Mobocertinib in Patients with EGFR Exon 20 Insertion-Positive Non-Small Cell Lung Cancer (MOON): An International Real-World Safety and Efficacy Analysis

Abstract

EGFR exon 20 (EGFR Ex20) insertion mutations in non-small cell lung cancer (NSCLC) are insensitive to traditional EGFR tyrosine kinase inhibitors (TKIs). Mobocertinib is the only approved TKI specifically designed to target EGFR Ex20. We performed an international, real-world safety and efficacy analysis on patients with EGFR Ex20-positive NSCLC enrolled in a mobocertinib early access program. We explored the mechanisms of resistance by analyzing postprogression biopsies, as well as cross-resistance to amivantamab. Data from 86 patients with a median age of 67 years and a median of two prior lines of treatment were analyzed. Treatment-related adverse events (TRAEs) occurred in 95% of patients. Grade ≥3 TRAEs were reported in 38% of patients and included diarrhea (22%) and rash (8%). In 17% of patients, therapy was permanently discontinued, and two patients died due to TRAEs. Women were seven times more likely to discontinue treatment than men. In the overall cohort, the objective response rate to mobocertinib was 34% (95% CI, 24-45). The response rate in treatment-naïve patients was 27% (95% CI, 8-58). The median progression-free and overall survival was 5 months (95% CI, 3.5-6.5) and 12 months (95% CI, 6.8-17.2), respectively. The intracranial response rate was limited (13%), and one-third of disease progression cases involved the brain. Mobocertinib also showed antitumor activity following EGFR Ex20-specific therapy and vice versa. Potential mechanisms of resistance to mobocertinib included amplifications in MET, PIK3CA, and NRAS. Mobocertinib demonstrated meaningful efficacy in a real-world setting but was associated with considerable gastrointestinal and cutaneous toxicity.

Keywords: EGFR exon 20 inhibitors; exon 20 insertion; mobocertinib; non-small cell lung cancer; real-world data.

Figure 1

Response of target lesions in treatment-naïve and pretreated patients. Best response to mobocertinib. Waterfall plot of maximum change in tumor size on imaging measured according to RECIST v1.1 in all target lesions between baseline and follow-up in pretreated and treatment-naïve patients. Both growth (+20%) and shrinkage (−30%) of tumors are indicated by the dashed lines. One patient experienced a tumor growth of 400%. For better illustration purposes, the Y-axis only shows 100%. Patients with no shrinkage or growth are shown with −1%. Eighteen patients are not shown because there was no measurable disease or because they had no adequate response for assessment. Four patients are not shown because retrospective measurement of lesions was not possible.

Figure 2

Progression-free and overall survival from the start of mobocertinib treatment. Kaplan–Meier plots with 95% confidence intervals. Black rectangles mark censoring. The median PFS was 5 months (95% CI, 3.5–6.5). The median OS was 12 months (95% CI, 6.8–17.2).

Figure 3

Mechanisms of resistance to mobocertinib. Upper panel: Potential mechanisms of resistance that have been detected in post- but not in pretreatment tissue biopsies. Paired samples were assessed by next-generation sequencing (NGS) using custom DNA and RNA panels developed for recurrent genetic alterations in lung cancer, by fluorescence in situ hybridization (FISH) for MET and HER2 amplification, and by a board-certified pathologist to identify transformation to small or squamous cell histology (lineage plasticity). Ampli: Amplification (superscript indicates whether it was detected by NGS or FISH). GCN: Gene copy number. MET/c7: Ratio of MET signals to centromere 7 signals in FISH. §: Posttreatment sample was a cerebrospinal fluid cytology specimen. Therefore, only a small NGS panel analysis could be performed. $: Pretreatment biopsy could only be analyzed for hotspots in BRAF, EGFR, and KRAS (Sanger sequencing). &: Both samples were liquid biopsies analyzed for plasma-circulating tumor DNA. Lower left panel: 3D-models of EGFR TK complexes with TKI mobocertinib (teal) and its competitive substrate ATP (magenta) obtained from MD: (A) p.768_770dup with mobocertinib: a flexible P-loop assumes a conformation when its F723 forms a hydrophobic contact with mobocertinib, shielding it from solvent; (B) p.768_770dup + p.G721S with mobocertinib: S721 forms hydrogen bond(s) with F723 and G724, rigidifying the P-loop and shifting F723 away from the binding pocket, exposing mobocertinib to the solvent; (C) p.768_770dup with ATP: the P-loop does not interact with ATP; (D) p.768_770dup + G721S with ATP: S721 forms an additional hydrogen bond with oxygens in ATP, either directly or mediated by water molecules. Key amino acid residues interacting with the ligands are shown as thin sticks. A magnesium ion in complexes with ATP is shown as a green sphere. Key hydrogen bonds formed by the Ser721 (black) are shown with green dashes. Non-polar hydrogen atoms and solvent molecules are hidden for better visibility. Images were created with PyMOL. Lower right panel: Analysis of the EGFR p.G721S mutation in the Ba/F3 cell model. Cells retrovirally transduced with overexpression vectors encoding the EGFR exon 20 insertion p.S768_D770dup alone or in combination with the acquired EGFR p.G721S variant; controls were deprived of IL-3 and treated with increasing concentrations of mobocertinib. Proportional proliferation compared to the DMSO control and half maximal inhibitory concentration (IC₅₀) for mobocertinib are provided.