Baseline Uncommon EGFR Mutations in Non-Small-Cell Lung Cancer: Structure-Function Classification and Clinical Evidence Interpretation.

In this narrative review, we examine baseline uncommon epidermal growth factor receptor (EGFR) mutations in non-small-cell lung cancer (NSCLC) through a structure-function classification framework and clinical evidence interpretation. Baseline uncommon EGFR mutations are molecularly heterogeneous, and exon-based classification alone cannot fully capture their structural diversity or variable sensitivity to EGFR-targeted therapy. By relating kinase-domain conformation to therapeutic sensitivity, structure-function classification offers a biologically grounded lens for interpreting this heterogeneity alongside sequence-based classification. This review focuses on baseline uncommon EGFR mutations, including EGFR exon 20 insertion (ex20ins) mutations, P-loop and αC-helix-compressing (PACC) mutations, and classical-like and long-tail non-ex20ins variants, while distinguishing them from acquired resistance alterations after EGFR tyrosine kinase inhibitor (EGFR-TKI) exposure. Typical loop-region ex20ins mutations can impose steric constraints near the ATP-binding pocket and are generally associated with limited activity of conventional EGFR-TKIs. Among ex20ins-directed approaches, amivantamab plus carboplatin-pemetrexed chemotherapy is supported by mature randomized first-line evidence in approved and accessible settings; oral ex20ins-directed EGFR-TKIs require interpretation according to treatment line, regulatory context, study design, and follow-up maturity. For PACC mutations, EGFR-TKI activity has been reported mainly in selected recurrent variants, whereas evidence remains uneven across agents, variants, compound mutations, and less common PACC-related alterations. Among classical-like and long-tail non-ex20ins variants, L861Q illustrates how clinically defined cohorts may combine mutations with distinct structural and pharmacologic profiles, whereas long-tail alterations remain dependent on limited variant-level evidence. Overall, the evidence base remains heterogeneous, spanning randomized trials, single-arm studies, retrospective cohorts, real-world datasets, conference abstracts, and early clinical reports. Taken together, the evidence supports a structure-informed approach to baseline uncommon EGFR-mutant NSCLC, while treatment interpretation and clinical application should remain grounded in evidence level, clinical context, and treatment availability. Prospective subtype-stratified studies, consistent evidence grading, and standardized molecular reporting with structure-informed annotations are needed to improve clinical interpretation and inform treatment strategies.