≥0.5; additional amplifications were captured by manual inspection. A secondary analysis assessed the correlation between paired tissue and plasma HER2 log2(CN) in pts with plasma and tissue NGS samples collected ≤30 days apart. The secondary analysis included pts irrespective of IHC sample collection date.

Results:Of 1578 pts with GC/GEJC and a curated IHC score evaluated in the primary analysis, 1507 and 305 pts underwent tissue and plasma NGS testing, respectively. Among pts with tissue and plasma NGS results, 902 (59%) and 201 (66%) had primary tumors of the stomach, respectively; all remaining pts had primary tumors of the cardia. According to IHC/ISH, 229 (15%) and 40 (13%) pts were HER2+ among those with tissue and plasma NGS testing, respectively. Of the pts identified as HER2 amplified by tissue and plasma NGS, 95% and 92% were HER2+ by IHC/ISH, respectively. Positive percent agreement (PPA) of HER2 amplification by tissue and plasma NGS with IHC/ISH HER2+ status was 70% and 58%, respectively. Concordance rates by PPA, negative PA (NPA), and overall PA (OPA) are summarized in Table. In the secondary analysis, a positive correlation in NGS HER2 log2(CN) was observed in pts with paired tissue and plasma NGS (mutant variant allele frequency >5%; R=0.85; P<2.2e−16; n=307).

Conclusions:Based on PPA, we observed moderate concordance between NGS HER2 amplification detection (plasma or tissue) and standard HER2 IHC assessment. NGS is not sufficient to detect all pts with HER2+ GC/GEJC. Further testing is warranted to determine if NGS testing could complement IHC/ISH in identifying pts who may benefit from HER2-directed therapies.