EGFR exon 20 insertions — the third resistance pocket

EGFR exon 20 insertions are the awkward middle child of EGFR-mutant lung cancer. They sit in the same kinase but deform the αC-helix into a configuration that classical TKIs cannot exploit, and for a decade they were the EGFR mutation no one had a drug for. That has finally changed. Amivantamab carries the frontline label, sunvozertinib cleared FDA in July 2025, and zipalertinib has an NDA on file with a February 2027 PDUFA. Here is the resistance landscape, the chemistry that finally cracked it, and where the field is going next.

Why exon 20 insertions are different

About 85% of EGFR-mutant NSCLC carries one of two “classical” activating mutations: L858R in exon 21, or exon 19 deletions. Both stabilize the active αC-in conformation of the kinase, expose the ATP pocket, and respond to gefitinib, erlotinib, afatinib, and osimertinib. Exon 20 insertions account for roughly 4–10% of EGFR-mutant cases — about 30,000–40,000 patients diagnosed globally each year — and behave nothing like the classical mutants.

Structurally, an in-frame insertion of one to four residues into the loop following the C-terminus of the αC-helix (residues 763–774, with the most common variants being V769_D770insASV, D770_N771insSVD, H773_V774insNPH, and A767_S768insTLA) wedges the αC-helix into a constitutively active position. The kinase is locked “on” without needing the conformational rearrangement that classical mutations stabilize. The downstream signal is the same. The pocket is not.

The drug-binding consequence is severe. The inserted residues crowd the ATP pocket entrance and reshape the P-loop, so the morpholino tail of gefitinib, the quinazoline scaffold of erlotinib, and the acrylamide warhead of osimertinib all find their preferred geometries either blocked or destabilized. Clinical ORRs of the approved EGFR TKIs against exon 20 insertions sit in single digits. In effect, this is a third resistance pocket the field had to design against from scratch.

What is approved today

• Amivantamab (Rybrevant, Janssen) — an EGFR-MET bispecific IgG1, first approved in May 2021 as monotherapy after platinum chemotherapy and subsequently moved to first line as amivantamab + carboplatin + pemetrexed on the basis of the PAPILLON phase 3 trial (Zhou et al., NEJM 2023). PAPILLON showed a median PFS of 11.4 months versus 6.7 months for chemo alone (HR 0.40). The drug is delivered IV with a notorious infusion-reaction burden on cycle 1 day 1, but the activity is the deepest seen to date in this population. It is the current frontline standard.
• Sunvozertinib (Zegfrovy, Dizal) — an orally bioavailable irreversible EGFR TKI specifically designed against the deformed exon 20 pocket. FDA granted accelerated approval on July 2, 2025 for patients who have progressed on platinum-based chemotherapy, supported by the WU-KONG6 phase 2 study: ORR 46%, median DOR 11.1 months, median PFS 13.8 months at the 300 mg cohort dose (Wang et al., JCO 2024). It had already been approved in China in August 2023. The companion diagnostic is the Oncomine Dx Express Test.
• Mobocertinib (Exkivity, Takeda) — worth a footnote. Granted accelerated approval in September 2021, voluntarily withdrawn from the US market in October 2023 after the EXCLAIM-2 confirmatory trial failed to demonstrate PFS benefit versus chemotherapy in the first-line setting. A reminder that accelerated approvals are real approvals only if the confirmatory data lands.

What is filed and what is coming

Zipalertinib (CLN-081, Cullinan/Taiho) is an orally bioavailable, irreversible, EGFR-mutant-selective TKI with the most differentiated mechanism in the field: it preferentially binds the inactive conformation of exon 20-mutated EGFR while sparing wild-type EGFR enough to avoid the rash/diarrhea ceiling that limited mobocertinib. The FDA accepted the NDA in 2025 with a PDUFA action date of February 27, 2027. The pivotal data (Piotrowska et al., presented ASCO 2025) showed an ORR of 35.2% across previously treated patients, with a median DOR of 8.8 months. The most interesting subset was prior-amivantamab-only patients — those who had progressed on the frontline standard — where zipalertinib showed a 30% ORR with a 14.7-month median DOR. That positions zipalertinib as the most likely second-line option after amivantamab-based induction.

Several other compounds are in earlier development. BAY 2927088 (Bayer) is in phase 1/2 with early ORRs in the 40s. Furmonertinib at higher doses (240 mg) has shown activity in some exon 20 subtypes. The next horizon is variant-specific chemistry — the response rate of an unselected exon 20 cohort masks substantial heterogeneity, with near-loop insertions (e.g. A763_Y764insFQEA) behaving more like classical mutants and far-loop insertions (anything past residue 770) being the harder ones.

What the docking shows

EGFR exon 20 insertions are an excellent case for mutation-aware molecular docking. Dock erlotinib or osimertinib against wild-type EGFR and you get a textbook Vina pose: quinazoline N1 hydrogen-bonded to the M793 hinge, anilino group reaching into the back pocket, an acceptable score around -9 to -10 kcal/mol. Re-dock the same compound against, for example, the D770_N771insSVD variant and the back-pocket geometry collapses — the inserted residues sterically clash with the anilino tail, and the score craters by 2–3 kcal/mol. That is the resistance signal, and ΔΔ between wild-type and the insertion variant is the readout that matters.

Sunvozertinib and zipalertinib were engineered around this pocket. Their hinge interactions are similar but the back-pocket region is trimmed and reshaped to accommodate the inserted residues. Docking sunvozertinib against an exon 20 insertion structure (a useful starting point is PDB 9BU9 for the V769_D770insASV variant) shows the back-pocket substituent threading around the insertion rather than colliding with it. The ΔΔ between wild-type and insertion variant is near zero — the drug binds both, with appropriate margin against wild-type to avoid dose-limiting rash.

Try the docking yourself

Open Studio and pick EGFR from the target catalog. The mutation chips include the major exon 20 insertion variants — V769_D770insASV, D770_N771insSVD, H773_V774insNPH, A763_Y764insFQEA — alongside the classical L858R and T790M. Dock erlotinib, osimertinib, sunvozertinib, and zipalertinib across the panel and compare the ΔΔ pattern: classical TKIs lose 2–3 kcal/mol against far-loop insertions and hold against the near-loop A763_Y764insFQEA, while the exon-20-specific binders look flat across the row. That side-by-side panel is the cleanest way to see why amivantamab-led regimens and exon-20-specific TKIs are not interchangeable with first-line EGFR TKIs.

Liganx is a free way to run molecular docking online against any of the major EGFR variants without installing AutoDock or maintaining your own receptor library. If you want a molecular docking sanity check on a candidate exon-20 binder before committing to chemistry, this is the fastest path.

Primary sources

• Zhou C, Tang KJ, Cho BC, et al. Amivantamab plus chemotherapy in NSCLC with EGFR exon 20 insertions. N Engl J Med 389, 2039-2051 (2023). doi:10.1056/NEJMoa2306441
• Wang M, Yang JC, Mitchell PL, et al. Sunvozertinib for the treatment of NSCLC with EGFR exon 20 insertion mutations: the phase 2 WU-KONG6 study. J Clin Oncol 42, 3781-3791 (2024). doi:10.1200/JCO.24.00858
• U.S. Food and Drug Administration. FDA grants accelerated approval to sunvozertinib for metastatic non-small cell lung cancer with EGFR exon 20 insertion mutations. July 2, 2025. FDA approval notice
• Piotrowska Z, Tan DSW, Smit EF, et al. Efficacy of zipalertinib in NSCLC patients with EGFR exon 20 insertion mutations who received prior platinum-based chemotherapy with or without amivantamab. J Clin Oncol 43, 16_suppl, 8503 (2025). doi:10.1200/JCO.2025.43.16_suppl.8503
• Robichaux JP, Elamin YY, Tan Z, et al. Mechanisms and clinical activity of an EGFR and HER2 exon 20-selective kinase inhibitor in non-small cell lung cancer. Nat Med 24, 638-646 (2018). doi:10.1038/s41591-018-0007-9