EGFR inhibitors

One target, two modalities

Products in this area act on the epidermal growth factor receptor, which drives signalling for proliferation and survival. Two structurally unrelated modalities do this. Small-molecule tyrosine kinase inhibitors bind the intracellular kinase domain and are taken orally. Monoclonal antibodies bind the extracellular domain and are given by infusion. They share a target and almost nothing else about their nonclinical requirements, which is the organizing point of this article.

Approved small-molecule inhibitors

The small-molecule inhibitors are conventionally described by generation, which reflects the binding mode and the mutation spectrum each was developed against. The approval histories are less straightforward than they appear, and are worth stating accurately because they bear on what a new entrant has to show.

Gefitinib

Target: a first-generation tyrosine kinase inhibitor binding the intracellular kinase domain of the receptor, developed for non-small cell lung cancer with activating mutations in the receptor gene. Approval history: the original application was granted accelerated approval in May 2003; a labeling supplement followed in June 2005; and the current approval rests on a separate application granted in July 2015, tied to use with a companion diagnostic.^[S1] A single approval date does not describe this product accurately.

Erlotinib

Target: a first-generation tyrosine kinase inhibitor binding the intracellular kinase domain reversibly. Availability: the branded product was approved in November 2004, and its listed formulations now carry a marketing status of discontinued, with a Federal Register determination that this was not for reasons of safety or effectiveness. Generic erlotinib hydrochloride remains available under several abbreviated applications.^[S2] The compound is therefore still in use; the brand is not the right way to refer to it.

Osimertinib

Target: a third-generation tyrosine kinase inhibitor, described as such in the peer-reviewed literature, that binds the kinase domain irreversibly and is active against the T790M resistance mutation as well as the common activating mutations.^[S6] Approval: granted in November 2015.^[S3] The nonclinical section of its label contains a carcinogenicity finding discussed below, which is the kind of result that separates small-molecule development from antibody development in this area.

A carcinogenicity finding worth knowing about

The osimertinib label records that the compound increased the incidences of hemangioma, and of combined hemangioma and hemangiosarcoma, in the mesenteric lymph node and whole body at 10 mg/kg per day.^[S3] Two things follow for a developer working on a small molecule in this area. Carcinogenicity assessment applies to small molecules in a way it does not to the antibodies against the same target, so the two modalities face different long-term study requirements from the outset. And a vascular tumour finding in rodents raises a specific interpretive question about human relevance that has to be argued on mechanism, not dismissed. Neither question is answerable late; both shape the study plan years before the data exist.

The problems specific to this class

Acquired resistance

A review of resistance in this setting describes it as developing most commonly through the T790M mutation, and, after osimertinib, through diverse mechanisms including the C797S mutation and amplification of other receptors such as MET and HER2.^[S6] The last point is the one most often missed in study design. Resistance that arises by amplifying a different receptor will not appear in a cell line selected only for mutations in the original target, so the panel has to include the bypass mechanisms as well as the on-target ones.

The mutation determines the population

Exon 19 deletions and the L858R substitution are the common activating mutations, and a compound's activity differs between them and against the less common variants. For nonclinical work this means the mutation status of every model used must be stated and verified, not inherited from a catalogue entry. Where the intended population is defined by a diagnostic, as it is for gefitinib in its current approval,^[S1] the models should cover the mutations that diagnostic detects.

Toxicity that comes from the target

The receptor is expressed in skin and gastrointestinal epithelium, and blocking it there produces the dermatologic and gastrointestinal effects characteristic of the class. These are consequences of the intended pharmacology, so improving selectivity for the target makes them more likely rather than less. The scale of the effect is documented: the panitumumab label reports dermatologic toxicities in 90% of patients, severe in 15%.^[S5]

The nonclinical work

Models that carry the resistance mechanisms

A resistant line made by prolonged exposure in culture and one derived from a patient after treatment are different objects. The first is clean and reproducible; the second reflects what actually happened in a person, including mechanisms nobody selected for. Both have a place, and the provenance of each line should be recorded in the study report rather than assumed. Where an organoid or patient-derived xenograft is used, the passage number and the molecular characterization at that passage belong in the record, because both drift.

Confirming the mechanism, not only the outcome

Binding to the target and inhibiting the pathway are separate observations, and inhibiting the pathway and slowing tumour growth are separate again. Measuring receptor phosphorylation and downstream pathway markers in the tumour establishes that the compound did what it was designed to do at the dose given. Without that, a growth effect cannot be attributed to the intended mechanism, and an absent effect cannot be diagnosed as insufficient exposure rather than an invalid hypothesis.

Pharmacokinetics and exposure at the site

A compound that fails in a model may have failed to reach it. Absorption, distribution, metabolism and excretion work establishes systemic exposure; measuring concentration in the tumour establishes whether the systemic figure translated. For lung cancer, where brain metastases are common, penetration of the blood-brain barrier is a specific question in this class, and it is answered by measuring brain concentration rather than by inferring it from physicochemical properties.

Safety assessment

Skin and gastrointestinal histopathology are where the target-related findings appear, and the toxicology protocol should be written to look there. For a small molecule, the package also includes the genotoxicity and carcinogenicity work that produced the osimertinib finding,^[S3] plus safety pharmacology including cardiac assessment. For an antibody against the same receptor, none of the carcinogenicity work applies, and immunogenicity is measured instead. This is the clearest illustration of why modality, not target, determines the package.

What to establish with the laboratory

Four things are worth establishing before award. Which resistant lines the laboratory holds, how each was derived, and whether the panel covers bypass mechanisms such as receptor amplification rather than only mutations in the target.^[S6] Whether mutation status is verified in-house or taken from the supplier. Whether pathway inhibition is measured in tumour tissue as a routine part of efficacy studies or offered as an extra. And, for a small molecule, whether the laboratory or a partner can carry the carcinogenicity and genotoxicity work, since that determines whether the package can be run in one place.^[S3]

The antibodies against this target are a different development problem

The article above covers small molecules, which is what the original text listed. Two approved antibodies act on the same receptor, and their labels show how far apart the two modalities are. The cetuximab label carries a boxed warning for infusion reactions and cardiopulmonary arrest.^[S4] The panitumumab label carries one for dermatologic toxicity, reported in 90% of patients and severe in 15%.^[S5] An infusion reaction is a property of administering a large protein intravenously and has no counterpart in an oral tablet; the dermatologic effect is the shared target-related toxicity, appearing in both modalities. A developer choosing between the two formats for this target is choosing between two nonclinical packages with little overlap: species selection and immunogenicity on one side, carcinogenicity and metabolite characterization on the other.

About this article

This is an independent editorial article for people who commission nonclinical work in the United States. Product descriptions, warnings and nonclinical findings are taken from the prescribing information and approval records as published, and approvals are recorded as facts; this site does not discuss the efficacy of any product or compare products. Findings in animals are reported as animal findings and are not statements about people. It does not recommend any laboratory. Last reviewed: September 3, 2026.

References

1. FDA — IRESSA (gefitinib) Approval Record (Original NDA 021399: Accelerated Approval May 5, 2003 / Label Amendment June 17, 2005; Current NDA 206995: July 13, 2015, with companion diagnostic) https://www.accessdata.fda.gov/scripts/cder/daf/ (accessed 2026-09-03)
2. FDA — TARCEVA (erlotinib) Approval Record and Marketing Status (NDA 021743, approved November 18, 2004). https://www.accessdata.fda.gov/scripts/cder/daf/ (accessed 2026-09-03)
3. FDA — TAGRISSO (osimertinib) full prescribing information (Nonclinical Toxicology) and approval records (NDA 208065, approved November 13, 2015). https://www.accessdata.fda.gov/scripts/cder/daf/ (accessed 2026-09-03)
4. FDA — ERBITUX (cetuximab) full prescribing information (Boxed Warning, Nonclinical Toxicology) and approval records. https://www.accessdata.fda.gov/scripts/cder/daf/ (accessed 2026-09-03)
5. FDA — VECTIBIX (panitumumab) full prescribing information (Boxed Warning) and approval records. https://www.accessdata.fda.gov/scripts/cder/daf/ (accessed 2026-09-03)
6. Resistance to EGFR Inhibitors in NSCLC: Mechanistic Insights and Emerging Therapies. https://pubmed.ncbi.nlm.nih.gov/ (accessed 2026-09-03)