HER2 exon 20 insertions in NSCLC: zongertinib and the YVMA problem

HER2 (ERBB2) mutations show up in roughly 2-4% of non-small-cell lung cancers, and unlike HER2 amplification in breast cancer, the lung-cancer version is driven by small in-frame insertions in the kinase domain. The single most common one, an insertion of four residues called YVMA, spent more than a decade defeating every tyrosine kinase inhibitor thrown at it. The drugs that hit the target hit wild-type EGFR just as hard, and the resulting rash and diarrhea capped the dose below what the tumor needed. Zongertinib, approved by the FDA in August 2025, is the first oral agent that cleanly separates the two.

What the YVMA insertion does to the kinase

The dominant HER2 exon 20 alteration is A775_G776insYVMA — an insertion of tyrosine-valine-methionine-alanine between residues 775 and 776, right at the end of the αC-helix in the kinase domain. It accounts for the majority of HER2 exon 20 insertions in lung cancer. Mechanistically it behaves like its EGFR exon 20 cousin: the extra residues wedge the αC-helix into the active “in” position, locking the kinase into a constitutively active conformation that no longer needs a dimerization partner to fire.

That same structural change narrows and reshapes the ATP-binding pocket. Bulky, reversible inhibitors designed against wild-type HER2 or EGFR simply do not fit well, which is why classical anti-HER2 antibodies and many first-generation TKIs underperformed against the insertion. The pocket is drug-accessible — it just demands a smaller, covalent warhead and exquisite shape complementarity.

The decade of near-misses

Before 2024 the HER2-mutant NSCLC toolbox was a list of compromises:

• Afatinib and dacomitinib (pan-HER covalent TKIs) — modest response rates, dose-limited by EGFR-driven diarrhea and rash because they hit wild-type EGFR with similar potency to HER2.
• Poziotinib — a small, rigid inhibitor designed for the constrained exon 20 pocket. It showed activity but a narrow therapeutic window and significant toxicity; it did not win a HER2-mutant NSCLC approval.
• Neratinib and pyrotinib — covalent pan-HER agents with real but limited single-agent activity and the same EGFR liability.
• Trastuzumab deruxtecan (T-DXd) — the antibody-drug conjugate that became the first HER2-directed therapy approved for HER2-mutant NSCLC (accelerated approval, 2024). It is given intravenously and carries an interstitial lung disease warning, leaving room for a well-tolerated oral option.

Why zongertinib is different

Zongertinib (development code BI 1810631, brand name Hernexeos) is an irreversible, covalent TKI engineered for HER2 selectivity. In biochemical assays it inhibits HER2 with an IC50 around 13 nM while sparing wild-type EGFR (IC50 around 579 nM), roughly a 40-fold window. That selectivity is the whole point: by leaving EGFR signaling in normal skin and gut largely intact, it dodges the on-target toxicity that capped the older pan-HER drugs, so the dose can actually reach the tumor.

The clinical readout matched the design. In the previously treated cohort of the phase 1b Beamion LUNG-1 trial, zongertinib produced a confirmed objective response rate in the low-to-mid 70s, the basis for the FDA’s August 2025 accelerated approval in non-squamous NSCLC with HER2 TKD activating mutations after prior therapy. In the treatment-naive cohort reported later, the confirmed response rate was 76% with a median duration of response of 15.2 months and median progression-free survival of 14.4 months. Notably, responses held up whether or not the tumor carried the YVMA insertion, and an intracranial response was seen in roughly half of patients with active brain metastases. Adverse events were predominantly low grade, and no interstitial lung disease signal of the kind seen with the ADC was reported.

What this means for design work

The HER2 exon 20 story is a clean lesson in selectivity over raw potency. The constrained pocket created by the YVMA insertion is druggable; the hard part was building a molecule potent enough on mutant HER2 while staying off wild-type EGFR. That is a difference of a few angstroms in two very similar ATP pockets, exactly the kind of question structure-based methods are meant to answer before you commit to synthesis.

Try the docking yourself

Open Studio and select HER2 (ERBB2) with an exon 20 insertion to dock a candidate against the mutant kinase domain, then re-dock the same molecule against wild-type EGFR to read out the selectivity gap. Doing molecular docking against both pockets side by side is how you spot an EGFR liability before it shows up as a dose-limiting rash. Liganx puts molecular docking online in the browser, so you can compare poses and interaction fingerprints across the HER2/EGFR pair without standing up a local pipeline.

Primary sources

• Heymach JV, et al. Zongertinib in Previously Treated HER2-Mutant Non-Small-Cell Lung Cancer. N Engl J Med (2025). doi:10.1056/NEJMoa2503704
• Wittlinger F, et al. Zongertinib (BI 1810631), an Irreversible HER2 TKI, Spares EGFR Signaling and Improves Therapeutic Response in Preclinical Models and Patients with HER2-Driven Cancers. Cancer Discov (2025). PMC11726021
• U.S. Food and Drug Administration. FDA grants accelerated approval to zongertinib for non-squamous NSCLC with HER2 (ERBB2) TKD activating mutations (August 8, 2025). fda.gov