HER2 (ERBB2) Mutation Testing in Non–Small Cell Lung Cancer: What the Assay Actually Detects

HER2 mutations turn up in roughly 2–4% of non–small cell lung cancers — a small slice, but a consequential one now that a targeted therapy exists for it. For years, the word "HER2" in a lung cancer report caused more confusion than clarity. Clinicians accustomed to breast and gastric practice reached instinctively for protein overexpression and gene amplification. That instinct, imported wholesale, was wrong for lung. The activating mutations that matter here are a distinct biological event, and conflating them with overexpression delayed the development of drugs that target them. This article walks through what the mutation test measures, how it's run, and what each result does — and doesn't — establish.

What the Test Measures

ERBB2, the gene encoding the HER2 receptor tyrosine kinase, can be altered in three mechanistically separate ways: point mutation, gene amplification, and protein overexpression. They are not interchangeable. The predictive biomarker of interest here is an activating HER2 mutation — most commonly small insertions in exon 20 of the kinase domain. These insertions lock the receptor into a constitutively active conformation, driving downstream proliferative signaling independent of ligand.

The key interpretive point here is that a mutation is a change in the DNA sequence. Amplification is an increase in gene copy number. Overexpression is an excess of protein on the cell surface. A tumor can carry any one without the others. In NSCLC, it's the mutation — not the amplified copy number or the strong membrane stain familiar from breast pathology — that currently gates the relevant targeted therapy [3].

How It's Tested

HER2 mutation status is determined by next-generation sequencing (NGS), not by immunohistochemistry and not by in situ hybridization. IHC measures protein. ISH measures copy number. Neither reliably captures an exon 20 insertion. Sequencing does.

Two specimen types are in routine use. The first is FFPE tissue — the formalin-fixed, paraffin-embedded block from a biopsy or resection. This remains the reference specimen. The main preanalytic constraints are familiar to anyone who works with archival material: adequate tumor cellularity, sufficient DNA yield, and the sequence artifacts that formalin fixation introduces, particularly deamination changes that can masquerade as low-level variants. Good laboratories account for these with appropriate variant-calling thresholds.

The second is ctDNA — circulating tumor DNA extracted from a plasma draw, sometimes called a liquid biopsy. Its appeal is obvious: no procedure, rapid turnaround, and access to tumor DNA when tissue is scarce. But a plasma assay depends on the tumor actually shedding DNA into the blood, and shedding varies with tumor burden and site. That's the central caveat. A negative ctDNA result doesn't exclude a HER2 mutation — it may simply mean too little tumor DNA reached the sample. When plasma comes back wild-type in a patient who otherwise fits the clinical picture, tissue confirmation is the sensible next move.

Scoring is binary in concept: the assay identifies an activating HER2 mutation or it does not. Worth emphasizing: the report should specify the mutation, because "HER2-positive" without qualification is exactly the ambiguity that caused trouble in the first place.

What Each Result Means

HER2-mutant. An activating ERBB2 mutation — typically an exon 20 insertion — was detected. This identifies the tumor as belonging to a molecularly defined subset with a specific therapeutic pathway.

Wild-type. No activating HER2 mutation was found. In tissue, this is a reasonably confident negative. In plasma, read it more cautiously — absence of a detectable mutation is not proof of its absence, particularly in low-shedding tumors. A wild-type result also says nothing about HER2 amplification or overexpression, which this assay isn't designed to assess.

What It Determines for Treatment Eligibility

A HER2-mutant result informs eligibility for HER2-directed antibody–drug conjugate (ADC) therapy — specifically trastuzumab deruxtecan (T-DXd). The evidence basis is the DESTINY-Lung02 trial, a randomized phase II study that evaluated T-DXd in patients with HER2-mutant metastatic NSCLC and supported its approval in this setting [3].

The framing matters. The mutation result is a gate to a class of therapy — it tells the clinical team the tumor carries the target the drug is built to engage. It is not, in itself, a treatment recommendation, and nothing here should be read as one. The pathologist's job ends at establishing the biomarker state; the therapeutic decision belongs to the treating oncologist and the patient, weighed against the full clinical picture.

Caveats and What's Evolving

The most persistent source of confusion is the HER2-low concept migrating from breast oncology. In breast cancer, HER2-low is an IHC-based category — tumors scoring 1+, or 2+ with negative ISH, that fall short of classic HER2-positive but still express enough protein to respond to certain ADCs. It's tempting to apply that logic to lung. Don't. HER2-low in NSCLC is not an established, harmonized predictive category, and the interpretive machinery differs fundamentally.

Here's the crux. IHC 1+ means faint, incomplete membrane staining; 2+/ISH-negative means moderate staining without gene amplification. These scores describe protein. They are orthogonal to the mutation the NGS assay detects. A tumor can be IHC 1+ and mutation wild-type, or mutation-positive with unremarkable protein staining. The cutoffs that define HER2-low are themselves contested, and the reproducibility of low-end IHC scoring is a known weak point even in tumor types where the category is validated. Applying a breast-derived protein cutoff to a lung tumor and calling it a HER2 result invites precisely the mutation-versus-overexpression error this whole field spent years untangling. The evidence for protein-based HER2 stratification in NSCLC remains immature, and readers should treat any such claim as provisional.

Two further points. FFPE variant-calling thresholds and ctDNA shedding limits both mean that a wild-type call carries context-dependent confidence — the specimen type belongs in the interpretation, not just the result. And the therapeutic landscape is moving quickly, with newer conjugates under evaluation across molecular targets in NSCLC [1, 2].

The conflation of mutation and overexpression cost lung cancer patients real time — a decade in which a druggable alteration was there to be found but often wasn't looked for correctly. A negative HER2 mutation NGS result, read properly, means one specific thing: no activating ERBB2 mutation was detected in the tissue examined. It does not close the door on other HER2 events, and it does not substitute for the molecular workup that increasingly defines how these tumors are classified from the moment the block reaches the bench.

---

References

1. Review. Lung Cancer: Targeted Therapy in 2025. Curr Oncol. 2025. PMC11941068.
2. Review. Seven new FDA approvals in 2025 across molecular targets in NSCLC. 2025. PMC12873738.
3. Li BT, et al. Trastuzumab Deruxtecan in Patients With HER2-Mutant Metastatic Non–Small-Cell Lung Cancer: Primary Results From the Randomized, Phase II DESTINY-Lung02 Trial. Journal of Clinical Oncology. 2023. PMID:37694347; PMCID:PMC10617843.