Abstract
Purpose
Epidermal growth factor receptor-tyrosine kinase inhibitors (EGFR-TKIs) serve as the standard first-line therapy for advanced EGFR-mutant non-small-cell lung cancer (NSCLC). However, resistance inevitably restricts their efficacy. Although well-documented acquired resistance mechanisms have been extensively characterized, approximately 30% of patients with EGFR-TKI resistance still present with unknown resistance (UR) mechanisms, and the driving factors behind primary resistance remain inadequately understood.
Methods
In this study, we conducted a large-scale, retrospective cohort study, performing target-capture deep sequencing (>400 genes) on 928 samples from 745 EGFR-TKI-resistant and 183 TKI-naïve EGFR-mutant NSCLC patients.
Results
The molecular landscape of EGFR-TKI resistance was systematically characterized. In the resistance cohort, 68.72% (512/745) harbored known resistance (KR) mechanisms, while 31.28% (233/745) were classified as UR. The landscape of KR was significantly influenced by EGFR-TKI generation. First-line third-generation TKI resistance demonstrated fewer EGFR-dependent events compared to first-generation TKIs (p < 0.01) and more RTK and PI3K-AKT-mTOR (PAM) pathway alterations (p < 0.05). In the UR cohort, pathway analysis revealed unique enrichment of atypical pathways, most notably ATP-dependent chromatin remodeling pathways. Differential gene analysis identified significant enrichment of MYC, BRCA1, EPHB, and PTPRD in the UR subgroup. To investigate primary resistance, we found that baseline IKZF1 amplification was significantly enriched in patients with short time-to-treatment-failure (TTF) and correlated with significantly poorer outcomes (HR = 3.02, 95% CI 1.32–6.90; log-rank p = 0.004).
Conclusions
This study provides a comprehensive real-world landscape of EGFR-TKI resistance, demonstrating distinct mechanisms influenced by TKI generation. ATP-dependent chromatin remodeling was identified as a distinctive pathway feature of unknown resistance, and baseline IKZF1 amplification emerged as a candidate biomarker of EGFR-TKI primary resistance. These findings suggest that epigenetic dysregulation may be a critical driver of both unknown and primary resistance to EGFR-TKIs.
Supplementary information
The online version contains supplementary material available at https://doi.org/10.1007/s13402-026-01253-y.
Keywords: Non-small-cell lung cancer (NSCLC), Epidermal growth factor receptor-tyrosine kinase inhibitor (EGFR-TKI), Acquired resistance, Primary resistance, Epigenetic dysregulation, IKZF1
Introduction
Lung cancer is the leading cause of cancer-related mortality in the world, with an estimated 1.8 million deaths in 2020 [1]. Non-small-cell lung cancer (NSCLC) is the common subtype of lung cancer, accounting for around 85% of all cases [2]. For patients with advanced or metastatic NSCLC, conventional chemotherapy yields limited efficacy [3]. Therefore, developing novel individualized therapeutic approaches is essential.
With the development of sequencing technologies, driver mutations in NSCLC have been progressively identified. These driver genes have enhanced our understanding of the molecular mechanisms of NSCLC and provided precise targets for the development of targeted therapies, thereby opening in the era of precision medicine for NSCLC [4, 5].
Among the major driver mutations in NSCLC, epidermal growth factor receptor (EGFR) mutations were the first to be discovered and successfully translated into targeted therapy, and these alterations account for >50% of cases among East Asian NSCLC patients [6]. EGFR-tyrosine kinase inhibitors (EGFR-TKIs) have become the standard first-line therapy for advanced EGFR-mutant NSCLC, significantly improving patient outcomes [7–9].
However, drug resistance limits the long-term efficacy of EGFR-TKIs [10]. Acquired resistance is almost inevitable [11]. In addition to acquired resistance, primary resistance also remains an urgent clinical problem [11].
The mechanisms of EGFR-TKI resistance are diverse, including EGFR-dependent and non-EGFR-dependent processes [6]. EGFR-dependent mechanisms mainly include secondary EGFR mutations (e.g., T790M, C797X) and EGFR amplification [6]. Non-EGFR-dependent mechanisms primarily involve bypass pathway activation (e.g., MET amplification, HER2 amplification) and activation of downstream signaling pathways (e.g., RAS/MAPK/ERK, PI3K/AKT) [12].
Nevertheless, aggregating these recognized mechanisms, the driver remains unidentified in around 30–50% of patients in clinical practice, needing further investigation [6, 12]. Additionally, primary resistance to EGFR-TKIs typically characterizes patients who experience disease progression early in treatment, including those who show no response to initial therapy or who initially benefit but relapse within six months [13]. The mechanisms underlying tumor-associated primary resistance are complex. On one hand, they are associated with tumor genomic profiles and the cell of origin, involving co-alterations in TP53, PIK3CA, PTEN, and BIM deletion polymorphism [13–16]. On the other hand, they involve the drug-induced transition of tumor cells to a drug-tolerant persister state, accompanied by transcriptomic, epigenomic, and metabolic reprogramming [13]. However, molecular mechanisms and prognostic biomarkers capable of predicting EGFR-TKI primary resistance remain limited.
Epigenetic regulation has played an important role in tumor development and progression and has been implicated in therapeutic resistance across multiple cancer types [17–20]. However, the specific epigenetic features that distinguish EGFR-TKI unknown resistance and primary resistance in advanced NSCLC remain incompletely characterized.
Therefore, based on this large real-world cohort, this analysis is the first to systematically compare the molecular landscapes of EGFR-TKI unknown resistance (UR) and primary resistance. Our study aimed to (1) systematically characterize the mutational landscape of EGFR-TKI resistance; (2) investigate the non-classical mechanisms in the UR cohort; and (3) identify novel baseline biomarkers to predict primary resistance, thereby informing precision treatment strategies for EGFR-TKI resistance.
Methods and materials
Patients and samples
In this retrospective study, clinical data and samples were collected from 928 patients with NSCLC. This cohort included 183 treatment-naïve baseline samples and 745 samples from patients with resistance to EGFR-TKIs. All 928 samples are independent, as paired pre- and post-resistance specimens from the same patients were not prospectively collected due to the real-world retrospective nature of the study. All sample underwent NGS using a targeted panel of > 400 tumor-related genes. All procedures were conducted in accordance with the Declaration of Helsinki. The study was approved by the ethics committee of Shanghai East Hospital (K20-288). Informed consent was obtained from all participants for genomic-level sequencing.
DNA-level next-generation sequencing (NGS)
Sample processing, DNA extraction, library preparation, hybridization, washing, sequencing, and bioinformatics analysis were performed as previously described [21]. Briefly, cell-free DNA (cfDNA) and genomic DNA (gDNA) were extracted from the peripheral blood, hydrothorax, cerebrospinal fluid, and FFPE samples using the CWE9600 Blood DNA kit (CoWin Biosciences, Cambridge, MA) and the Maxwell 16 FFPE Plus LEV DNA Purification kit (Promega, Madison, WI)) according to the manufacturer’s instructions. The KAPA Library Preparation Kit (Kapa Biosystems, Wilmington, MA, USA) was used to prepare indexed NGS libraries. DNA libraries were then hybridized to a custom-designed biotinylated oligonucleotide probe covering > 400 cancer-related genes (Integrated DNA Technology, Coralville, IA). DNA sequencing was then performed using the NextSeq CN500 system (Illumina, Inc.), and the MGISEQ-2000 (MGI, Shenzhen, China).
After removing the adaptor and low-quality reads, the sequenced reads were mapped to the reference human genome (hg19) with Burrows-Wheel Aligner (version 0.7.15-r1140). Somatic single nucleotide variants (SNVs) and small insertions and deletions (InDels) were identified through the application of MuTect2 (v1.1.4) and GATK (v3.6–0-g89b7209). Copy number variations (CNVs) and structural variations (SVs) were detected by Contra (2.0.8).
In this study, Tumor mutational burden (TMB) was calculated by the number of non-synonymous mutation per megabase (mut/Mb). The threshold of TMB-H for this study was determined to be ≥ 9 mut/Mb [22]. Samples that did not meet the evaluation criteria of TMB were marked as TMB-unknown (TMB-U).
Patient stratification and clinical endpoints
Patients were stratified based on the EGFR-TKI resistance mechanisms identified by NGS profile at the time of progression. To focus on the clearest biological contrast between early and late resistance, we purposefully enforced an extreme-phenotype comparison strategy comparing 1st-line with ≥3rd-line resistance, intentionally excluding intermediate 2nd-line resistance states to minimize confounding variables.
The Known EGFR-TKI Resistance Mechanisms subgroup (KR) comprised patients with well-defined, known function in conferring EGFR-TKI resistance. These known mechanisms were classified into four major categories: (1) EGFR-dependent mechanisms, covering EGFR T790M, C797S, EGFR 20ins, and EGFR amplification; (2) bypass receptor tyrosine kinase (RTK) activation, covering MET, HER2/ERBB2, HER3/ERBB3, FGFR1-4, and RET alterations, together with ALK, ROS1, and NTRK1-3 fusions; (3) activation of the RAS/RAF signaling pathway, covering KRAS, NRAS, HRAS, and BRAF alterations; and (4) activation of the PI3K-Akt-mTOR signaling pathway, covering PIK3CA, PTEN, AKT1, TSC1/2, and other pathway-related alterations. The complete gene-level inclusion list and patient-level detection counts per category were provided in Table S1. The Unknown EGFR-TKI resistance mechanisms subgroup (UR) included all patients whose resistant samples lacked any of these defined resistance mechanisms. Combined TP53 and RB1 alterations (SNV and/or CNV loss in both genes) were assessed as a genomic surrogate for susceptibility to SCLC transformation, as previously described [23].
In this study, time-to-treatment-failure (TTF) was used to assess patient prognosis. TTF was defined as the time from initiation of therapy to treatment discontinuation for any reason.
To identify the prognostic molecular biomarkers, patients from baseline cohort were stratified into distinct prognostic subgroups based on TTF. Patients who experienced treatment failure within 6 months were defined as the poor prognosis subgroup (n = 44) [14, 24, 25]. Patients with a TTF ≥ 12 months (including those without treatment failure at 12 months of follow-up) were defined as the favorable prognosis subgroup (n = 58).
External validation cohort
To assess the reproducibility of IKZF1 amplification in independent EGFR-mutant lung adenocarcinoma (LUAD) populations, three public cohorts were queried via cBioPortal: TCGA LUAD PanCancer Atlas, CPTAC LUAD, and OncoSG LUAD East Asian [26–28]. Within each cohort, EGFR-mutant patients were identified on the basis of somatic EGFR alteration status. The six IKZF1-altered baseline samples in our cohort all had an absolute copy number ≥ 4, corresponding to high-level amplification. To match this threshold across external cohorts, GISTIC score = 2 (high-level amplification) was used as the matched IKZF1 amplification category in TCGA, CPTAC, and OncoSG.
Statistical analysis
Differences on categorical variables were examined with Chi-square test. Categorical variables were presented in counts and percentages, and continuous variables were described as median with interquartile range (IQR). Differences in continuous variables between two independent groups were tested using Mann-Whitney test. Differences in continuous variables among three or above subgroups were evaluated using Kruskal-Wallis test.
Mutation profile, lollipop plots, and somatic interaction analyses among the top 20 mutated genes was assessed using the R package maftools (2.24.0). Volcano plots and stacked bar charts were created using R package “ggplot2” (3.5.2). R package clusterProfiler (4.16.0) was applied to perform the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis. Log-rank test was used to examine the difference among subgroups and Kaplan-Meier survival curves were used to display the survival status, using R packages “survival” (3.8–3) and “surminer” (0.5.0). Univariate Cox proportional hazards regression was used to estimate the hazard ratio (HR) and 95% confidence interval (CI) for baseline IKZF1 amplification. For survival analyses with sparse events, Firth-penalized univariate Cox regression (R package coxphf v1.13.4) was used. A two-sided value of p value < 0.05 was considered statistically significant. Statistical analyses were performed using R (v4.5.0).
Results
Characteristics of the study
A total of 928 samples were successfully profiled by NGS, including 183 baseline EGFR-TKI naïve samples and 745 EGFR-TKI resistance samples (Fig. 1). Among resistant samples, a resistance mechanism was identified in 512 (68.72%; KR cohort), whereas 233 (31.28%; UR cohort) lacked an identifiable resistance mechanism. Within the KR cohort, 287 samples (56.05%) were collected at first-line resistance and 225 (43.95%) at ≥3rd-line resistance. Within the UR cohort, 148 (63.52%) and 85 (36.48%) samples were collected at resistance to first-line and ≥3rd-line EGFR-TKI, respectively.
Fig. 1.
Flow diagram illustrating patient stratification in this study
The sample type of overall cohort predominantly consisted of tissue (43.43%) and plasma (36.53%) samples (Table 1). Sample types within the KR cohort were relatively balanced (tissue: 39.06%; plasma: 33.98%). In contrast, the UR cohort was enriched for plasma samples (56.22%), whereas tissue samples were more common in the baseline cohort (78.14%). These differences in sample type distribution between UR and baseline cohort was statistically significant (p = 0.0002). Sequencing quality metrics were comparable between tissue and plasma specimens (Fig. S1). In the overall, KR, and UR cohorts, plasma specimens showed significantly higher mean deduplicated sequencing depth than tissue specimens (all p < 0.05), while tissue specimens showed higher max VAF and mean VAF, reflecting higher tumor-cell content in solid biopsies rather than reduced detection sensitivity in plasma.
Table 1.
Clinical and molecular characteristics in the entire cohor and within the defined subgroups
| ALL (N = 928) |
Known resistance (N = 512) |
Unknown resistance (N = 233) |
EGFR-TKI naïve (N = 183) |
P.value* | ||
|---|---|---|---|---|---|---|
| Sample type | 0.0002** | |||||
| Tissue | 403 (43.43%) | 200 (39.06%) | 60 (25.75%) | 143 (78.14%) | ||
| Plasma | 339 (36.53%) | 174 (33.98%) | 131 (56.22%) | 34 (18.58%) | ||
| Hydrothorax | 100 (10.78%) | 74 (14.45%) | 21 (9.01%) | 5 (2.73%) | ||
| Cerebrospinal fluid | 73 (7.87%) | 53 (10.35%) | 19 (8.15%) | 1 (0.55%) | ||
| Others | 13 (1.40%) | 11 (2.15%) | 2 (0.86%) | 0 (0.00%) | ||
| Gender | 0.3965 | |||||
| Female | 538 (57.97%) | 293 (57.23%) | 133 (57.08%) | 112 (61.20%) | ||
| Male | 390 (42.03%) | 219 (42.77%) | 100 (42.92%) | 71 (38.80%) | ||
| Age | 58.00 [50.00;66.00] | 57.00 [49.00;65.00] | 59.00 [50.00;67.00] | 61.00 [53.00;68.00] | 0.1489 | |
| Smoking | 0.4530 | |||||
| Yes | 233 (25.11%) | 128 (25.00%) | 60 (25.75%) | 45 (24.59%) | ||
| No | 550 (59.27%) | 285 (55.66%) | 140 (60.09%) | 125 (68.31%) | ||
| Unknown | 145 (15.62%) | 99 (19.34%) | 33 (14.16%) | 13 (7.10%) | ||
| Metastatic site | 0.0023** | |||||
| One | 242 (26.08%) | 116 (22.66%) | 56 (24.03%) | 70 (38.25%) | ||
| Multiple | 564 (60.78%) | 332 (64.84%) | 142 (60.94%) | 90 (49.18%) | ||
| Unknown | 122 (13.15%) | 64 (12.50%) | 35 (15.02%) | 23 (12.57%) | ||
| Treatment Phase | <0.0001** | |||||
| Pre-treatment | 183 (19.72%) | 0 (0.00%) | 0 (0.00%) | 183 (100.00%) | ||
| Post-first-line resistance | 435 (46.88%) | 287 (56.05%) | 148 (63.52%) | 0 (0.00%) | ||
| 1st-gen resistance | 325(74.71%) | 213(74.22%) | 112(75.68%) | 0 (100.00%) | ||
| 2nd-gen resistance | 27(6.21%) | 17(5.92%) | 10(6.76%) | 0 (100.00%) | ||
| 3rd-gen resistance | 83(19.08%) | 57(19.86%) | 26(17.57%) | 0 (100.00%) | ||
| Post-third-generation resistance | 310 (33.41%) | 225 (43.95%) | 85 (36.48%) | 0 (0.00%) | ||
| 3rd-gen resistance | 310 (100.00%) | 225 (100.00%) | 85 100.00%) | 0 (100.00%) | ||
| Somatic mutation number | 7.00 [4.00;12.00] | 9.50 [7.00;16.00] | 4.00 [2.00;8.00] | 6.00 [4.00;8.00] | 0.0017** | |
| TMB | 0.0168* | |||||
| TMB-H | 107 (11.53%) | 78 (15.23%) | 19 (8.15%) | 10 (5.46%) | ||
| TMB-L | 723 (77.91%) | 431 (84.18%) | 193 (82.83%) | 99 (54.10%) | ||
| TMB-U | 25 (2.69%) | 3 (0.59%) | 21 (9.01%) | 1 (0.55%) | ||
| Unknown | 73 (7.87%) | 0 (0.00%) | 0 (0.00%) | 73 (39.89%) | ||
| TMB value | 4.80 [2.88;6.72] | 4.80 [2.88;7.68] | 2.88 [1.92;5.76] | 4.80 [3.00;6.72] | <0.0001** | |
| Resistance | 0.0002** | |||||
| EGFR-Dependent | 294 (31.68%) | 255 (49.80%) | 0 (0.00%) | 39 (21.31%) | ||
| RAS/RAF | 16 (1.72%) | 14 (2.73%) | 0 (0.00%) | 2 (1.09%) | ||
| RTK | 49 (5.28%) | 45 (8.79%) | 0 (0.00%) | 4 (2.19%) | ||
| PAM | 60 (6.47%) | 44 (8.59%) | 0 (0.00%) | 16 (8.74%) | ||
| Multi | 158 (17.03%) | 154 (30.08%) | 0 (0.00%) | 4 (2.19%) | ||
| Others | 351 (37.82%) | 0 (0.00%) | 233 (100.00%) | 118 (64.48%) |
*TMB, tumor mutational burden; TMB-L, TMB-low; TMB-H, TMB-high; TMB-U, TMB-unknown; RTK, receptor tyrosine kinases; PAM, PI3K-AKT-mTOR. Statistical comparisons between unknown resistance cohort and baseline cohort were performed using the Chi-square test or Mann-Whitney test
Across the entire cohort, the majority of patients were females (57.97%) and never-smokers (59.27%). Most patients (60.78%) presented with multiple metastatic sites. The median somatic mutation number per sample was 7, and the majority of patients (77.91%, 723/928) were classified as TMB-L. In the KR subgroup, approximately half of the patients (49.80%, 255/512) were classified as having EGFR-dependent resistance. Other identified resistance mechanisms included alterations in the receptor tyrosine kinases (RTK) pathway (8.79%), the PI3K-AKT-mTOR (PAM) pathway (8.59%), and RAS/RAF pathway (2.73%). Moreover, 30.08% (154/512) of KR cohort harbored two or more concurrent resistance mechanisms. Notably, no significant differences in gender, age, or smoking history were observed between the UR and baseline subgroups (all p > 0.05), suggesting comparable clinical backgrounds for subsequent analyses. Moreover, the prevalence of TP53 and RB1 co-alteration was 8.74% (16/183) in the baseline cohort, 6.44% (15/233) in the UR cohort, and 11.52% (59/512) in the KR cohort (three-group Chi-square p = 0.0835; baseline vs. UR p = 0.484), indicating that SCLC-transformation-like genomic profiles were not enriched in the UR subgroup (Fig. S2).
Resistance mechanisms stratified by EGFR-TKI therapy history
Among 435 patients who developed resistance to first-line EGFR-TKIs (325 treated with first-generation, 27 with second-generation, and 83 with third-generation agents), alterations were identified most frequently in EGFR (100%), TP53 (248/435, 57.01%), MYC (49/435, 11.26%), RB1 (46/435, 10.57%), LRP1B (44/435, 10.11%), and MET (42/435, 9.66%) (Fig. 2a). To explicitly highlight acquired secondary and tertiary resistance mechanisms, resistance-associated EGFR alterations were tracked separately. In this first-line resistance cohort, T790M was detected in 38.62% (168/435), EGFR amplification in 22.07% (96/435), C797S in 1.38% (6/435), and L718Q in 1.15% (5/435).
Fig. 2.
Molecular profile of first-line EGFR-TKI resistance stratified by TKI generation. (a) Oncoplot visualizing the mutation profile of the first-line EGFR-TKI resistance cohort. Dedicated rows display the resistance-associated EGFR alterations (T790M, C797S, L718Q, and EGFR amplification) separately from the overall EGFR row. (b) Stacked bar chart comparing the distribution of major resistance mechanism categories (EGFR-dependent, bypass RTK, RAS/RAF, and PAM pathway) across first-, second-, and third-generation first-line EGFR-TKI treatment. (c) Volcano plot showing differentially enriched genes between first/second-generation and third-generation TKI resistance within the first-line KR cohort. The x-axis represents log2(OR) and the y-axis represents −log10(p-value); genes with p < 0.05 are highlighted. Statistical significance in (b) was determined by the Chi-square test. RTK, receptor tyrosine kinases; PAM, PI3K-AKT-mTOR
When dividing resistance mechanisms by TKI generation (Fig. 2b), EGFR-dependent events were more common after first- versus third-generation EGFR-TKI treatment (55.69% vs. 33.73%; p = 0.001). Specifically, in first-generation-TKI resistance (n = 325), T790M was 44.62%, amplification 18.77%, C797S 1.23%, and L718Q 0.62%. In second-generation-TKI resistance (n = 27), T790M was 29.63% and amplification 25.93%. In first-line third-generation-TKI resistance (n = 83), the predominant resistance alterations were EGFR amplification (33.73%), L718Q (3.61%), and C797S (2.41%). In contrast, RTK pathway alterations were enriched after second- and third-generation agents compared with first-generation therapy (22.22% and 21.69% vs. 8.62%; p = 0.001). PAM pathway alterations were also observed more frequently after third- than first-generation EGFR-TKI treatment (27.71% vs. 14.15%; p = 0.014). RAS/RAF pathway events were numerically higher after second- versus first- and third-generation therapy (7.41% vs. 3.08% vs. 4.82%), but the difference was not statistically significant (p > 0.05).
To explore generation-specific resistance mechanism, 287 KR samples with defined resistance alterations were analyzed and stratified into first/second-generation EGFR-TKI resistance (n = 230; first: n = 213; second: n = 17) versus third-generation resistance (n = 57). Differential gene analysis revealed that all significantly enriched genes were observed in the third-generation resistance group (p < 0.05, Fig. 2c). These genes were primarily involved in bypass activation (e.g., MET, HGF), cell-cycle regulation (CDKN2A, CDKN2B, CDK6, CDK12), DNA damage-repair pathways (FANCF, FANCG), and epigenetic regulation (SMARCA4).
Given that later-line targeted therapy predominantly involves third-generation EGFR-TKIs, analysis was then focused on patients resistant to third-generation EGFR-TKIs. This cohort was stratified by treatment line (first-line, n = 83; later-line, n = 310). The most frequently altered genes in this set were EGFR (100%), TP53 (64.38%), MET (18.32%), RB1 (15.52%), CDKN2A (13.49%), and CDKN2B (12.98%) (Fig. 3a). Furthermore, evaluating the initial TKI for patients with ≥3rd-line resistance revealed that 96.77% initiated targeted therapy with a first-generation EGFR-TKI, highlighting relative homogeneity in their initial-generation exposure (Fig. S3). In the third-generation-TKI-resistant cohort, C797S reached 12.72% (50/393) and EGFR amplification reached 27.99% (110/393). There were no significant differences in the distribution of classic resistance pathways between the first-line and later-line groups (all p > 0.05, Fig. 3b). To further explore differences in defined resistance mechanisms, we analyzed the 282 KR patients treated with third-generation EGFR-TKIs, stratifying them into a first-line (n = 57) and later-line (n = 225) resistance subgroup. Differential gene analysis identified significant enrichment only in the first-line group (p < 0.05), including genes involved in bypass activation (e.g., ARAF, CTNNB1, RICTOR) and DNA damage repair (e.g., BLM, FANCF, FANCG, PALB2, ERCC5) (Fig. 3c). No genes were significantly enriched in the later-line resistance group, potentially due to the increased tumor heterogeneity in this population.
Fig. 3.
Molecular profile of third-generation EGFR-TKI resistance stratified by line of therapy. (a) Oncoplot visualizing the mutation profile of the overall third-generation EGFR-TKI-resistant cohort. Dedicated rows display the resistance-associated EGFR alterations (T790M, C797S, L718Q, and EGFR amplification). (b) Stacked bar chart comparing the distribution of major resistance mechanism categories between first-line and later-line third-generation EGFR-TKI resistance subgroups. (c) Volcano plot showing differentially enriched genes between first-line and later-line resistance within the third-generation KR subgroup. Statistical significance in (b) was determined by the Chi-square test. In (c), the x-axis represents log2(OR) and the y-axis represents −log10(p-value); genes with p < 0.05 are highlighted. RTK, receptor tyrosine kinases; PAM, PI3K-AKT-mTOR
Molecular landscape of EGFR-dependent versus non-EGFR-dependent resistance
To explore the molecular landscape of EGFR-dependent resistance mechanism, patients were classified as EGFR-dependent (n = 345) and non-EGFR-dependent (n = 129) subgroups. The remaining 38 patients within the KR cohort were excluded from this specific analysis due to a functionally unclassified mechanism (EGFR T790M detected after third-generation TKI resistance).
In the EGFR-dependent subgroup, EGFR alterations (100%) and TP53 co-alterations (63.77%) were the most frequent events, with acquired T790M and EGFR amplification predominating (60.87% and 51.59%, respectively) (Fig. 4a). Pathway analysis confirmed enrichment of canonical pathways (e.g., ErbB signaling pathway, PI3K-AKT/mTOR, and RAS-MAPK), apoptosis, and cell-cycle programs (Fig. 4b). Notably, EGFR dependent subgroup showed unique enrichment signals for mitophagy, NOD-like receptor, and Hippo pathways, suggesting metabolic/oxidative stress and impaired growth-suppressive signaling during the EGFR-TKI resistance.
Fig. 4.
Molecular landscape of the EGFR-dependent resistance subgroup. (a) Oncoplot showing the mutation profile of the EGFR-dependent resistance subgroup. Dedicated rows display the resistance-associated EGFR alterations (T790M, C797S, L718Q, and EGFR amplification). (b) KEGG pathway enrichment analysis for the EGFR-dependent subgroup
In the non-EGFR-dependent subgroup, beyond the primary EGFR mutations (100%), the most frequent co-alterations were TP53 (72.09%), MET (31.78%) and PIK3CA (31.78%) (Fig. 5a). KEGG enrichment similarly highlighted canonical resistance pathways (PI3K-AKT/mTOR, RAS-MAPK, and ErbB signaling pathway) together with apoptosis and cell-cycle pathways. Distinctively, enrichment of lysine degradation and primary immunodeficiency pathways suggested additional metabolic and immune-dysregulation features (Fig. 5b).
Figure 5.
Molecular landscape of the non-EGFR-dependent resistance subgroup. (a) Oncoplot showing the mutation profile of the non-EGFR-dependent resistance subgroup. Dedicated rows display the resistance-associated EGFR alterations (T790M, C797S, L718Q, and EGFR amplification). (b) KEGG pathway enrichment analysis for the non-EGFR-dependent subgroup
Co-mutation and mutual exclusivity analyses revealed distinct patterns between EGFR-dependent and non-EGFR-dependent resistance (Fig. 6). Resistance mutations were largely concentrated within the EGFR protein tyrosine kinase domain (Fig. 6a). The EGFR-dependent group exhibited a higher number of significant co-occurrence gene pairs, including TP53 with MYC and RB1, and CDKN2B with SDHA (p < 0.01, Fig. 6b). In the non-EGFR-dependent group, showed fewer significant co-occurrence events (Fig. 6c). Mutual exclusivity between TP53 and MDM2 was observed in both groups, while the TP53 and IFNG pair, and PIK3CA and MET pair were exclusively mutual in the non-EGFR-dependent group ((all p < 0.01)
Fig. 6.
Molecular landscape of EGFR-dependent versus non-EGFR-dependent resistance. (a) Lollipop plot showing the distribution of EGFR mutations in EGFR-dependent group; (b) Co-mutation and mutual exclusivity analysis plot for frequently altered genes in EGFR-dependent resistance group; (c) Co-mutation and mutual exclusivity analysis plot for frequently altered genes in non-EGFR-dependent resistance group; (d) Volcano plot showing differentially enriched genes between EGFR-dependent and non-EGFR-dependent resistance group within KR cohort
Finally, differential gene analysis identified distinct mutation difference between EGFR-dependent and non-EGFR-dependent resistance (Fig. 6d). CDKN2A, IKZF1, RBM10, KDM6A, IL7R, FGF3, FGF19, and NTRK1 were significantly enriched in the EGFR-dependent group (p < 0.05), whereas MET, KRAS, PIK3CA, TSC1, and ERBB2 were preferentially enriched in the non-EGFR-dependent group (p < 0.05).
Genomic characterization of the UR cohort
To further characterize the genomic landscape of patients with unknown EGFR-TKI resistance mechanisms, this study analyzed 233 samples from the UR cohort and 183 EGFR-TKI naïve (baseline) samples with evaluable clinical outcomes. As shown in Fig. 7a, in the UR cohort, no patient harbored T790M, C797S, or EGFR amplification. Four UR patients (1.72%) harbored L718Q. The five most frequently altered genes in both cohorts were EGFR, TP53, LRP1B, RB1, and RBM10. The frequencies of these alterations in the baseline versus UR cohorts were: EGFR (100.00% vs 100.00%), TP53 (64.48% vs 52.36%), LRP1B (11.48% vs 7.73%), RB1 (9.84% vs 9.01%), and RBM10 (9.84% vs 7.30%).
Fig. 7.
Genomic landscape of the unknown resistance cohort. (a) Oncoplot showing the mutation profile of the unknown EGFR-TKI resistance cohort and baseline cohort. (b) KEGG pathway enrichment analysis showing pathways uniquely enriched in UR cohort compared to the baseline cohort. (c) Volcano plot showing differentially enriched genes between the UR and baseline cohorts. UR: unknown resistance
To explore the biological mechanisms of NSCLC with unknown EGFR-TKI resistance mechanism, KEGG pathway enrichment profiles were compared between the baseline and UR cohorts. As shown in Fig. 7b, the overall enrichment patterns were highly concordant. Notably, the UR cohort displayed unique enrichment in several pathways, including ATP-dependent chromatin remodeling, IL-17 signaling, ubiquitin-mediated proteolysis, and base excision repair pathways. These results indicate that tumors with unknown resistance mechanisms retain pretreatment pathway abnormalities but may also acquire unique and non-canonical resistance features through the activation of chromatin remodeling, inflammatory, and proteolysis pathways. To assess whether the observed pathway enrichments were influenced by specimen type, KEGG enrichment analysis was repeated separately in tissue-derived (n = 60) and plasma-derived (n = 131) UR samples. The four pathways remained enriched in the UR cohort across both subgroups (Fig. S4 and Table S2).
Differential gene analysis between the UR and baseline cohorts exhibited distinct signatures (Fig. 7c). MYC, BRCA1, EPHB, and PTPRD were significantly represented in the UR subgroup. In contrast, FOXA1, PTEN, ROS1, MET, PIK3CA, NKX2-1, TP53, POM121L12, STAT3, MTOR, NFKBIA, and IKZF1 were enriched in the EGFR-TKI naïve cohort.
Baseline genomic biomarkers of EGFR-TKI treatment benefit
To assess whether baseline alterations correlate with clinical benefit from EGFR-TKIs, the baseline cohort (n = 183) was stratified by TTF into a long-TTF subgroup (TTF ≥ 12 months; n = 58) and a short-TTF subgroup (TTF ≥ 6 months; n = 44). As expected, the two groups showed a highly significant difference (p < 0.001; Fig. 8a).
Fig. 8.
Baseline genomic predictors of EGFR-TKI treatment benefit. (a) Kaplan-Meier curve comparing the differences in the subgroups of long-TTF and short-TTF in baseline cohort. Statistical significance was determined by the log-rank test. (b) KEGG pathway enrichment analysis showing the unique pathways enriched in the short-TTF subgroup; (c) KEGG pathway enrichment analysis showing the unique pathways enriched in the long-TTF subgroup; (d) Volcano plot showing differentially enriched gene between the short-TTF and long-TTF survival groups; (e) Kaplan-Meier curve comparing TTF for patients in the baseline cohort based on the presence or absence of the baseline IKZF1 amplification
Separate KEGG enrichment analyses within each subgroup revealed distinct pathway signatures. The short-TTF subgroup was enriched for pathways involved in ATP-dependent chromatin remodeling, basal transcription factors, nucleotide excision repair, NF-κB signaling, and Hedgehog signaling (Fig. 8b). In contrast, the long-TTF subgroup showed enrichment for leukocyte transendothelial migration, mismatch repair, ABC transporters, and the phosphatidylinositol signaling system (Fig. 8c). Differential gene analysis between these survival subgroups identified IKZF1 alterations, all of which are amplifications, as significantly enriched in the short-TTF subgroup (p < 0.05; Fig. 8d). Consistently, in the overall baseline cohort, the presence of a baseline IKZF1 amplification was associated with a significantly shorter TTF (HR = 3.02; 95% CI, 1.32–6.90; p = 0.004; Fig. 8e).
To assess the reproducibility of IKZF1 amplification as a recurrent event in EGFR-mutant LUAD, three public cohorts were queried via cBioPortal. Using the high-level amplification as the matched threshold, IKZF1 amplification frequency in EGFR-mutant patients was 10.4% (7/67) in TCGA LUAD, 2.6% (1/38) in CPTAC LUAD, and 5.5% (8/145) in OncoSG LUAD, compared with 3.3% (6/183) in our baseline cohort (Table S3).
Discussion
Acquired resistance to EGFR-TKIs remains a major challenge in the management of advanced EGFR-mutant NSCLC. While known resistance mechanisms (KR) (e.g., EGFR T790M and MET amplification) have been relatively well studied, approximately 30–50% of clinical resistance events still lack an identifiable mechanism (unknown resistance, UR), highlighting the need for deeper investigation. Moreover, a subset of patients exhibits primary resistance in the beginning of TKI treatment, highlighting the urgent need for baseline biomarkers to predict this poor outcome. Based on a large real-world cohort, our retrospective analysis provided a comprehensive view of the resistance landscape to EGFR-TKIs across KR and UR cohorts. Notably, we reported that the UR subset exhibited a distinctive, atypical resistance pattern characterized by epigenetic features. Moreover, baseline IKZF1 amplification merged as a candidate, hypothesis-generating potential primary resistance biomarker which was also associated with epigenetic dysregulation.
The distribution of EGFR-TKI resistance mechanisms is strongly influenced by TKI treatment history. Consistent with previous reports on EGFR-dependent resistance proportions [6, 29], EGFR-dependent mechanisms accounted for a higher fraction after resistance to first-generation EGFR-TKIs (55.69%). In contrast, after resistance to first-line third-generation EGFR-TKIs, the proportion of EGFR-dependent resistance decreased markedly (33.73%). In addition, our study observed a corresponding increase in the activation of RTK and PAM pathways.
A further comparison of resistance after first-line versus later-line treatment with third-generation EGFR-TKIs revealed distinct patterns. First-line resistance samples exhibited more evident pathway enrichment (e.g., bypass activation and DNA damage repair pathways). In contrast, patients with later-line resistance showed no significantly enriched differential genes, suggesting that tumors exposed to multiple lines of therapy display greater heterogeneity during evolution. In this context, NGS testing becomes more important for biomarker-driven decision-making in later-line treatment [6].
We further analyzed the biological mechanisms of EGFR-dependent and non-EGFR-dependent subgroups. The non-EGFR-dependent subgroup was significantly enriched for key genes such as MET, KRAS, PIK3CA, and ERBB2, suggesting that combination therapy with EGFR-TKIs and corresponding targeted inhibitors may be a promising future direction for these patients. For EGFR-dependent patients, treatment should prioritize selecting EGFR-TKIs based on the specific resistance mutation, or the targeted development of next-generation TKIs [6].
In this study, approximately one-third of patients (31.28%) had unknown resistance mechanisms, a proportion consistent with prior report [6, 29, 30]. The proportion of unknown resistance can also be influenced by factors such as the line of therapy. Although the UR cohort did not differ significantly from the EGFR-naïve baseline cohort in clinical characteristics (age, sex, and smoking history), it displayed a distinct molecular landscape.
Pathway enrichment analysis revealed that the UR cohort was distinctively enriched in atypical pathways, including IL-17 signaling and ATP-dependent chromatin remodeling. The enrichment of the IL-17 pathway is consistent with previous study [31], which demonstrated that IL-17 promotes TKI resistance by enhancing the phosphorylation of EGFR and MET. More importantly, research linking ATP-dependent chromatin remodeling to EGFR-TKI resistance is limited. While prior work has reported that SWI/SNF chromatin remodeling complexes promote tyrosine kinase inhibitor resistance in EGFR-mutant lung cancer [19], the specific role of ATP-dependent chromatin remodeling in EGFR-TKI resistance has remained unclear. ATP-dependent chromatin remodeling plays a key role in oncogenic genomic reprogramming during tumor development and anti-cancer treatment, and this pathway may modulate the tumor microenvironment by epigenetically activating pro-inflammatory gene expression [32]. Together, these observations suggest that epigenetic dysregulation may be associated with the development of unknown resistance, although direct functional validation is required before causal conclusions can be drawn.
Differential gene analysis further supported the above findings. Enrichment of the transcription factor MYC has been reported to be associated with EGFR-TKI resistance and shortened PFS in prior studies [33, 34]. Furthermore, the combination of anlotinib with osimertinib has been shown to reverse osimertinib resistance by inhibiting the c-MET/MYC/AXL axis [35]. In addition, PTPRD alterations were found to be enriched in the UR cohort. Together with MYC amplification, PTPRD alterations have been considered as ALK-independent resistance-related genetic alterations in Chinese lung adenocarcinoma patients [36]. These findings suggest that PTPRD alterations may represent an overlooked novel target in EGFR-TKI resistance that warrants further exploration.
To explore primary resistance, baseline patients were classified into short-benefit (TTF ≤ 6 months) and long-benefit (TTF ≥ 12 months) cohorts, based on the definition of acquired resistance and previous studies [14, 24, 25]. IKZF1 amplification was found to be significantly enriched in the short-benefit group and was significantly associated with worse TTF (HR = 3.02). IKZF1 is associated with chromatin remodeling and encodes the lymphoid transcription factor IKAROS. IKZF1 alterations are linked to poor outcomes in B-progenitor acute lymphoblastic leukemia (ALL) [37]. In NSCLC, IKZF1 has also been reported to be associated with EGFR amplification and poor response to EGFR-TKIs [38, 39]. In our baseline cohort, all IKZF1 alterations consisted of copy number gains spanning exons 2 to 8E on chromosome 7. This pattern is consistent with prior studies reporting that higher Ikaros expression correlates with tumor progression in lung adenocarcinoma [40]. Notably, this finding stands in contrast to the role of IKZF1 loss in driving therapeutic resistance in ALL [41]. This discrepancy may be explained by the context-dependent function of Ikaros, which acts either as a transcriptional repressor or activator depending on its associated chromatin complexes: it can mediate tumor suppression by recruiting the nucleosome remodeling and deacetylase complex, or achieve gene activation by integrating into the SWI/SNF complex [42]. This dual regulatory capacity offers a conceptual basis for how IKZF1 amplification might influence transcriptional reprogramming under EGFR-TKI pressure. However, a causal tumor-cell-intrinsic role cannot be established from correlative genomic data and requires direct functional validation.
Cross-cohort analyses supported the biological reproducibility of IKZF1 amplification in EGFR-mutant LUAD. Comparable low-frequency amplification rates (2.6–10.4%) were observed across TCGA, CPTAC, and OncoSG EGFR-mutant cohorts, consistent with the 3.3% rate in our own baseline cohort.
Interestingly, the ATP-dependent chromatin remodeling pathway was significantly enriched in both the primary resistance group and the UR group. These results suggest that abnormalities in epigenetic mechanisms related to chromatin remodeling may span the entire course of EGFR-TKI treatment. It may function as a primary resistance factor that leads to poor outcomes and as an acquired mechanism that drives atypical resistance.
This study has several limitations. First of all, as a retrospective real-world study, our analysis is subject to potential selection bias, which may limit the generalization of the results. Secondly, matched pre- and post-TKI specimens from the same patients were not systematically available, limiting our ability to definitively trace intra-patient clonal evolution and differentiate pre-existing from strictly acquired alterations. Third, the absence of rebiopsy-confirmed histologic transformation data prevented direct assessment of SCLC or squamous transformation, although our genomic surrogate analysis suggests SCLC transformation does not primarily drive the UR cohort. Fourth, complete line-by-line chemotherapy records were not uniformly available for late-line resistance cases. Fifth, differences in variant detection sensitivity between plasma and tissue biopsies may contribute to variability in variant detectability, although our depth of coverage analyses support the validity of the integrated results. Finally, our findings linking epigenetic dysregulation and IKZF1 amplification to EGFR-TKI resistance are based on pathway enrichment and correlative analyses, and thus required validation in larger prospective cohorts and dedicated functional experiments, including in vitro IKZF1 knockdown/overexpression in EGFR-mutant NSCLC cell lines and single-cell RNA-sequencing of EGFR-mutant tumors.
Conclusion
In summary, based on a large real-world cohort, this study not only provides a comprehensive landscape of EGFR-TKI resistance but also reported novel insights into unknown resistance and primary resistance mechanisms. Our results suggest that epigenetic dysregulation may be a potential contributor of both unknown and primary resistance to EGFR-TKIs. IKZF1 amplification emerged as a candidate, hypothesis-generating biomarker of primary resistance that requires validation in larger prospective cohorts and dedicated functional studies, which together may inform new therapeutic strategies for the approximately 30% of patients with unknown resistance.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Author contributions
Songchen Zhao, Shuchun Wang and Caicun Zhou designed the study. Shuchun Wang and Songchen Zhao organized the clinical data. Shuchun Wang and Wenshuai Chen performed the bioinformatics and statistical analyses and prepared the figures and tables. Songchen Zhao and Shuchun Wang drafted the manuscript. Songchen Zhao, Fei Zhou , Shuchun Wang and Caicun Zhou contributed to the interpretation of data and critically revised the manuscript. All authors contributed to the discussion of the results and approved the final version of the manuscript.
Funding
This study was supported by National Natural Science Foundation of China (82530089 and 82521002), New Quality Clinical Specialty Program of High-end Medical Disciplinary Construction in Shanghai Pudong New Area (2025TS1404201), the Healthcare Talents Youth Program of Shanghai Pudong New Area (2025-PWXZ-02).
Data availability
The data supporting this study’s findings are available from the corresponding author upon reasonable request.
Declarations
Ethical approval
Samples were collected from patients after they had provided written informed consent. All methods were carried out in accordance with relevant guidelines and regulations. All procedures were conducted in accordance with the Declaration of Helsinki. All experimental protocols were approved by the ethics committee of Shanghai Pulmonary Hospital (K20-288).
Competing interests
Wenshuai chen and Shuchun Wang are employees of Geneplus-Beijing. All other authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Shuchun Wang, Email: wangshch@geneplus.org.cn.
Caicun Zhou, Email: caicunzhoudr@163.com.
References
- 1.C.Li, S.Lei, L.Ding, Y.Xu, X.Wu, H.Wang, Z.Zhang, T.Gao, Y.Zhang, L.Li, Global burden and trends of lung cancer incidence and mortality. Chin. Med. J. (Engl). 136(13), 1583–1590 (2023). 10.1097/CM9.0000000000002529 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.R.S.Herbst, D.Morgensztern, C.Boshoff, The biology and management of non-small cell lung cancer. Nature. 553(7689), 446–454 (2018). 10.1038/nature25183 [DOI] [PubMed] [Google Scholar]
- 3.J.H.Schiller, D.Harrington, C.P.Belani, C.Langer, A.Sandler, J.Krook, J.Zhu, D.H.Johnson, Comparison of four chemotherapy regimens for advanced non–small-Cell lung cancer. N. Engl. J. Med. 346(2), 92–98 (2002). 10.1056/NEJMoa011954 [DOI] [PubMed] [Google Scholar]
- 4.J.W, Z.L, Non-small cell lung cancer targeted therapy: drugs and mechanisms of drug resistance. Int. J. Mol. Sci. [Internet]. 23. [cited 2025 Oct 31]. Available from: 10.3390/ijms232315056. https://pubmed.ncbi.nlm.nih.gov/36499382/ [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.T.Mitsudomi, K.Suda, Y.Yatabe, Surgery for NSCLC in the era of personalized medicine. Nat. Rev. Clin. Oncol. 10(4), 235–244 (2013). 10.1038/nrclinonc.2013.22 [DOI] [PubMed] [Google Scholar]
- 6.A.Passaro, P.A.Jänne, T.Mok, S.Peters, Overcoming therapy resistance in EGFR-mutant lung cancer. Nat. Cancer. 2(4), 377–391 (2021). 10.1038/s43018-021-00195-8 [DOI] [PubMed] [Google Scholar]
- 7.X.Nan, C.Xie, X.Yu, J.Liu, EGFR TKI as first-line treatment for patients with advanced EGFR mutation-positive non-small-cell lung cancer. Oncotarget. 8(43), 75712–75726 (2017). 10.18632/oncotarget.20095 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.W.-H.Hsu, J.C.-H.Yang, T.S.Mok, H.H.Loong, Overview of current systemic management of EGFR-mutant NSCLC. Ann. Oncol. Elsevier. 29, i3–9 (2018). 10.1093/annonc/mdx702 [DOI] [PubMed]
- 9.J.-C.Soria, Y.Ohe, J.Vansteenkiste, T.Reungwetwattana, B.Chewaskulyong, K.H.Lee, A.Dechaphunkul, F.Imamura, N.Nogami, T.Kurata, I.Okamoto, C.Zhou, B.C.Cho, Y.Cheng, E.K.Cho, P.J.Voon, D.Planchard, W.-C.Su, J.E.Gray, S.-M.Lee, R.Hodge, M.Marotti, Y.Rukazenkov, S.S.Ramalingam, Osimertinib in untreated EGFR -Mutated advanced non–small-Cell lung cancer. N. Engl. J. Med. 378(2), 113–125 (2018). 10.1056/NEJMoa1713137 [DOI] [PubMed] [Google Scholar]
- 10.L.E.L.Hendriks, J.Remon, C.Faivre-Finn, M.C.Garassino, J.V.Heymach, K.M.Kerr, D.S.W.Tan, G.Veronesi, M.Reck, Non-small-cell lung cancer. Nat. Rev. Dis. Primers. 10(1), 71 (2024). 10.1038/s41572-024-00551-9 [DOI] [PubMed] [Google Scholar]
- 11.P.S.Hammerman, P.A.Jänne, B.E.Johnson, Resistance to epidermal growth factor receptor tyrosine kinase inhibitors in non–small cell lung cancer. Clin. Cancer Res. 15(24), 7502–7509 (2009). 10.1158/1078-0432.CCR-09-0189 [DOI] [PubMed] [Google Scholar]
- 12.K.Fu, F.Xie, F.Wang, L.Fu, Therapeutic strategies for EGFR-mutated non-small cell lung cancer patients with osimertinib resistance. J. Hematol. Oncol. 15(1), 173 (2022). 10.1186/s13045-022-01391-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.K.Kobayashi, Primary resistance to EGFR tyrosine kinase inhibitors (TKIs): contexts and comparisons in EGFR-Mutated lung cancer. JoR. 3(4), 223–236 (2023). 10.3390/jor3040021 [DOI] [Google Scholar]
- 14.Y.Xu, X.Tong, J.Yan, X.Wu, Y.W.Shao, Y.Fan, Short-term responders of non–small cell lung cancer patients to EGFR tyrosine kinase inhibitors display high prevalence of TP53 mutations and primary resistance mechanisms. Transl. Oncol. 11(6), 1364–1369 (2018). 10.1016/j.tranon.2018.08.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.J.A.Engelman, T.Mukohara, K.Zejnullahu, E.Lifshits, A.M.Borrás, C.-M.Gale, G.N.Naumov, B.Y.Yeap, E.Jarrell, J.Sun, S.Tracy, X.Zhao, J.V.Heymach, B.E.Johnson, L.C.Cantley, P.A.Janne, Allelic dilution obscures detection of a biologically significant resistance mutation in EGFR-amplified lung cancer. J. Clin. Investigation. 116(10), 2695–2706 (2006). 10.1172/JCI28656 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.C.Yamamoto, Y.Basaki, A.Kawahara, K.Nakashima, M.Kage, H.Izumi, K.Kohno, H.Uramoto, K.Yasumoto, M.Kuwano, M.Ono, Loss of PTEN expression by blocking nuclear translocation of EGR1 in gefitinib-resistant lung cancer Cells harboring epidermal growth factor receptor–activating mutations. Cancer Res. 70(21), 8715–8725 (2010). 10.1158/0008-5472.CAN-10-0043 [DOI] [PubMed] [Google Scholar]
- 17.K.Imran, M.J.Iqbal, M.M.Ahmed, A.Khalid, H.Cortés, O.D.Reyes-Hernández, G.Figueroa-González, G.Leyva-Gómez, L.Falzone, M.Libra, F.Longo, J.Sharifi-Rad, D.Calina, Epigenetic dysregulation in cancer: mechanisms, diagnostic biomarkers and therapeutic strategies. Med. Oncol. 42(8), 359 (2025). 10.1007/s12032-025-02905-z [DOI] [PubMed] [Google Scholar]
- 18.S.Eustermann, A.B.Patel, K.-P.Hopfner, Y.He, P.Korber, Energy-driven genome regulation by ATP-dependent chromatin remodellers. Nat. Rev. Mol. Cell. Biol. 25(4), 309–332 (2024). 10.1038/s41580-023-00683-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.F.J.de Miguel, C.Gentile, W.W.Feng, S.J.Silva, A.Sankar, F.Exposito, W.L.Cai, M.A.Melnick, C.Robles-Oteiza, M.M.Hinkley, J.A.Tsai, A.-V.Hartley, J.Wei, A.Wurtz, F.Li, M.I.Toki, D.L.Rimm, R.Homer, C.B.Wilen, A.Z.Xiao, J.Qi, Q.Yan, D.X.Nguyen, P.A.Jänne, C.Kadoch, K.A.Politi, Mammalian SWI/SNF chromatin remodeling complexes promote tyrosine kinase inhibitor resistance in EGFR-mutant lung cancer. Cancer Cell. 41(8), 1516–1534.e9 (2023). 10.1016/j.ccell.2023.07.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.F.Hao, Y.Zhang, J.Hou, B.Zhao, Chromatin remodeling and cancer: the critical influence of the SWI/SNF complex. Epigenet. Chromatin. 18(1), 22 (2025). 10.1186/s13072-025-00590-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.H.Zhang, Y.Wang, Q.Ji, H.Cai, X.Liang, J.Xie, H.Li, J.Wang, G.Zhu, E.Tian, L.Zhu, M.Yuan, R.Chen, M.Zhao, Clinicopathological and molecular characteristics of patients with hypermutant lung cancer: a retrospective cohort study. Oncol. Lett. 21(4), 329 (2021). 10.3892/ol.2021.12590 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Y.Zhang, L.Chang, Y.Yang, W.Fang, Y.Guan, A.Wu, S.Hong, H.Zhou, G.Chen, X.Chen, S.Zhao, Q.Zheng, H.Pan, L.Zhang, H.Long, H.Yang, X.Wang, Z.Wen, J.Wang, H.Yang, X.Xia, Y.Zhao, X.Hou, Y.Ma, T.Zhou, Z.Zhang, J.Zhan, Y.Huang, H.Zhao, N.Zhou, X.Yi, L.Zhang, The correlations of tumor mutational burden among single-region tissue, multi-region tissues and blood in non-small cell lung cancer. J. Immunother. Cancer. 7(1), 98 (2019). 10.1186/s40425-019-0581-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.M.Offin, J.M.Chan, M.Tenet, H.A.Rizvi, R.Shen, G.J.Riely, N.Rekhtman, Y.Daneshbod, A.Quintanal-Villalonga, A.Penson, M.D.Hellmann, M.E.Arcila, M.Ladanyi, D.Pe’er, M.G.Kris, C.M.Rudin, H.A.Yu, Concurrent RB1 and TP53 alterations define a subset of EGFR-mutant lung cancers at risk for histologic transformation and inferior clinical outcomes. J. Thorac. Oncol. 14(10), 1784–1793 (2019). 10.1016/j.jtho.2019.06.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.F.Morgillo, C.M.Della Corte, M.Fasano, F.Ciardiello, Mechanisms of resistance to EGFR-targeted drugs: lung cancer. ESMO Open. 1(3), e000060 (2016). 10.1136/esmoopen-2016-000060 [DOI] [PMC free article] [PubMed]
- 25.Z.Song, Y.Zhang, Treatment and prognosis after progression in long-term responders to EGFR-tyrosine kinase inhibitor in advanced non-small cell lung cancer. Aoms. 1, 107–111 (2016). 10.5114/aoms.2016.57586 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.K.A. Hoadley, C. Yau, T. Hinoue, D.M. Wolf, A.J. Lazar, E. Drill, et al., Cell-of-origin patterns dominate the molecular classification of 10, 000 tumors from 33 types of cancer. Cell. 173(2), 291–304.e6 (2018). 10.1016/j.cell.2018.03.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.M.A.Gillette, S.Satpathy, S.Cao, S.M.Dhanasekaran, S.V.Vasaikar, K.Krug, F.Petralia, Y.Li, W.-W.Liang, B.Reva, A.Krek, J.Ji, X.Song, W.Liu, R.Hong, L.Yao, L.Blumenberg, S.R.Savage, M.C.Wendl, B.Wen, K.Li, L.C.Tang, M.A.MacMullan, S.C.Avanessian, M.H.Kane, C.J.Newton, M.Cornwell, R.B.Kothadia, W.Ma, S.Yoo, R.Mannan, P.Vats, C.Kumar-Sinha, E.A.Kawaler, T.Omelchenko, A.Colaprico, Y.Geffen, Y.E.Maruvka, F.da Veiga Leprevost, M.Wiznerowiczet al., Proteogenomic characterization reveals therapeutic vulnerabilities in lung adenocarcinoma. Cell. 182(1), 200–225.e35 (2020). 10.1016/j.cell.2020.06.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.J.Chen, H.Yang, A.S.M.Teo, L.B.Amer, F.G.Sherbaf, C.Q.Tan, J.J.S.Alvarez, B.Lu, J.Q.Lim, A.Takano, R.Nahar, Y.Y.Lee, C.Z.J.Phua, K.P.Chua, L.Suteja, P.J.Chen, M.M.Chang, T.P.T.Koh, B.-H.Ong, D.Anantham, A.A.L.Hsu, A.Gogna, C.W.Too, Z.W.Aung, Y.F.Lee, L.Wang, T.K.H.Lim, A.Wilm, P.S.Choi, P.Y.Ng, C.K.Toh, W.-T.Lim, S.Ma, B.Lim, J.Liu, W.L.Tam, A.J.Skanderup, J.P.S.Yeong, E.-H.Tan, C.L.Creasyet al., Genomic landscape of lung adenocarcinoma in East asians. Nat. Genet. 52(2), 177–186 (2020). 10.1038/s41588-019-0569-6 [DOI] [PubMed] [Google Scholar]
- 29.L.V.Sequist, B.A.Waltman, D.Dias-Santagata, S.Digumarthy, A.B.Turke, P.Fidias, K.Bergethon, A.T.Shaw, S.Gettinger, A.K.Cosper, S.Akhavanfard, R.S.Heist, J.Temel, J.G.Christensen, J.C.Wain, T.J.Lynch, K.Vernovsky, E.J.Mark, M.Lanuti, A.J.Iafrate, M.Mino-Kenudson, J.A.Engelman, Genotypic and histological evolution of lung cancers acquiring resistance to EGFR inhibitors. Sci Transl Med. 3(75), 75ra26 (2011). 10.1126/scitranslmed.3002003 [DOI] [PMC free article] [PubMed]
- 30.X.Zhou, L.Zeng, Z.Huang, Z.Ruan, H.Yan, C.Zou, S.Xu, Y.Zhang, Strategies Beyond 3rd EGFR-TKI acquired resistance: opportunities and Challenges. Cancer Med. 14(9), e70921 (2025). 10.1002/cam4.70921 [DOI] [PMC free article] [PubMed]
- 31.K.-L.Lee, T.-C.Lai, W.-J.Lee, Y.-C.Chen, K.-H.Ho, W.-Y.Hung, Y.-C.Yang, M.-H.Chan, F.-K.Hsieh, C.-L.Chung, J.-H.Chang, M.-H.Chien, Sustaining the activation of EGFR signal by inflammatory cytokine IL17A prompts cell proliferation and EGFR-TKI resistance in lung cancer. Cancers (Basel). 15(13), 3288 (2023). 10.3390/cancers15133288 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.F.-L.Zhang, D.-Q.Li, Targeting chromatin-remodeling factors in cancer Cells: promising molecules in cancer therapy. IJMS. 23(21), 12815 (2022). 10.3390/ijms232112815 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.L.Zhu, Z.Chen, H.Zang, S.Fan, J.Gu, G.Zhang, K.D.-Y.Sun, Q.Wang, Y.He, T.K.Owonikoko, S.S.Ramalingam, S.-Y.Sun, Targeting c-myc to overcome acquired resistance of EGFR mutant NSCLC Cells to the third-generation EGFR tyrosine kinase inhibitor, Osimertinib. Cancer Res. 81(18), 4822–4834 (2021). 10.1158/0008-5472.CAN-21-0556 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.G.Wang, T.Li, Y.Wan, Q.Li, MYC expression and fatty acid oxidation in EGFR-TKI acquired resistance. Drug. Resist. Updat. 72, 101019 (2024). 10.1016/j.drup.2023.101019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.T.Lei, T.Xu, N.Zhang, X.Zou, Z.Kong, C.Wei, Z.Wang, Anlotinib combined with osimertinib reverses acquired osimertinib resistance in NSCLC by targeting the c-MET/MYC/AXL axis. Pharmacol. Res.188, 106668 (2023). 10.1016/j.phrs.2023.106668 [DOI] [PubMed] [Google Scholar]
- 36.G.Hua, X.Zhang, M.Zhang, Q.Wang, X.Chen, R.Yu, H.Bao, J.Liu, X.Wu, Y.Shao, B.Liang, K.Lu, Real-world circulating tumor DNA analysis depicts resistance mechanism and clonal evolution in ALK inhibitor-treated lung adenocarcinoma patients. ESMO Open. 7(1), 100337 (2022). 10.1016/j.esmoop.2021.100337 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.M.L.Churchman, J.Low, C.Qu, E.M.Paietta, L.H.Kasper, Y.Chang, D.Payne-Turner, M.Althoff, G.Song, S.-C.Chen, J.Ma, M.Rusch, D.McGoldrick, M.Edmonson, P.Gupta, Y.-D.Wang, W.Caufield, B.Freeman, L.Li, J.Panetta, S.Baker, Y.-L.Yang, K.Roberts, K.McCastlain, I.Iacobucci, J.Peters, V.Centonze, F.Notta, S.Dobson, S.Zandi, J.Dick, L.Janke, J.Peng, K.Kodali, V.Pagala, J.Min, A.Mayasundari, R.Williams, C.Willman, J.Roweet al., EFFICACY OF RETINOIDS IN IKZF1-MUTATED BCR-ABL1 ACUTE LYMPHOBLASTIC LEUKEMIA. Cancer Cell. 28(3), 343–356 (2015). 10.1016/j.ccell.2015.07.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.H.Bai, Y.Zhou, W.Liu, W.Xu, L.Cheng, Y.Huo, H.Ji, L.Xiong, Genetic mutation profiling reveals biomarkers for targeted therapy efficacy and prognosis in non-small cell lung cancer. Heliyon. 10(6), e27633 (2024). 10.1016/j.heliyon.2024.e27633 [DOI] [PMC free article] [PubMed]
- 39.Y.Xu, J.Yan, C.Zhou, L.Wu, H.Wang, J.Zhao, M.Zhou, J.Wang, X.Zheng, L.Zhang, K.Jiang, X.Zheng, Q.Miao, S.Wu, Z.Zou, R.Lian, Y.He, R.Chen, S.Yang, Y.Li, S.Chen, G.Lin, Genomic characterisation of de novo EGFR copy number gain and its impact on the efficacy of first-line EGFR-tyrosine kinase inhibitors for EGFR mutated non-small cell lung cancer. Eur. J. Cancer. 188, 81–89 (2023). 10.1016/j.ejca.2023.04.009 [DOI] [PubMed] [Google Scholar]
- 40.W.Zhao, T.Chen, H.Wang, Ikaros is heterogeneously expressed in lung adenocarcinoma and is involved in its progression. J. Int. Med. Res. 48(8), 0300060520945860 (2020). 10.1177/0300060520945860 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.R.Marke, F.N.van Leeuwen, B.Scheijen, The many faces of IKZF1 in B-cell precursor acute lymphoblastic leukemia. Haematologica. 103(4), 565–574 (2018). 10.3324/haematol.2017.185603 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.R.Xia, Y.Cheng, X.Han, Y.Wei, X.Wei, Ikaros proteins in tumor: current perspectives and New developments. Front. Mol. Biosci. [Internet]. Frontiers, 2021[cited 2025 Nov 3];8. Available from: https://www.frontiersin.org/journals/molecular-biosciences/articles/10.3389/fmolb.2021.788440/full [DOI] [PMC free article] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data supporting this study’s findings are available from the corresponding author upon reasonable request.