Abstract
Introduction
Transformed small-cell lung cancer (tSCLC) is a clinically important resistance mechanism to EGFR tyrosine kinase inhibitors in EGFR-mutant non–small cell lung cancer. This study characterizes clinical features, treatment outcomes, and biomarker profiles in patients with tSCLC.
Methods
Data from 45 patients with EGFR-mutant NSCLC who developed tSCLC between 2014 and 2023 were analyzed. Demographic characteristics, treatment histories, and delta-like ligand 3 (DLL3) and B7-H3 expression were collected. Objective response rate, progression-free survival (PFS), and posttransformation survival (PTS) were assessed. Spatial transcriptomic profiling was performed in selected cases.
Results
Most patients were women (60%) and never-smokers (75.6%). Exon 19 deletion was the predominant EGFR mutation (57.8%). Median PFS and PTS were 3.3 and 9.2 months, respectively. Etoposide plus platinum (EP) was the predominant first-line regimen (69.8%), with 23.2% of the patients receiving EP plus immune checkpoint or tyrosine kinase inhibitors. EP-based combination regimens yielded a numerically higher objective response rate and a significantly longer PFS than EP alone (7.5 versus 2.8 months, p = 0.002). PTS was longer with EP-based regimens than with other regimens (10.4 versus 6.4 months, p = 0.035). DLL3 and B7-H3 were expressed in 87.5% and 66.7% of tumors, respectively, without prognostic significance. Multivariable analysis identified brain metastasis and liver progression at transformation as adverse prognostic factors. Spatial transcriptomic analysis revealed neuroendocrine lineage reprogramming, stromal depletion, and immune exclusion.
Conclusions
tSCLC remains an aggressive resistance phenotype with poor outcomes. EP-based combination strategies may provide clinical benefit, whereas frequent DLL3 expression supports further evaluation of targeted therapies.
Keywords: Epidermal growth factor receptor, Small cell transformation, Non–small cell lung cancer, Delta-like ligand 3, Immune checkpoint inhibitor
Introduction
Histologic transformation is a well-recognized mechanism of acquired resistance in advanced cancers, particularly NSCLC, which often transforms to neuroendocrine carcinoma, mostly SCLC.1, 2, 3 Approximately 3% to 14% of patients with NSCLC harboring EGFR mutations progress to small cell carcinoma (SCLC) after developing resistance to EGFR-tyrosine kinase inhibitors (TKIs).4, 5, 6 Patients with EGFR-mutant NSCLC who develop transformed SCLC (tSCLC) tend to have a poor prognosis, with a median survival duration ranging from 10 to 12.4 months.6, 7, 8 Neuroendocrine transformation typically occurs at around 17.8 months after the initial NSCLC diagnosis and is often associated with mutations in RB1, TP53, and PIK3CA.6 Emerging evidence suggests that neuroendocrine transformation is primarily driven by transcriptional reprogramming rather than new mutational events.9,10
Etoposide plus platinum (EP) chemotherapy remains the standard treatment for patients with tSCLC, leading to response rates similar to those observed in patients with de novo SCLC. In a study conducted in North America,6 immune checkpoint inhibitors (ICIs) exhibited limited efficacy when used alone in this setting. However, in recent phase 3 randomized trials, ICIs such as atezolizumab and durvalumab markedly improved overall survival (OS) when administered alongside EP chemotherapy as the first-line treatment for de novo extensive-stage SCLC.11,12 Despite these findings, the efficacy of chemoimmunotherapy in treating tSCLC remains unclear because of the scarcity of relevant studies. Although a study highlighted the efficacy of PD-L1 inhibitor–based combination therapies in managing tSCLC, the sample size was small.13 Therefore, real-world data on the effectiveness of various treatment regimens after small cell transformation remain limited, particularly in Asian populations, necessitating further investigation.
Several emerging therapies have been evaluated for treating relapsed de novo SCLC. Lurbinectedin, an oncogenic transcription inhibitor, has reported efficacy as a second-line treatment for SCLC.14 Therapeutic agents targeting delta-like ligand 3 (DLL3) have also proven effective, with tarlatamab—a bispecific T-cell engager directed against DLL3—achieving an objective response rate (ORR) of 40% in patients with relapsed or refractory SCLC.15 Furthermore, antibody–drug conjugates, such as those targeting B7-H3, have yielded promising results, with early trials reporting ORRs of up to 52%.16 Despite these findings, the expression patterns and predictive value of biomarkers associated with these novel therapies for tSCLC remain poorly characterized. Therefore, further research is required to clarify the biomarker landscape of tSCLC and its potential role in guiding personalized treatment.
In this study, we evaluated the clinical efficacy of EP alone and EP plus ICIs or EGFR-TKIs in Asian patients with EGFR-mutant tSCLC. In addition, we measured the expression levels of DLL3 and B7-H3 to determine their prognostic value in EGFR-mutant tSCLC. Furthermore, we performed a spatial transcriptomic analysis using paired tumor specimens collected before and after histological transformation to tSCLC to identify gene expression patterns and changes in biomarker expression.
Methods
Study Design and Cohort
Data on patients with EGFR-mutant NSCLC who developed tSCLC between July 2014 and November 2023 were retrospectively reviewed. Patients with tSCLC have been prospectively and consecutively enrolled since January 2022. Clinical data, such as demographic characteristics, molecular profile, cancer stage at diagnosis, metastasis site, systemic treatments received before and after histologic transformation, and survival status, were extracted from the patients’ electronic medical records. Patients were considered eligible for inclusion if they had an initial diagnosis of advanced EGFR-mutant NSCLC, had received at least one line of EGFR-TKI therapy, and had undergone a repeat biopsy at the time of acquired resistance. Transformation to small-cell histology was confirmed when an SCLC component was identified in the repeat biopsy specimen, as per previously established criteria.1 Treatment responses were assessed per the Response Evaluation Criteria in Solid Tumors version 1.1. Disease status and treatment outcomes were documented, with the last follow-up conducted on June 30, 2025. This study adhered to the ethical principles outlined in the Declaration of Helsinki. All enrolled patients received a full explanation of the study and provided written informed consent to participate in the prospective cohort. The study protocol was approved by the Institutional Review Board of Taipei Veterans General Hospital, Taipei, Taiwan (2022-08-017B, 2023-06-016CC, and 2024-02-012A).
Immunohistochemistry for SCLC Subtyping and DLL3/B7-H3 Expression Evaluation
Formalin-fixed, paraffin-embedded tumor samples were subjected to immunohistochemical (IHC) staining to identify SCLC markers, such as ASCL1, NEUROD1, POU2F3, and YAP1.17, 18, 19 Established protocols were used to evaluate the expression of DLL3 and B7-H3, two emerging therapeutic biomarkers in neuroendocrine tumors. The details of the primary antibodies and staining conditions used are outlined in Supplementary Table 1. Marker expression was quantified using the H-score method, in which the percentage of positive tumor cells is combined with the staining intensity to generate a score ranging from 0 to 300.17,18 SCLC molecular subtypes were determined based on the relative expression levels of ASCL1, NEUROD1, POU2F3, and YAP1. Classification criteria adapted from the study by Park et al. were used to assign tumors to ASCL1-dominant, NEUROD1-dominant, or ASCL1/NEUROD1-codominant (AN) categories. Tumors negative for all four markers were classified as triple-negative.20
Spatial Transcriptomic Analysis
Spatial transcriptomic profiling was performed using the 10× Genomics Visium platform (10× Genomics, Pleasanton, CA). For this, we used four paired tumor specimens from two patients (P1 and P2) with EGFR-mutant NSCLC with subsequent small cell transformation. For each patient, a tumor specimen was collected before transformation, and another specimen was collected after histological confirmation of SCLC. Formalin-fixed, paraffin-embedded samples (5 μm) were mounted on Visium Spatial Gene Expression slides (10× Genomics). Tissue sections were then stained with hematoxylin–eosin and imaged to guide area selection for RNA capture. Libraries were prepared using the Visium CytAssist Spatial Gene Expression Kit (10× Genomics). Sequencing was performed on a NovaSeq platform (Illumina, San Diego, CA) following the manufacturer’s protocol. The Space Ranger pipeline (v.3.1.3; 10× Genomics) was used for human reference genome (GRCh38) alignment, feature counting, and spatial barcode assignment. The resulting data were visualized and analyzed using Loupe Browser (v.8.1.2; 10× Genomics).
Tumor-enriched regions were annotated by a pathologist (YCY) and selected for downstream analysis. Unsupervised clustering and dimensionality reduction were performed using Uniform Manifold Approximation and Projection to explore spatial transcriptional heterogeneity. A differential gene expression (DEG) analysis was performed to compare pretransformation NSCLC regions with posttransformation tSCLC regions. Genes that were commonly upregulated or downregulated in both P1 and P2 were identified using the following thresholds: log2 fold change greater than or equal to 1 and p less than 0.05. Next, a gene ontology (GO) enrichment analysis was performed to identify the biological processes associated with transcriptional reprogramming during histologic transformation. Finally, a broader union DEG set, comprising genes that met the significance criteria in either P1 or P2, was generated to capture potentially heterogeneous or patient-specific alterations.
Statistical Analyses
Standard descriptive statistics were used to describe the baseline clinical characteristics. For between-group comparisons, continuous variables were analyzed using either an independent t-test or the Mann–Whitney U test, depending on the data distribution. Categorical variables were compared using the chi-square test or Fisher’s exact test, as appropriate. Kaplan–Meier survival analysis was performed to estimate OS and progression-free survival (PFS) after transformation, with between-group differences assessed using the log-rank test. Univariable and multivariable Cox proportional-hazards models were used to identify factors associated with survival outcomes. Variables with a p value less than 0.1 in the univariable analysis were included in the multivariable model. A two-sided p value of less than 0.05 was considered statistically significant. All statistical analyses were performed using SPSS (v.26.0; IBM, Armonk, NY). A GO enrichment analysis for upregulated and downregulated genes was performed using the “clusterProfiler” package (v.4.6.2) in R (v.4.2.2; R Foundation for Statistical Computing, Vienna, Austria). Gene symbols were converted into Entrez IDs using data from the org. Hs.eg.db database. Biological process terms were identified using enrichGO (p < 0.05) and visualized as dot plots displaying gene ratio, gene count, and false discovery rate–adjusted p values.
Results
Patient Characteristics
A total of 1974 patients were diagnosed with EGFR-mutant NSCLC between 2014 and 2023 at our institution. Of these patients, 45 (2.3%) were subsequently diagnosed with tSCLC after EGFR-TKI treatment. Table 1 presents the baseline clinicodemographic characteristics of the study cohort. The patients’ median age at the time of small cell transformation was 62.3 (range: 34.3–85.6) years. Most of the patients were women (60%) and never-smokers (75.6%). Adenocarcinoma was the predominant initial histological subtype (88.9%). Most patients (93.4%) were diagnosed with stage IV disease before EGFR-TKI treatment, with 35.6% having stage IVA disease and 57.8% having stage IVB disease.
Table 1.
Patient Characteristics (n = 45)
| Age, years, median (range) | 62.3 (34.3 – 85.6) |
| Sex, n (%) | |
| Male | 18 (40) |
| Female | 27 (60) |
| Smoking, n (%) | |
| Never smoker | 34 (75.6) |
| Ever smoker | 11 (24.4) |
| Stage before EGFR-TKI, n (%) | |
| III | 3 (6.7) |
| IVA | 16 (35.6) |
| IVB | 26 (57.8) |
| Baseline EGFR mutation, n (%) | |
| Exon 19 deletion | 26 (57.8) |
| L858R | 17 (37.8) |
| L861Q | 2 (4.4) |
| First-line EGFR-TKI, n (%) | |
| Gefitinib | 12 (26.7) |
| Erlotinib | 6 (13.3) |
| Afatinib | 20 (44.4) |
| Osimertinib | 7 (15.6) |
| First-line TKI PFS, median (range) | 13.2 (0.43 – 48.9) |
| Treatment lines before transformation, median (range) | 2 (1-8) |
| Number of EGFR-TKI before transformation, median (range) | 1 (1-3) |
| Time from initial diagnosis to transformation, months, median (range) | 23.9 (5.6 – 81.73) |
| Time from 1L TKI to transformation, months, median (range) | 20.5 (5.4 – 64.76) |
| Last treatment before transformation | |
| EGFR-TKI | 32 (71.1) |
| Gefitinib | 4 (8.9) |
| Erlotinib | 3 (6.7) |
| Afatinib | 11 (24.4) |
| Osimertinib | 14 (31.1) |
| Chemotherapy | 13 (28.9) |
| Pathology report at transformation | |
| tSCLC | 29 (64.4) |
| Combined histology, including NSCLC and SCLC | 16 (35.6) |
| Brain metastasis at transformation | 20 (44.4) |
| Liver metastasis at transformation | 13 (28.9) |
| Brain as progression site at transformation | 4 (8.9) |
| Liver as progression site at transformation | 13 (28.9) |
tSCLC, transformed small cell lung cancer; TKI, tyrosine kinase inhibitor.
The most common EGFR mutations were exon 19 deletions (57.8%) and L858R (37.8%). First-line EGFR-TKIs included afatinib (44.4%), gefitinib (26.7%), osimertinib (15.6%), and erlotinib (13.3%). The median PFS with first-line EGFR-TKI treatment was 13.2 months (range: 0.43–48.9). Patients received a median of two lines of systemic therapy (range: 1–8) and one line of EGFR-TKI therapy (range: 1–3) before small cell transformation. The median duration from first-line TKI treatment to transformation was 20.5 months.
Before transformation, 71.1% of all patients were on EGFR-TKIs, mostly osimertinib (31.1%). At the time of tSCLC diagnosis, 64.4% of the patients had SCLC alone, and 35.6% patients had combined histology. At the time of transformation, brain and liver metastases were identified in 44.4% and 28.9% of the patients, respectively; liver progression was more frequent than brain progression (28.9% versus 8.9%). Thirty-two patients (71.1%) underwent post-transformation molecular testing; among them, 81.3% retained their original EGFR mutation, and only two patients (4.4% of the total cohort) had acquired the T790M mutation.
Treatment Pattern and Outcomes After SCLC Transformation
Table 2 presents the treatment patterns in patients with tSCLC. A total of 43 patients (95.6%) received systemic treatment, with the majority receiving either one (31.1%) or two (35.6%) lines of therapy. EP-based regimens, including EP alone (46.5%), EP plus ICIs (11.6%), and EP plus EGFR-TKIs (11.6%), were the most common first-line systemic regimens (69.8%) after small cell transformation. Other first-line regimens included paclitaxel-based chemotherapy with or without anti–vascular endothelial growth factor and ICIs (14%), TKI monotherapy (14%), and TKI plus oral etoposide (2.3%). Treatment was heterogeneous among the 28 patients (62.2%) who received second-line therapy after small cell transformation. EP-based regimens were used in 21.4% of all patients, whereas non-EP-based chemotherapeutic regimens, such as topotecan or paclitaxel combinations, were used in 35.7% of the patients. TKI-based regimens were also common (42.9%), either in combination with chemotherapy (25%) or as monotherapy (17.9%). Overall, 22.2% and 57.8% of patients received ICIs and EGFR-TKIs, respectively, after small cell transformation. EP-based regimens were used in 75.6% of patients, and a combination of TKIs and chemotherapy in 33.3% of patients.
Table 2.
Treatment Patterns After Transformation
| Number of lines, n (%) | |
| Without treatment | 2 (4.4) |
| 1 line | 14 (31.1) |
| 2 lines | 16 (35.6) |
| 3 lines | 7 (15.6) |
| ≥ 4 lines | 6 (13.3) |
| First-line systemic treatment category, n (%)a | |
| Total patients who received first-line treatment | 43 |
| EP-based regimen | 30 (69.8) |
| EP | 20 (46.5) |
| EP + ICI | 5 (11.6) |
| EP + TKI | 5 (11.6) |
| Other chemotherapy | 6 (14) |
| Paclitaxel + Cisplatin/Carboplatin | 1 (2.3) |
| Paclitaxel + Cisplatin/Carboplatin + ICI + anti-VEGF | 1 (2.3) |
| Paclitaxel | 1 (2.3) |
| Pemetrexed | 1 (2.3) |
| Oral etoposide | 2 (4.7) |
| TKI only | 6 (14) |
| TKI + oral etoposide | 1 (2.3) |
| Second-line systemic treatment category, n (%)b | |
| Total patients who received second-line treatment | 28 |
| EP-based regimen | 6 (21.4) |
| EP only | 5 (17.9) |
| EP + ICI | 1 (3.6) |
| Other chemotherapy | 10 (35.7) |
| Topotecan | 1 (3.6) |
| Paclitaxel + Cisplatin/Carboplatin | 5 (17.9) |
| Paclitaxel + Cisplatin/Carboplatin + ICI + anti-VEGF | 2 (7.1) |
| Paclitaxel | 2 (7.1) |
| Pemetrexed | 1 (3.6) |
| TKI + chemotherapy | 7 (25) |
| TKI + paclitaxel | 5 (17.9) |
| TKI + oral etoposide | 2 (7.1) |
| TKI only | 5 (17.9) |
| Ever receiving ICI after tSCLC, n (%) | 10 (22.2) |
| Ever receiving TKI after tSCLC, n (%) | 26 (57.8) |
| Ever receiving EP after tSCLC, n (%) | 34 (75.6) |
| Ever combining TKI with chemotherapy after tSCLC | 15 (33.3) |
Percentage represents the proportion of patients receiving this treatment among those who received first-line treatment (n = 43).
Percentage represents the proportion of patients receiving this treatment among those who received second-line treatment (n = 28). EP, etoposide plus platinum; ICI, immune checkpoint inhibitor; TKI, tyrosine kinase inhibitor; anti-VEGF, anti–vascular endothelial growth factor therapy; tSCLC, transformed small cell lung cancer.
Supplementary Table 2 presents the efficacy of first-line systemic therapy in treating tSCLC. Among 43 patients who received first-line therapy after small cell transformation, the ORR was 16.3%, and the disease control rate (DCR) was 53.5%. Patients who received EP-based regimens (n = 30) exhibited numerically higher ORR (20.0%) and DCR (60.0%) than those who received non-EP-based regimens (ORR = 7.7%, DCR = 38.5%), but the difference was not significant. A small subset of patients receiving EP plus ICIs (n = 5) exhibited numerical increases in ORR (40.0%) and DCR (80.0%), with those receiving EGFR-TKIs plus EP (n = 5) achieving disease control in all cases (DCR = 100.0%), despite having a modest ORR (20.0%). In patients receiving first-line systemic therapy, the median PFS duration and the median posttransformation survival (PTS) duration were 3.3 months (95% confidence interval [CI]: 1.9–4.9) and 9.2 months (95% CI: 7.7–10.7), respectively. The median OS from the initial NSCLC diagnosis was 35.9 months (range = 8.0–93.1 months). The overall median PFS duration was 2.8 months in patients receiving EP alone, 5.2 months in those receiving EP plus ICIs, and 11.2 months in those receiving EP plus TKIs. PFS was longer in patients receiving EP-based combination regimens than in those receiving EP alone (7.5 versus 2.8 months, p = 0.003; Fig. 1A). Furthermore, patients receiving EP-based regimens had a significantly longer PTS than those receiving other therapeutic regimens (10.4 versus 6.4 months, p = 0.035; Fig. 1B). No significant differences in PFS or PTS were observed between pure SCLC and combined histology (Supplementary Fig. 1).
Figure 1.
Post-transformation survival outcomes by treatment group. (A) Progression-free survival following first-line treatment with EP (etoposide plus platinum) combined with either an immune checkpoint inhibitor, a tyrosine kinase inhibitor, or EP alone. (B) Post-transformation overall survival comparing EP-based regimens versus non–EP-based regimens.
Efficacy of Subsequent Treatment and Histological Changes After Disease Progression
Supplementary Table 3 presents the efficacy of second-line systemic treatment for tSCLC. Paclitaxel-based regimens were the most frequently used second-line treatments (32.1%), yielding a modest ORR of 11.1% and a median PFS duration of 2.9 months (95% CI: 1.9–4.3). Other regimens, such as topotecan, EGFR-TKIs with or without chemotherapy, and miscellaneous chemotherapeutic treatments, conferred limited clinical benefit. Four patients received lurbinectedin after their previous treatments failed, and no response was observed (DCR = 25%).
A total of 17 patients underwent a repeat biopsy after disease progression following systemic treatment for tSCLC. SCLC remained the dominant histological subtype in eight of these patients (47%), with only two patients (11.8%) having a persistent combined histology. Notably, seven patients (41.2%) exhibited a reversion to adenocarcinoma histology. Overall, these findings confirmed dynamic histological changes and intrapatient heterogeneity after small cell transformation (Supplementary Table 4).
SCLC Subtype and Novel Biomarkers
IHC subtyping was successfully performed in 18 patients. Among them, six patients (33.3%) had the ASCL1-predominant subtype (SCLC-A), six (33.3%) had the NEUROD1-predominant subtype (SCLC-N), and three (16.7%) had the ASCL1–NEUROD1-codominant subtype (SCLC-AN). Three patients (16.7%) tested negative for all NAPY markers (NEUROD1, ASCL1, POU2F3, and YAP1). None of the patients exhibited predominance of POU2F3 or YAP1. No significant difference in PTS was observed among the IHC-defined subtypes.
As shown in Supplementary Table 5, DLL3 and B7-H3 expression at any level was detected in 87.5% and 66.7% of all tumors, respectively. The mean H-score was 77.5 for DLL3 and 24.4 for B7-H3. For tumors with both markers evaluated, the median H-score was significantly higher for DLL3 than for B7-H3 (40 versus 5, p = 0.002, Wilcoxon signed rank test). No significant differences in clinical characteristics or treatment outcomes were observed between DLL3-low/negative and DLL3-high (H-score > 50) patients or between B7-H3-positive and B7-H3-negative patients (Supplementary Tables 6 and 7). Patients with high DLL3 expression had a longer PTS than those with low or negative expression (median = 11.5 versus 7.1 months), but the difference was not significant (log-rank p = 0.405). Similarly, no significant difference in PTS was observed between B7-H3-positive and B7-H3-negative tumors (median = 9.5 versus 8.4 months, p = 0.611; Fig. 2).
Figure 2.
(A) Post-transformation survival between different molecular subtypes based on DLL3 (B) B7-H3.
Factors Associated With PTS
In the univariable analysis, the absence of brain metastases at transformation, the use of an EP-based regimen as the first-line treatment, the use of EGFR-TKIs after transformation, and the absence of the liver as the site of progression at transformation were associated with prolonged PTS. In the multivariable analysis, the presence of brain metastases (hazard ratio [HR] = 2.331, p = 0.019) and liver progression (HR = 1.674, p = 0.017) at transformation emerged as independent predictors of reduced PTS. Conversely, the use of an EP-based regimen (HR = 0.355, p = 0.011) and the administration of EGFR-TKIs after transformation (HR = 0.427, p = 0.012) were independently associated with improved survival (Supplementary Table 8).
Spatial Transcriptomic Profiles
Figure 3A presents the spatial transcriptomic profile of a patient with EGFR L858R-mutant lung adenocarcinoma (LUAD) who developed tSCLC after 18 months of treatment with afatinib. Tumor IHC analysis after transformation confirmed the upregulation of ASCL1 and DLL3 (80%) and the downregulation of B7-H3 (Fig. 3B). Spatial mapping revealed distinct histologic regions, with LUAD morphology in the pretransformation specimens and small-cell morphology in the posttransformation specimens (Fig. 3C). Transcriptomic analysis revealed upregulated ASCL1 and DLL3 expression in the tSCLC region, with CD276 expression lower than that observed in the LUAD region (Fig. 3D).
Figure 3.
IHC and spatial transcriptomics of a case with small cell carcinoma transformation. (A) Brief history (B) IHC for ASCL1, NeuroD1, DLL3 and B7-H3 (scale bar: 50 μm) (C) ST map pre- and posttransformation. (D) Gene expression of ASCL1, DLL3, and CD276 (B7-H3) after SCLC transformation. IHC, Immunohistochemistry; ST, Spatial transcriptomic; SCLC, small cell carcinoma.
In total, 33 and 110 genes were significantly upregulated and downregulated, respectively, after transformation from EGFR-mutant LUAD to tSCLC in both patient samples (Supplementary Table 9). The upregulated genes included neuroendocrine lineage-associated regulators, such as FOXA1, NKX2-1, and NCAM1, alongside metabolic and signaling modulators, such as ABAT, PRMT3, and PLXNA2. By contrast, the downregulated genes primarily included immune-related, stromal, and epithelial markers, such as C1QA, CD74, IFI27, MMP2, TIMP2, SFRP2, and KRT19. These findings indicate the broad suppression of inflammatory signaling, extracellular matrix remodeling, and epithelial lineage identity after transformation.
GO enrichment analysis revealed that the upregulated genes were significantly associated with nervous system development, cranial nerve development, and protein secretion (Fig. 4A), consistent with neuroendocrine reprogramming. It further revealed that the downregulated genes were associated with pathways related to the humoral immune response, extracellular matrix organization, complement activation, and B-cell-mediated immunity (Fig. 4B), highlighting a marked transition toward an immune-suppressed and mesenchymal-depleted state.
Figure 4.
GO Pathway analysis of transformed SCLC by spatial transcriptomics. GO, Gene Ontology.
Union-based analysis of genes significantly altered in either P1 or P2 revealed a broader DEG set associated with small cell transformation. This set comprised 1047 upregulated genes and 963 downregulated genes. GO analysis of the upregulated genes revealed enrichment of cell cycle, mitotic spindle, and chromosome segregation pathways, indicating an increase in proliferative activity (Supplementary Fig. 2A). By contrast, GO analysis of the downregulated genes revealed enrichment of extracellular matrix organization, wound healing, and immune regulation pathways, consistent with stromal depletion and an immune-cold tumor microenvironment (Supplementary Fig. 2B).
Discussion
In this large-scale cohort study, we explored treatment patterns, survival outcomes, and biomarker profiles in Taiwanese patients with EGFR-mutant tSCLC after EGFR-TKIs failure. These patients had a median PTS duration of 9.2 months, consistent with the aggressive nature of this resistance mechanism. EP-based regimens remained the most common therapeutic options for these patients, yielding modest response rates and survival benefits, particularly when combined with EGFR-TKIs or ICIs. Brain metastases and liver progression emerged as independent negative prognostic factors. Beyond clinical outcomes, we observed a high prevalence of the therapeutic target DLL3 (87.5%). Spatial transcriptomic profiling provided novel insights into the tSCLC microenvironment, defining a distinct cold immune landscape marked by neuroendocrine lineage reprogramming, considerable immune exclusion, and stromal depletion.
A study reported that neuroendocrine transformation served as a mechanism of acquired resistance in 2% to 15% of all patients with lung cancer.1 In our cohort, the incidence of tSCLC was 2.3%, which is consistent with the aforementioned finding. However, this consistency reflects an underestimation, as not all patients in our study underwent a repeat biopsy at disease progression; therefore, the actual incidence may have been higher. Currently, we do not have access to noninvasive diagnostic platforms, such as liquid biopsy, to reliably detect histologic transformation. Novel approaches, such as epigenomic cell-free DNA assays or circulating tumor RNA analyses, may facilitate earlier identification of patients with tSCLC.21,22
To the best of our knowledge, no optimal treatment strategies have yet been developed for patients with small cell transformation. EP has traditionally been used as the standard treatment, with a prior study indicating a clinical response rate of 54%.6 Recent studies have evaluated the efficacy of EP plus ICIs in this context and noted potential survival benefits.13,23 In the present study, the ORR was only 15% in patients receiving EP alone, whereas it was 40% in those receiving EP plus ICIs; this finding underscores the importance of combination therapy. Our findings are consistent with those of a Taiwanese study that reported the limited efficacy of EP alone in patients with tSCLC.24 Whether EGFR-TKIs or ICIs should be prioritized in combination options remains unclear. Additional clinical factors, such as the presence of brain metastases, which may favor TKI use, or PD-L1 expression levels, which may influence ICI responsiveness, should be considered when tailoring treatment strategies. In our cohort, we observed substantial histologic heterogeneity. Specifically, 41.2% of patients who underwent a repeat biopsy after disease progression following systemic treatment for tSCLC reported adenocarcinoma histology. This finding suggests dynamic histologic evolution or coexisting subclones within the tumor, which underscores the importance of repeat biopsy. Hence, a combinatorial strategy may be required to effectively address both the small-cell and non–small cell components of the disease.
In addition to conventional regimens, several novel therapeutic options have been explored for the treatment of tSCLC. Lurbinectedin is an oncogenic transcription inhibitor that has proven effective in preclinical tSCLC models. Although combining lurbinectedin with osimertinib has reported promising preliminary clinical activity in this setting,25,26 prospective trials are still required to validate the safety and efficacy of this combined strategy in patients with EGFR-mutant tSCLC. Tarlatamab has reported superiority over second-line chemotherapy in patients with relapsed SCLC, emerging as a novel standard of care following progression after first-line treatment.27 However, its efficacy in patients with tSCLC appears to be limited, and further research is required to evaluate its clinical benefit—either as monotherapy or in combination with other agents—in this population.28 A clinical trial evaluating the efficacy of tarlatamab plus etoposide, carboplatin, or atezolizumab in the treatment of tSCLC is currently ongoing (ClinicalTrials.gov identifier: NCT06830694). Taken together, the findings underscore the urgent need for more effective biomarker-informed, individualized treatment strategies for patients with EGFR-mutant NSCLC who exhibit histologic small cell transformation.
SCLC has historically been regarded as a single disease entity. However, the latest research suggests biological heterogeneity, leading to the proposal of four molecular subtypes based on key transcription factors: ASCL1 (SCLC-A), NEUROD1 (SCLC-N), POU2F3 (SCLC-P), and YAP1 (SCLC-Y).17,29 These subtypes also appear to apply to tSCLC.9 In our previous IHC-based study, we identified SCLC-A as the most common subtype when we followed a single-marker dominance approach.19 However, we observed the frequent coexpression of transcription factors, which complicated the classification process. In the present study, we used an updated IHC subtyping framework adapted from the study by Park et al. to accommodate codominant expression (SCLC-AN).20 Using this framework, we identified an equal distribution of SCLC-A and SCLC-N (n = 6 each), with additional cases classified as SCLC-AN (n = 3) and triple-negative (n = 3). No cases with POU2F3 or YAP1 predominance were noted. Notably, no significant correlation was observed between IHC-defined subtypes and PTS in our cohort, which warrants further investigation in larger data sets.
In addition to subtyping, we evaluated the expression of two surface markers with therapeutic relevance: DLL3 and B7-H3. To the best of our knowledge, this is the first study to assess these two markers specifically in tSCLC specimens. DLL3 was detected in 87.5% of all tumors, consistent with the findings of studies reporting expression rates of 80% to 93% in patients with tSCLC.23,30 B7-H3 expression was also observed in most of the cases. Notably, DLL3 was associated with significantly higher median H-scores than B7-H3; this finding highlights DLL3 as a neuroendocrine lineage marker and potential therapeutic target. Despite their frequent expression, neither DLL3 nor B7-H3 was associated with PTS in our cohort. A study indicated that patients with tSCLC who have stronger neuroendocrine features have a poorer prognosis.31 Therefore, although DLL3 and B7-H3 represent promising therapeutic targets, particularly in the context of emerging antibody–drug conjugates and T-cell engagers, they do not currently serve as reliable prognostic biomarkers in tSCLC. Larger prospective studies are required to validate their clinical utility in risk stratification and treatment selection.
Our spatial transcriptomic analysis revealed the tumor microenvironmental shifts accompanying transformation from EGFR-mutant LUAD to tSCLC. The observed transcriptional reprogramming, characterized by the upregulation of neuroendocrine lineage drivers (e.g., FOXA1, NKX2-1, and NCAM1) and the concurrent downregulation of immune and epithelial genes (e.g., C1QA, CD74, and KRT19), supports a marked shift toward an immune-excluded phenotype. GO pathway analysis revealed a loss of stromal components (MMP2 and TIMP2), complement activation, and B-cell activity, confirming immune suppression and stromal depletion in the tSCLC state. In this context, the marked immune suppression identified by spatial transcriptomics provides a biological explanation for the modest benefit of ICIs observed in the present and previous studies. The enrichment observed in mitotic and cell cycle pathways underscores the highly proliferative nature of tSCLC, which suggests a potential vulnerability to cytotoxic agents. These spatial transcriptomic findings are consistent with our clinical observations and highlight the need for therapeutic strategies that address both the neuroendocrine lineage dominance and the immune desert phenotype of tSCLC.
This study has several limitations. First, it was conducted at a single institution, which may limit the generalizability of our findings. Second, the restaging assessments conducted and the treatment decisions made after histologic transformation were not standardized but were rather left to the discretion of the attending physicians. Consequently, the therapeutic regimens were heterogeneous, and only a small number of patients received EP-based combination therapy, which may have influenced the efficacy data. Finally, our biomarker analysis focused on selected IHC and therapeutic markers; other promising targets, such as SEZ6 and Trop2, were not evaluated. Multicenter studies involving protocol-driven treatment and expanded biomarker panels are required to validate and extend our findings.
Conclusions
This real-world study provides a comprehensive overview of patients with EGFR-mutant NSCLC who exhibit small cell transformation after failure of EGFR-TKI treatment. Despite the frequent use of EP-based chemotherapy, clinical outcomes remain poor in these patients, particularly in those with brain metastases or liver progression. Although the addition of EGFR-TKIs or ICIs may offer incremental benefit, the optimal treatment sequence remains unclear. Translational analyses in this study revealed a distinct “cold” tumor microenvironment characterized by neuroendocrine reprogramming and marked immune exclusion, which provides a mechanistic rationale for the limited efficacy of ICIs observed in this setting. We also observed highly prevalent DLL3 and B7-H3 expression, which represents a promising avenue for biomarker-directed therapeutic options such as T-cell engagers or antibody–drug conjugates. Future precision medicine strategies aimed at tSCLC should prioritize these novel molecular targets and explore combination approaches designed to overcome the immunosuppressive microenvironment and improve outcomes for this challenging and heterogeneous disease.
CRediT Authorship Contribution Statement
Chi-Lu Chiang: Conceptualization, Methodology, Formal analysis, Investigation, Data curation, Writing - original draft, Writing - review & editing, Visualization, Funding acquisition.
Ying-Ting Liao: Investigation, Data curation, Formal analysis, Funding acquisition, Writing - review & editing.
Yi-Chen Yeh: Investigation, Formal analysis, Writing - review & editing.
Ruei-Lin Sun: Investigation, Writing - review & editing.
Hsu-Ching Huang: Investigation, Writing - review & editing.
Yung-Hung Luo: Investigation, Writing - review & editing.
Chia-I Shen: Investigation, Writing - review & editing.
Heng-Sheng Chao: Investigation, Writing - review & editing.
Yen-Han Tseng: Investigation, Writing - review & editing.
Tsu-Hui Shiao: Investigation, Writing - review & editing.
Ruihong Wang: Investigation, Writing - review & editing.
Haitao Wang: Investigation, Writing - review & editing.
David S. Schrump: Investigation, Writing - review & editing.
Yuh-Min Chen: Conceptualization, Investigation, Resources, Writing - review & editing, Supervision, Funding acquisition.
Disclosure
CLC received honoraria from Amgen, ACT Genomic, AstraZeneca, Bristol Myers Squibb, Boehringer Ingelheim, Chugai, Daiichi Sankyo, Johnson & Johnson, Merck KGaA, Merck Sharp & Dohme, Pfizer and Roche. YTL received honoraria from AstraZeneca, Merck Sharp & Dohme, Pfizer, Chugai, Orient EuroPharma, Boehringer Ingelheim, and Roche. CIS received honoraria from Amgen, AstraZeneca, Boehringer Ingelheim, and Roche. YHL received honoraria from AstraZeneca, Boehringer Ingelheim, Pfizer, Roche, and Bristol Myers Squibb. YMC received honoraria from Boehringer Ingelheim, Eli Lilly, Roche, Chugai, Merck Sharp & Dohme, Pfizer, Novartis, Bristol Myers Squibb, Ono Pharmaceutical, AstraZeneca, and Takeda Oncology. The remaining authors declare no conflicts of interest that may be relevant to the content of this manuscript.
Acknowledgments
The authors thank the Single Cell and Spatial Core (SCSC) of the Center for Cancer Research (CCR), National Cancer Institute (NCI); the Frederick National Laboratory for Cancer Research (FNLCR); and the Molecular Histopathology Laboratory (MHL) for their support in library preparation and spatial transcriptomics profiling. In addition, the authors thank the National Genomics Center for Clinical and Biotechnological Applications, a national core facility operated by the Cancer and Immunology Research Center at National Yang Ming Chiao Tung University, for providing technical services in Visium data analysis. This core facility is supported by the National Core Facility for Biopharmaceuticals (NCFB). This manuscript was edited by Wallace Academic Editing.
Funding
This work was supported by Taipei Veterans General Hospital (grant number V112E-002-5 to CLC and V113A-003 to YTL), the Yen Tjing Ling Medical Foundation (grant numbers: CI-114-8 and CI-115-14 to CLC), and the National Science and Technology Council of Taiwan (grant number 111-2926-I-075-501-G to YMC).
Ethics Approval
The study protocol was approved by the Institutional Review Board of Taipei Veterans General Hospital, Taipei, Taiwan (2022-08-017B, 2023-06-016CC and 2024-02-012A).
Consent for Publication
Not applicable. All data were anonymized.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Footnotes
Cite this article as: Chiang CL, Liao YT, Yeh YC, et al. EGFR-mutant non–small cell lung cancer with small cell transformation: clinicopathological features, treatment landscape, and biomarker profiles. JTO Clin Res Rep 2026;7:101050
Note: To access the supplementary material accompanying this article, visit the online version of the JTO Clinical and Research Reports at www.jtocrr.org and at 10.1016/j.jtocrr.2026.101050.
Supplementary Data
References
- 1.Sato Y., Saito G., Fujimoto D. Histologic transformation in lung cancer: when one door shuts, another opens. Ther Adv Med Oncol. 2022;14 doi: 10.1177/17588359221130503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Huang J., Zhang S.L., Zhou C., et al. Genomic and transcriptomic analysis of neuroendocrine transformation in ALK-rearranged lung adenocarcinoma after treatments with sequential ALK inhibitors: a brief report. JTO Clin Res Rep. 2022;3 doi: 10.1016/j.jtocrr.2022.100338. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Lin J.J., Langenbucher A., Gupta P., et al. Small cell transformation of ROS1 fusion-positive lung cancer resistant to ROS1 inhibition. NPJ Precis Oncol. 2020;4:21. doi: 10.1038/s41698-020-0127-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Sequist L.V., Waltman B.A., Dias-Santagata D., et al. Genotypic and histological evolution of lung cancers acquiring resistance to EGFR inhibitors. Sci Transl Med. 2011;3 doi: 10.1126/scitranslmed.3002003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Yu H.A., Arcila M.E., Rekhtman N., et al. Analysis of tumor specimens at the time of acquired resistance to EGFR-TKI therapy in 155 patients with EGFR-mutant lung cancers. Clin Cancer Res. 2013;19:2240–2247. doi: 10.1158/1078-0432.CCR-12-2246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Marcoux N., Gettinger S.N., O’Kane G., et al. EGFR-mutant adenocarcinomas that transform to small-cell lung cancer and other neuroendocrine carcinomas: clinical outcomes. J Clin Oncol. 2019;37:278–285. doi: 10.1200/JCO.18.01585. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Zhang B., Lewis W., Stewart C.A., et al. Brief report: comprehensive clinicogenomic profiling of small cell transformation from EGFR-mutant NSCLC informs potential therapeutic targets. JTO Clin Res Rep. 2024;5 doi: 10.1016/j.jtocrr.2023.100623. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ferrer L., Giaj Levra M., Brevet M., et al. A brief report of transformation from NSCLC to SCLC: molecular and therapeutic characteristics. J Thorac Oncol. 2019;14:130–134. doi: 10.1016/j.jtho.2018.08.2028. [DOI] [PubMed] [Google Scholar]
- 9.Quintanal-Villalonga A., Taniguchi H., Zhan Y.A., et al. Multiomic analysis of lung tumors defines pathways activated in neuroendocrine transformation. Cancer Discov. 2021;11:3028–3047. doi: 10.1158/2159-8290.CD-20-1863. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ishioka K., Yasuda H., Hamamoto J., et al. Upregulation of FGF9 in lung adenocarcinoma transdifferentiation to small cell lung cancer. Cancer Res. 2021;81:3916–3929. doi: 10.1158/0008-5472.CAN-20-4048. [DOI] [PubMed] [Google Scholar]
- 11.Horn L., Mansfield A.S., Szczęsna A., et al. First-line atezolizumab plus chemotherapy in extensive-stage small-cell lung cancer. N Engl J Med. 2018;379:2220–2229. doi: 10.1056/NEJMoa1809064. [DOI] [PubMed] [Google Scholar]
- 12.Paz-Ares L., Dvorkin M., Chen Y., et al. Durvalumab plus platinum-etoposide versus platinum-etoposide in first-line treatment of extensive-stage small-cell lung cancer (Caspian): a randomised, controlled, open-label, phase 3 trial. Lancet. 2019;394:1929–1939. doi: 10.1016/S0140-6736(19)32222-6. [DOI] [PubMed] [Google Scholar]
- 13.Zhang C.Y., Sun H., Su J.W., et al. A potential treatment option for transformed small-cell lung cancer on PD-L1 inhibitor-based combination therapy improved survival. Lung Cancer. 2023;175:68–78. doi: 10.1016/j.lungcan.2022.11.016. [DOI] [PubMed] [Google Scholar]
- 14.Trigo J., Subbiah V., Besse B., et al. Lurbinectedin as second-line treatment for patients with small-cell lung cancer: a single-arm, open-label, phase 2 basket trial. Lancet Oncol. 2020;21:645–654. doi: 10.1016/S1470-2045(20)30068-1. [DOI] [PubMed] [Google Scholar]
- 15.Ahn M.J., Cho B.C., Felip E., et al. Tarlatamab for patients with previously treated small-cell lung cancer. N Engl J Med. 2023;389:2063–2075. doi: 10.1056/NEJMoa2307980. [DOI] [PubMed] [Google Scholar]
- 16.Johnson M., Awad M., Koyama T., et al. OA05.05 ifinatamab deruxtecan (I-DXd; DS-7300) in patients with refractory SCLC: a subgroup analysis of a phase 1/2 study. J Thorac Oncol. 2023;18:S54–S55. [Google Scholar]
- 17.Baine M.K., Hsieh M.S., Lai W.V., et al. SCLC subtypes defined by ASCL1, NEUROD1, POU2F3, and YAP1: a comprehensive immunohistochemical and histopathologic characterization. J Thorac Oncol. 2020;15:1823–1835. doi: 10.1016/j.jtho.2020.09.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Qu S., Fetsch P., Thomas A., et al. Molecular subtypes of primary SCLC tumors and their associations with neuroendocrine and therapeutic markers. J Thorac Oncol. 2022;17:141–153. doi: 10.1016/j.jtho.2021.08.763. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Chiang C.L., Huang H.C., Luo Y.H., et al. Clinical utility of immunohistochemical subtyping in patients with small cell lung cancer. Lung Cancer. 2024;188 doi: 10.1016/j.lungcan.2024.107473. [DOI] [PubMed] [Google Scholar]
- 20.Park S., Hong T.H., Hwang S., et al. Comprehensive analysis of transcription factor-based molecular subtypes and their correlation to clinical outcomes in small-cell lung cancer. EBioMedicine. 2024;102 doi: 10.1016/j.ebiom.2024.105062. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.El Zarif T., Meador C.B., Qiu X., et al. Detecting small cell transformation in patients with advanced EGFR mutant lung adenocarcinoma through epigenomic cfDNA profiling. Clin Cancer Res. 2024;30:3798–3811. doi: 10.1158/1078-0432.CCR-24-0466. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Nesselbush M.C., Luca B.A., Jeon Y.J., et al. An ultrasensitive method for detection of cell-free RNA. Nature. 2025;641:759–768. doi: 10.1038/s41586-025-08834-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Saalfeld F.C., Möller J., Christopoulos P., et al. Small cell transformation in EGFR-mutated non-small cell lung cancer: DLL3 expression and efficacy of immune checkpoint inhibitors or tyrosine kinase inhibitors combined with chemotherapy. Eur J Cancer. 2024;213 doi: 10.1016/j.ejca.2024.115065. [DOI] [PubMed] [Google Scholar]
- 24.Chen C.Y., Ko H.W., Hu P.W., et al. Impact of subsequent treatment on clinical outcomes in patients with EGFR-mutant non-small cell lung cancer transformed to small-cell lung cancer. Lung Cancer (Auckl) 2025;16:25–37. doi: 10.2147/LCTT.S516527. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Singh A., Sridhar A., Poddar A., et al. Combination of lurbinectedin and osimertinib for treatment of EGFR-mutated transformed small cell lung cancer: a brief report. JTO Clin Res Rep. 2025;6 doi: 10.1016/j.jtocrr.2025.100807. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Chakraborty S., Coleman C., Manoj P., et al. De novo and histologically transformed small-cell lung cancer is sensitive to lurbinectedin treatment through the modulation of EMT and NOTCH signaling pathways. Clin Cancer Res. 2023;29:3526–3540. doi: 10.1158/1078-0432.CCR-23-0471. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Mountzios G., Sun L., Cho B.C., et al. Tarlatamab in small-cell lung cancer after platinum-based chemotherapy. N Engl J Med. 2025;393:349–361. doi: 10.1056/NEJMoa2502099. [DOI] [PubMed] [Google Scholar]
- 28.Cooper A.J., Herzberg B., Parma M., et al. Multi-institution analysis of tarlatamab for the treatment of EGFR-mutant transformed small cell lung cancer. JCO Precis Oncol. 2025;9 doi: 10.1200/PO-25-00538. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Rudin C.M., Poirier J.T., Byers L.A., et al. Molecular subtypes of small cell lung cancer: a synthesis of human and mouse model data. Nat Rev Cancer. 2019;19:289–297. doi: 10.1038/s41568-019-0133-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Aredo J.V., Singhal S., Berry G.J., et al. DLL3 immunohistochemical expression in neuroendocrine-transformed EGFR-mutant lung cancer and two cases of tarlatamab therapy. JTO Clin Res Rep. 2025;6 doi: 10.1016/j.jtocrr.2025.100913. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Chen L., Zhang X.H., Mao Z.J., et al. Clinical outcomes and neuroendocrine features of transformed versus primary small-cell lung cancer. Lung Cancer. 2025;207 doi: 10.1016/j.lungcan.2025.108714. [DOI] [PubMed] [Google Scholar]
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 that support the findings of this study are available from the corresponding author upon reasonable request.