Skip to main content
NCBI home page
Scientific Reports logoLink to Scientific Reports
. 2025 May 4;15:15592. doi: 10.1038/s41598-025-00222-z

RETRACTED ARTICLE: Novel prognostic biomarkers in small cell lung cancer reveal mutational signatures, genomic mutations, and immune implications

Yuting Li 1,#, Chen Song 2,#, Haijun Wang 3, Wenyu Di 3, Yangyang Chen 4, Yuanyuan Hu 1, Peiheng Li 4, Jie Chen 4, Yanfeng Ren 5,✉, Jing Gong 6,7,✉, Qinghua Wang 1,5,✉
PMCID: PMC12050310  PMID: 40320401

Abstract

Small cell lung cancer (SCLC) is a highly malignant lung cancer subtype with a dismal prognosis and limited treatment options. This study aimed to identify new prognostic molecular biomarkers for SCLC and explore their immune-related implications for treatment strategies. We analyzed 200 SCLC samples via whole-exome sequencing (WES) and 313 samples by targeted sequencing. A smoking-related SBS4 mutational signature was linked to poorer prognosis and lower tumor mutational burden (TMB), while the APOBEC-mediated SBS13 signature was associated with better prognosis and higher TMB. We identified a molecular subtype with the worst outcomes and lowest TMB in both cohorts. Among 38 high-frequency mutated genes associated with SCLC prognosis, only UNC13A mutations were beneficial. Patients with UNC13A mutations had favorable immune infiltration and tumor immunogenicity. Additionally, TP53 splice site mutations were related to the worst survival outcomes. In conclusion, we discovered new molecular biomarkers for SCLC prognosis. Our findings on their immunological characteristics offer insights for developing novel SCLC treatment strategies.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-00222-z.

Keywords: SCLC, Mutational signatures, Molecular subtypes, Frequently mutated genes, Prognosticators, Immunotherapeutic potential

Subject terms: Small-cell lung cancer, Tumour biomarkers, Mutation

Introduction

Small cell lung cancer (SCLC) accounts for approximately 15% of all lung cancers, characterized by abnormally high proliferation rates, marked early metastatic tendencies, and poor prognosis1. SCLC is closely associated with exposure to tobacco carcinogens. Most patients present with metastases at diagnosis, and only one-third are potentially curable1. Although clinical progress in SCLC has been slow, insights into disease biology have identified new vulnerabilities and potential therapeutic strategies2. The recent introduction of immune checkpoint inhibitor (ICI) therapy has brought new hope to SCLC patients, with a subset achieving long-term benefit3. Currently, the most urgent need is to identify stable prognostic molecular markers and develop therapeutic strategies for patients most likely to respond to treatment, thereby enabling more patients to benefit from antitumor immunotherapy.

Tumor mutational signatures are specific mutation patterns in the genome caused by endogenous and exogenous factors, which are presented through 96 mutation patterns formed by six main substitutions and their 3’ and 5’ base contexts4. Mutational signatures have been proven to be closely related to tumor biological characteristics, survival benefits, and immune treatment response4. Abbas et al. pointed out that the evolution of esophageal adenocarcinoma was linked with the dynamic changes of multiple mutation signatures5. A recent study showed that age-related mutational signature was significantly enriched in triple-negative breast cancer patients with poor survival and reduced immune activity6, and this mutation pattern exhibited resistance to ICI treatment in non-small cell lung cancer (NSCLC) and melanoma7. In addition, the presence of multiple mutational signatures has been demonstrated to be related to the efficacy of tumor immunotherapy.

Driven by precision individualized therapy, the exploration of tumor molecular subtypes has become extremely important. Currently, tumor molecular classification mainly relies on transcriptome data, such as mRNA expression and LncRNA expression. Recent studies have pointed out that with the help of mutational signature activity, potential subtypes related to prognosis and treatment sensitivity could also be identified. A recent study has confirmed the prognostic molecular subtypes relevant to mutation activity in esophageal squamous cell carcinoma8. Utilizing the extracted six mutational signature activities, Zhang et al. conducted consensus clustering analysis and uncovered two molecular subtypes in melanoma. These subtypes demonstrated substantial differences in the prognosis of melanoma patients undergoing ICI treatment9. Shi et al. identified female-specific immunotherapy subtypes in melanoma via the clustering of four mutational signature activities10.

In this work, we integrated somatic mutation data and clinical information from SCLC samples derived from whole-exome sequencing and targeted sequencing. We explored molecular indicators connected with SCLC prognosis from three aspects: mutational signatures, molecular subtypes, and frequently mutated genes. Furthermore, we provided theoretical support for improving clinical treatment strategies through in-depth analysis of immunological features.

Methods

Genomic data and clinical information of SCLC samples

A total of 200 SCLC samples that underwent whole-exome sequencing (WES) and their corresponding clinical information from two recently published studies11,12 were included in this study as the discovery dataset. Among them, only the study by George et al. included overall survival follow-up data11. In addition, the 80 SCLC samples sourced from the study by George et al. contained transcriptome mRNA expression data, based on which we conducted subsequent transcriptome-related analyses.

A total of 313 SCLC samples that underwent MSK-IMPACT targeted mutation sequencing from the MSKCC center13 were also included in this study for validation of relevant results. MSK-IMPACT is one of the commonly used sequencing methods, capable of detecting mutations in over 400 driver genes of common and rare tumor types and obtaining the final mutational landscape14. Detailed study design was exhibited in Fig. 1.

Fig. 1.

Fig. 1

Open in a new tab

Flowchart of this study. Integration of somatic mutational profiles and clinical information to discover novel indicators of SCLC prognosis and treatment potentials.

Detection of mutational signatures operative in SCLC

Bayesian non-negative matrix factorization (NMF)15 was employed to extract mutational signatures present in the WES SCLC cohort. Firstly, the mutation data was integrated into a feature matrix format. Then, the NMF algorithm decomposed the matrix into two non-negative submatrices, one representing the extracted detailed mutation signatures and the other representing the corresponding mutation activities. Finally, the extracted mutational signatures were compared with the known single base substitutions (SBS) signatures in the COSMIC database (v3.4)16 based on cosine similarity to determine the final results. In this study, all mutational signatures were treated as binary variables, following a previous study17: if the activity of a mutational signature exceeded 25%, it was considered to be present in the sample.

For the targeted sequencing cohort, the SigProfilerAssignment tool18 was used to extract specific mutational signatures. SigProfilerAssignment is a new tool developed by the Alexandrov lab. This tool assigns previously known mutational signatures to individual sequencing samples, utilizing a custom-implemented forward stagewise algorithm and non-negative least squares.

Consensus clustering and molecular subtyping

Unsupervised class discovery is a highly valuable technique in cancer research, where it may reveal intrinsic groups sharing biological features. The consensus clustering method provides quantitative and visual stability evidence for estimating the number of unsupervised classes in a dataset. In this study, we utilized the ConsensusClusterPlus R package19 with consensus clustering functionality, which extends novel features and visualization capabilities, including item tracking, item consensus, and cluster consensus plots. Specifically, based on the extracted mutation feature activities, we employed the partition around medoids (PAM) method for clustering and performed 500 bootstraps, each including 80% of the samples, using Euclidean distance as the distance metric. To achieve more accurate results, we simultaneously explored the classification effects under scenarios with cluster numbers ranging from 2 to 10, and ultimately determined the result based on relevant indicators.

GSEA and GSVA

Gene set enrichment analysis (GSEA) was utilized to explore significantly enriched signaling pathways between specific groups. Initially, we conducted a whole-genome differential expression analysis based on distinct groups and then extracted each gene and its corresponding fold change (FC) value from the differential results for subsequent analysis. The GSEA function in the clusterProfiler package20 was employed to perform pathway enrichment analysis based on the FC values sorted in descending order. Signaling pathways from four databases, including KEGG, GO BP, HALLMARK, and REACTOME, were used as background annotations.

The single sample GSEA (ssGSEA) function from the gene set variation analysis (GSVA) package21 was employed to calculate the enrichment levels of specific molecular signatures or immune cells. Based on the acquired characteristic gene sets of molecular signatures or cell subtypes, ssGSEA was able to evaluate the enrichment scores corresponding to each sample. The ssGSEA algorithm ranks genes in a sample by expression level and calculates an enrichment score based on the relative representation of genes in a set compared to the whole genome using a running-sum statistic.

Immunocyte infiltration analysis and immune-related signatures

The method proposed by Charoentong et al.22 was used to evaluate the abundance of tumor immune cell infiltration, which identified 28 immune cell subtypes that can be classified into three types: anti-tumor, tumor-promoting, and neutral cells. Based on the characteristic gene sets corresponding to each immune cell (Table S1), the ssGSEA method was employed to estimate the infiltration abundance.

We have integrated recently reported representative immune response and suppression molecular signatures (Table S2), including T cell-inflamed signature23, cytolytic activity signature24, and WNT TGFβ signature25. The enrichment scores of each signature were evaluated using ssGSEA, and the differences between distinct subgroups were explored.

Statistical and mathematical analysis

Statistical and mathematical analysis methods as well as related graphic visualizations were primarily conducted using R software. Detailed mutation patterns of mutational signatures and high-frequency mutated genes were achieved through the maftools package. Heatmaps representing the differences in immune cell infiltration abundance between different groups were generated using the pheatmap package. Survival curves were plotted using the Kaplan-Meier method, and differences between subgroups were tested using the log-rank method. The multivariate Cox regression models for correcting clinical confounding factors were implemented using the forestmodel package. P values less than 0.05 were considered statistically significant.

Results

Two mutational signatures linked with SCLC outcome in the WES cohort

In the integrated WES cohort, a total of 46,583 somatic mutations were detected. These mutations were characterized by C > T dominated SNP missense mutations (Figure S1). The median mutation count for all samples was 207. We compared the tumor mutational burden (TMB) of this SCLC cohort with other tumor types in the TCGA project, revealing that SCLC had a relatively high mutation burden (Figure S2).

Using the Bayesian NMF algorithm, we extracted potential mutational signatures present in the SCLC genome. The cophenetic metric plot indicated that when the number of signatures was five, the line declined most rapidly (Fig. 2A), suggesting the possible existence of five mutational signatures in the WES cohort. Further, we compared the extracted mutational signatures with known SBS signatures in the COSMIC database, ultimately identifying SBS4 (smoking related), SBS5 (unknown etiology), SBS13 (APOBEC cytidine deaminase related), SBS40 (unknown etiology), and SBS60 (sequencing artefact) signatures (Fig. 2B). Their detailed mutation characteristics and mutational activities were presented in Fig. 2C and Table S3, respectively. We then explored the prognostic abilities of the above five mutational signatures. Kaplan-Meier survival analysis demonstrated that SCLC patients carrying the SBS4 signature exhibited significantly inferior survival outcomes (Log-rank test P = 0.026; Fig. 2D). This association was further confirmed by a multivariate Cox model incorporating age, gender, stage, ethnicity, and sample type (HR: 1.82, 95% CI: 0.97–3.42, P = 0.064; Fig. 2E). Meanwhile, we found significantly improved clinical outcomes in SCLC patients carrying the SBS13 signature (Log-rank test P = 0.021; Fig. 2F), and this result was also verified by a multivariate adjusted Cox regression model (HR: 0.37, 95% CI: 0.14–0.96, P = 0.039; Fig. 2G). The associations of SBS5, SBS40, and SBS60 mutational signatures with SCLC survival risk were presented in Figure S3 A-S3 C, with no statistically significant survival differences observed (Log-rank test all P > 0.05).

Fig. 2.

Fig. 2

Open in a new tab

Determination of two mutational signatures associated with SCLC prognosis in the WES cohort. (A) Cophenetic metric plot of mutational signature numbers determined by the NMF algorithm. (B) Comparison of detected five mutational signatures with validated SBS signatures in the COSMIC database. (C) Detailed mutational features of detected five SBS signatures. (D) Kaplan-Meier survival curves stratified by the SBS4 signature status. (E) Multivariate Cox regression model of SBS4 signature was conducted with several clinical factors taken into consideration. (F) Kaplan-Meier survival curves stratified by the SBS13 signature status. (G) Multivariate Cox regression model of SBS13 signature was conducted with several clinical factors taken into consideration.

Validation of prognostic abilities of SBS4 and SBS13 and tumor immunogenicity

A total of 3231 somatic mutations were included in the targeted sequencing cohort. Unlike the WES cohort, missense SNP mutations in this cohort were primarily C > A substitutions (Figure S4). We also extracted aforementioned five mutational signatures (Table S4) and assessed their association with SCLC prognosis. Kaplan-Meier survival analysis revealed that the presence of the SBS4 signature was associated with significantly reduced SCLC survival outcomes (Log-rank test P = 0.003; Fig. 3A). Multivariate Cox regression analysis, incorporating age, gender, and sample type, yielded consistent findings (HR: 2.07, 95% CI: 1.28–3.35, P = 0.003; Fig. 3B). Furthermore, based on both univariate (Log-rank test P = 0.024; Fig. 3C) and multivariate survival analyses (HR: 0.64, 95% CI: 0.42–0.97, P = 0.039; Fig. 3D), we also observed a correlation between the presence of the SBS13 signature and favorable prognosis. In both the WES and targeted sequencing SCLC cohorts, we incorporated the two extracted mutational signatures, SBS4 and SBS13, into the multivariate Cox regression models to obtain more accurate results. The results indicated that the presence of the SBS4 signature was still associated with a poorer prognosis, while the SBS13 signature was associated with a better prognosis in SCLC (Figure S5A, S5B). We subsequently used the Benjamini-Hochberg method to correct all the P values of the multivariable Cox regression model in Figure S5. The correction results of the WES SCLC dataset showed that the corrected P values of the SBS4 and SBS13 signatures were 0.109 and 0.096, respectively. In the targeted sequencing dataset, the corrected P values of the SBS4 and SBS13 signatures were 0.046 and 0.079, respectively.

Fig. 3.

Fig. 3

Open in a new tab

Validation of prognostic abilities of SBS4 and SBS13 signatures in the targeted SCLC cohort and their tumor immunogenicity. (A) Kaplan-Meier survival curves stratified by the SBS4 signature status in the targeted cohort. (B) Multivariate Cox regression model of SBS4 signature was conducted with several clinical factors taken into consideration. (C) Kaplan-Meier survival curves stratified by the SBS13 signature status in the targeted cohort. (D) Multivariate Cox regression model of SBS13 signature was conducted with several clinical factors taken into consideration. (E) Associations of SBS4 signature with TMB in WES and targeted cohorts. (F) Associations of SBS13 signature with TMB in WES and targeted cohorts.

Previous studies have indicated that mutational signatures are often accompanied by genomic instability. Therefore, we explored the relationship of SBS4 and SBS13 with TMB. The results showed that the SBS4 signature, which was associated with poor prognosis, exhibited significantly reduced TMB levels in both the WES and targeted sequencing cohorts (Wilcoxon rank-sum test both P < 0.001; Fig. 3E). However, elevated TMB was observed to be enriched in SCLC patients carrying the SBS13 signature (Wilcoxon rank-sum test both P < 0.05; Fig. 3F).

The above findings suggest that SBS4 and SBS13 signatures have predictive values for the prognosis of SCLC, prompting us to consider whether these two mutational signatures play a role in the progression of NSCLC. Therefore, utilizing the TCGA NSCLC cohort, five mutational signatures were extracted (Figure S6 A), namely SBS4, SBS5, SBS6, SBS7b, and SBS13 (Figure S6B). Their mutational patterns and activities were presented in Figure S6C and Table S5, separately. Survival analyses indicated that NSCLC patients carrying SBS4 and SBS13 did not exhibit meaningful reductions or improvements in prognosis (Figure S6D and S6E).

Identification of a molecular subtype associated with SCLC inferior outcome

In the WES SCLC cohort, based on the mutational activity profile of the extracted five mutational signatures, we employed consensus clustering analysis to identify potential molecular subtypes. The clustering results showed that when the number of clusters was three, the maximum slope of the straight line was observed, suggesting the possibility of three potential SCLC subtypes (Fig. 4A). The consensus clustering bar plot also demonstrated that the three clusters showed better consistency when the number of clusters was three (Fig. 4B). Further survival analysis revealed that, compared to the other two subtypes, one subtype (cluster 2) exhibited the worst prognosis (Log-rank test P = 0.039; Fig. 4C), and the correlation between this subtype and the worst outcome remained statistically significant after adjusting for clinical confounding factors (HR: 1.63, 95% CI: 1.01–2.36, P = 0.029; Fig. 4D). Genomic characterization analysis showed that patients in this subtype also exhibited the lowest TMB (Wilcoxon rank-sum test P < 0.001; Fig. 4E).

Fig. 4.

Fig. 4

Open in a new tab

Identification of a molecular subtype with poorer SCLC outcome. (A) Association between clustering effects and clustering numbers. (B) Clustering consensus with clustering number selecting from 2 to 10. (C) Survival analysis of identified three SCLC molecular subtypes in the WES cohort. (D) Multivariate Cox adjusted analysis was performed to confirm the link between subtypes and prognosis. (E) Distinct TMB levels across three molecular subtypes in the WES cohort. (F) Survival analysis of identified three SCLC molecular subtypes in the targeted cohort. (G) Multivariate Cox adjusted analysis was performed to confirm the association. (H) Distinct TMB levels across three molecular subtypes in the targeted cohort.

In the targeted sequencing cohort, the clustering results still suggested the potential existence of three SCLC subtypes (Figure S7 A). The clustering tracking plot was presented in Figure S7B, and the consensus clustering bar plot was shown in Figure S7 C. Consistent with the results from the WES cohort, the cluster 2 subtype in the targeted sequencing cohort also exhibited the worst survival outcome (Log-rank test P = 0.002; Fig. 4F) and was validated by a multivariate Cox regression model (HR: 2.05, 95% CI: 1.22–3.46, P = 0.007; Fig. 4G). Similarly, the lowest TMB level was observed to be enriched in this subtype (Wilcoxon rank-sum test P < 0.001; Fig. 4H). In both the WES and targeted sequencing SCLC cohorts, we incorporated clinical factors, identified three clusters and TMB into the multivariate Cox regression models to obtain more accurate results. The results indicated that the cluster 2 still exhibited the poorer prognoses (Figure S8 A-S8B).

Specific gene mutations connected with SCLC outcome

Considering that targeted sequencing only covers more than 400 genes, in order to explore the gene mutations related to SCLC prognosis comprehensively, we performed whole-genome mutation analysis on the WES cohort. Firstly, we conducted univariate Cox regression analysis and multivariate Cox regression analysis corrected for clinical confounding factors on all mutated genes. Subsequently, we screened for genes with both statistical significances and a mutation frequency greater than 5%. Finally, 38 high-frequency mutated SCLC prognostic genes were identified (Table S6) and their mutational patterns were illustrated with a waterfall plot (Fig. 5A). Surprisingly, among all 38 genes, only UNC13 A mutation was associated with a better outcome in SCLC, while mutations in other genes showed poorer prognosis.

Fig. 5.

Fig. 5

Open in a new tab

Frequently mutated genes linked with SCLC outcome. (A) Waterfall plot exhibition of detailed mutational patterns of identified frequently mutated genes connected with SCLC outcomes. Green gene mutations were linked with inferior outcome and red gene mutations were linked with favorable outcome. (B) Co-mutation or mutual exclusion between prognostic mutated genes. (C) Kaplan-Meier survival analysis of co-mutations of NAALAD2 and OR14I1.

We further explored the co-mutation and mutual exclusion of the above 38 prognostic genes in the SCLC regulatory network. The results showed that there were 48 pairs of genes with meaningful co-mutations, while no significant mutually exclusive gene mutations were found (Fig. 5B and Table S7). Among all co-mutations, the co-mutation of NAALAD2 and OR14I1 was the most significant, and co-mutated SCLC patients harbored an inferior survival outcome (Log-rank test P = 0.008; Fig. 5C). Subsequent multivariate Cox regression model with clinical confounding factors and TMB taken into account still obtained this association (HR: 2.35, 95% CI: 1.28–5.19, P = 0.016; Fig. 5C).

Immune infiltration and tumor immunogenicity behind UNC13 A mutation

Previous results indicated that most genetic mutations were associated with poor outcomes in SCLC. However, patients with UNC13 A mutations exhibited significantly improved survival outcomes in both univariate survival analysis (Log-rank test P = 0.039; Fig. 6A) and multivariate Cox regression model (HR: 0.36, 95% CI: 0.16–0.81, P = 0.013; Fig. 6B). Recent studies have reported that this genetic mutation was related to immune regulation26,27, prompting us to further explore its underlying immunological molecular characteristics. Through the ssGSEA algorithm, we evaluated the infiltration abundance of 28 types of immune cells (Fig. 6C) and found that in the subgroup with UNC13 A mutations, the abundance of activated/memory CD8 T cells and dendritic cells was significantly increased, while the infiltration abundance of immunosuppressive cells, such as regulatory T cells and plasmacytoid dendritic cells, was reduced (all P < 0.05).

Fig. 6.

Fig. 6

Open in a new tab

Immune infiltration and pathway enrichment of UNC13 A mutations. (A) Kaplan-Meier survival curves stratified by wild-type and mutated UNC13 A. (B) Multivariate Cox regression model of UNC13 A mutations with clinical confounding factors taken into account. (C) Heatmap representation of infiltration abundance differences of 28 immunocyte subtypes between UNC13 A two subgroups. Immunocytes highlighted with red were significantly enriched in UNC13 A mutated patients, whereas green immunocytes were enriched in UNC13 A wild-type patients. GSEA analyses of UNC13 A mutations based on the signaling pathways from (D) KEGG and (E) GO BP databases. Pathways highlighted with purple were immune response relevant. (F) Distinct TMB levels between UNC13 A wild-type and mutated groups. (G) T cell-inflamed signature, (H) cytolytic activity, and (I) WNT TGFβ signature enrichment scores between UNC13 A wild-type and mutated groups.

The GSEA results revealed that immune activation signaling pathways, such as graft versus host disease, antigen processing presentation, and cytokine receptor interaction in KEGG (Fig. 6D), and immune effector process in GO BP (Fig. 6E), were significantly enriched in SCLC patients carrying UNC13 A mutations. Furthermore, results based on the HALLMARK and REACTOME databases further validated the association between UNC13 A mutations and activated immune pathways (Figure S9 A and S9B).

A significantly elevated TMB level was also observed in the UNC13 A mutation group (Fig. 6F). The enrichment scores of immunogenic-related molecular signatures, T cell-inflamed signature and cytolytic activity, were significantly increased in the UNC13 A mutant group (Fig. 6G and H), while the immunosuppressive WNT TGFβ signature was distant from this group (Fig. 6I).

TP53 splice site mutation involved in worse SCLC outcome

TP53, as a known driver gene for SCLC, exhibited a very high mutation rate, and its mutations could lead to shortened median survival for SCLC patients (Fig. 7A and B). In this work, we analyzed in detail the impact of distinct TP53 mutation types on the prognosis of SCLC. The results indicated that, compared to other mutation types, TP53 splice site mutations resulted in a worse survival outcome (Log-rank test P = 0.072; Fig. 7C), and this association remained consistent even after including factors such as age, gender, stage, ethnicity, and sample type in a multivariate Cox model (HR: 4.20, 95% CI: 1.08–16.30, P = 0.043; Fig. 7D).

Fig. 7.

Fig. 7

Open in a new tab

Association of TP53 splice mutations with SCLC outcome. (A) Survival analysis of SCLC patients with and without TP53 mutations. (B) Multivariate Cox regression model of TP53 mutations. (C) Kaplan-Meier survival curves stratified by TP53 wild-type, splice mutation, and non-splice mutation subgroups. (D) Multivariate Cox regression model of TP53 three subgroups.

Discussion

The prognosis of SCLC is extremely poor, and there are no effective molecular markers for predicting prognosis. The emergence of immunotherapy has provided new possibilities for improving the survival of tumor patients. This study explored the mutational signatures, molecular subtypes, and frequently mutated genes of SCLC from the perspective of somatic mutations, and assessed their relationships with outcomes and immune molecular characteristics, providing theoretical basis for clinical prognosis assessment and improvement of treatment strategies.

Mutational signatures play a significant role in assessing tumor prognosis and predicting the efficacy of immunotherapy. Smoking-related SBS4 signature has been proven to be associated with a better response to ICI treatment in NSCLC28, whereas in melanoma, the presence of this signature led to a poorer ICI outcome9, suggesting the heterogeneity of tumor treatment. Currently, no studies have reported the relationship between this mutational signature and tumor prognosis. In this work, we found that SCLC patients carrying SBS4 harbored an inferior survival outcome and exhibited lower TMB, which is a promising biomarker for predicting immunotherapy benefits29. These results suggest that SBS4 signature may mediate a poor outcome in SCLC and may be associated with immunotherapy resistance.

The SBS13 mutational signature is widely present in various tumors and associated with better anti-PD1 treatment efficacy30,31. Currently, no studies have indicated its correlation with SCLC prognosis. We observed that the SBS13 signature was linked with favorable survival outcomes in SCLC patients. Subsequent genomic characterization analysis revealed that this signature also resulted in an elevated TMB. Therefore, we hypothesize that SCLC patients carrying the SBS13 mutation signature may achieve better response rates and prognosis when receiving immunotherapy.

In our study, we were cognizant of such factors and took steps to mitigate their impact, such as incorporating multiple clinical variables (age, gender, stage, ethnicity, and sample type) into multivariate Cox regression models to isolate the true effect of the SBS13 signature on prognosis. Nevertheless, several potential confounders remain unaccounted for. Treatment regimens vary among patients, with different combinations of chemotherapy, radiotherapy, and immunotherapy, and patients with the SBS13 signature might have received more effective treatments by chance, thus confounding the relationship. Additionally, undetected co-mutations or genetic alterations could co-occur with the SBS13 signature and influence tumor biology, immune response, and patient survival. Lifestyle factors like smoking history, beyond its link to the SBS4 signature, might also play a role, as patients with and without the SBS13 signature could have different smoking habits or carcinogen exposures. Future studies could address these issues through more comprehensive data collection, including detailed treatment information, in-depth genetic analysis using advanced sequencing, and better lifestyle factor characterization to precisely define the relationship between the SBS13 mutational signature and SCLC prognosis.

The principle of precision individualized therapy drives us to search for more subdivided molecular subtypes of SCLC. Relying on mutation signature activity and consensus clustering, we identified three molecular subtypes of SCLC in the WES and targeted sequencing cohorts, and survival analysis indicated that one subtype had the worst prognosis. Further analysis of the subtype with the worst prognosis revealed that it also harbored the lowest TMB level. Based on mutation information, we discovered a potential molecular subtype of SCLC that may be associated with poor prognosis and treatment response.

Most genomic mutations can contribute to poor prognosis, but in this study, we found that UNC13 A mutations were associated with survival benefit in SCLC. Further immunological characterization revealed that patients with UNC13 A mutations exhibited activated immune microenvironment and immune response pathways, resulting in enhanced immunogenicity. Recent studies have demonstrated close relationships of single-gene mutations with high immunogenicity and immunotherapy efficacy, such as FAT132, TP5333, MUC1634, and POLE35. Therefore, UNC13 A mutations may serve as a potential indicator of response to immunotherapy in SCLC. Prospective cohorts are needed to confirm these findings.

In summary, this study identified several molecular markers related to SCLC prognosis by integrating somatic mutation profiles, mRNA transcriptome data, and clinical information. At the same time, we clarified underlying molecular immunological characteristics and provided clues for the development of SCLC clinical trials and the updating of treatment methods.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1 (2.2MB, docx)
Supplementary Material 2 (601.5KB, xlsx)

Author contributions

QW, JG, and YR designed this study; YL, CS, HW, WD, YH, and YC collected and integrated the related data; YL, HW, PL, and JC conducted main data analysis; YL, JG, and QW composed and corrected the manuscript.

Funding

This research was funded by the Provincial Natural Science Foundation of Henan Province (No. 252300421637) and Provincial Natural Science Foundation of Shandong Province (No. ZR2022MH127).

Data availability

The datasets analyzed during the current study are available in the cBioPortal repository [http://www.cbioportal.org/study/summary? id=msk_met_2021] and recently published two studies [doi:10.1038/nature14664 and 10.1038/ng.2405].

Declarations

Competing interests

The authors declare no competing interests.

Footnotes

This article has been retracted. Please see the retraction notice for more detail: https://doi.org/10.1038/s41598-026-40190-6

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Yuting Li and Chen Song authors contributed equally to this work.

Change history

2/18/2026

This article has been retracted. Please see the Retraction Notice for more detail: 10.1038/s41598-026-40190-6

Contributor Information

Yanfeng Ren, Email: renyf@sdsmu.edu.cn.

Jing Gong, Email: gongjing1990@163.com.

Qinghua Wang, Email: wangqinghua@sdsmu.edu.cn.

References

  • 1.Megyesfalvi, Z. et al. Clinical insights into small cell lung cancer: tumor heterogeneity, diagnosis, therapy, and future directions. CA Cancer J. Clin.73, 620–652. 10.3322/caac.21785 (2023). [DOI] [PubMed] [Google Scholar]
  • 2.Lee, J. H., Saxena, A. & Giaccone, G. Advancements in small cell lung cancer. Semin. Cancer Biol.93, 123–128. 10.1016/j.semcancer.2023.05.008 (2023). [DOI] [PubMed] [Google Scholar]
  • 3.Meijer, J. J. et al. Small cell lung cancer: novel treatments beyond immunotherapy. Semin. Cancer Biol.86, 376–385. 10.1016/j.semcancer.2022.05.004 (2022). [DOI] [PubMed] [Google Scholar]
  • 4.Koh, G., Degasperi, A., Zou, X., Momen, S. & Nik-Zainal, S. Mutational signatures: emerging concepts, caveats and clinical applications. Nat. Rev. Cancer21, 619–637. 10.1038/s41568-021-00377-7 (2021). [DOI] [PubMed] [Google Scholar]
  • 5.Abbas, S. et al. Mutational signature dynamics shaping the evolution of oesophageal adenocarcinoma. Nat. Commun.14, 4239. 10.1038/s41467-023-39957-6 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Chen, H. et al. Age-related mutational signature negatively associated with immune activity and survival outcome in triple-negative breast cancer. Oncoimmunology9, 1788252. 10.1080/2162402X.2020.1788252 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Chong, W. et al. Association of clock-like mutational signature with immune checkpoint inhibitor outcome in patients with melanoma and NSCLC. Mol. Ther. Nucleic Acids23, 89–100. 10.1016/j.omtn.2020.10.033 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Li, X. C. et al. A mutational signature associated with alcohol consumption and prognostically significantly mutated driver genes in esophageal squamous cell carcinoma. Ann. Oncol.29, 938–944. 10.1093/annonc/mdy011 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Zhang, W. et al. Novel molecular determinants of response or resistance to immune checkpoint inhibitor therapies in melanoma. Front. Immunol.12, 798474. 10.3389/fimmu.2021.798474 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Shi, F. et al. Sex disparities of genomic determinants in response to immune checkpoint inhibitors in melanoma. Front. Immunol.12, 721409. 10.3389/fimmu.2021.721409 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.George, J. et al. Comprehensive genomic profiles of small cell lung cancer. Nature524, 47–53. 10.1038/nature14664 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Rudin, C. M. et al. Comprehensive genomic analysis identifies SOX2 as a frequently amplified gene in small-cell lung cancer. Nat. Genet.44, 1111–1116. 10.1038/ng.2405 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Nguyen, B. et al. Genomic characterization of metastatic patterns from prospective clinical sequencing of 25,000 patients. Cell185, 563–575. 10.1016/j.cell.2022.01.003 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Zehir, A. et al. Mutational landscape of metastatic cancer revealed from prospective clinical sequencing of 10,000 patients. Nat. Med.23, 703–713. 10.1038/nm.4333 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Kim, J. et al. Somatic ERCC2 mutations are associated with a distinct genomic signature in urothelial tumors. Nat. Genet.48, 600–606. 10.1038/ng.3557 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Tate, J. G. et al. COSMIC: the catalogue of somatic mutations in cancer. Nucleic Acids Res.47, D941–D947. 10.1093/nar/gky1015 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Alexandrov, L. B. et al. Signatures of mutational processes in human cancer. Nature500, 415–421. 10.1038/nature12477 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Diaz-Gay, M. et al. Assigning mutational signatures to individual samples and individual somatic mutations with sigprofilerassignment. Bioinformatics10.1093/bioinformatics/btad756 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Wilkerson, M. D. & Hayes, D. N. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics26, 1572–1573. 10.1093/bioinformatics/btq170 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Wu, T. et al. ClusterProfiler 4.0: A universal enrichment tool for interpreting omics data. Innovation2, 100141. 10.1016/j.xinn.2021.100141 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Hanzelmann, S., Castelo, R. & Guinney, J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinform.10.1186/1471-2105-14-7 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Charoentong, P. et al. Pan-cancer immunogenomic analyses reveal genotype-immunophenotype relationships and predictors of response to checkpoint Blockade. Cell. Rep.18, 248–262. 10.1016/j.celrep.2016.12.019 (2017). [DOI] [PubMed] [Google Scholar]
  • 23.Ayers, M. et al. IFN-gamma-related mRNA profile predicts clinical response to PD-1 blockade. J. Clin. Investig.127, 2930–2940. 10.1172/JCI91190 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Rooney, M. S., Shukla, S. A., Wu, C. J., Getz, G. & Hacohen, N. Molecular and genetic properties of tumors associated with local immune cytolytic activity. Cell160, 48–61. 10.1016/j.cell.2014.12.033 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Fuxe, J. & Karlsson, M. C. TGF-beta-induced epithelial-mesenchymal transition: a link between cancer and inflammation. Semin. Cancer Biol.22, 455–461. 10.1016/j.semcancer.2012.05.004 (2012). [DOI] [PubMed] [Google Scholar]
  • 26.Feng, P., Li, Z., Li, Y., Zhang, Y. & Miao, X. Characterization of different subtypes of immune cell infiltration in glioblastoma to aid immunotherapy. Front. Immunol.13, 799509. 10.3389/fimmu.2022.799509 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Ansari, U. et al. Role of the UNC13 family in human diseases: A literature review. AIMS Neurosci.10, 388–400. 10.3934/Neuroscience.2023029 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Rizvi, N. A. et al. Cancer immunology. Mutational landscape determines sensitivity to PD-1 Blockade in non-small cell lung cancer. Science348, 124–128. 10.1126/science.aaa1348 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Samstein, R. M. et al. Tumor mutational load predicts survival after immunotherapy across multiple cancer types. Nat. Genet.51, 202–206. 10.1038/s41588-018-0312-8 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Wang, S., Jia, M., He, Z. & Liu, X. S. APOBEC3B and APOBEC mutational signature as potential predictive markers for immunotherapy response in non-small cell lung cancer. Oncogene37, 3924–3936. 10.1038/s41388-018-0245-9 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Miao, D. et al. Genomic correlates of response to immune checkpoint Blockade in microsatellite-stable solid tumors. Nat. Genet.50, 1271–1281. 10.1038/s41588-018-0200-2 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Zhang, W. et al. Favorable immune checkpoint inhibitor outcome of patients with melanoma and NSCLC harboring FAT1 mutations. NPJ Precis. Oncol.6, 46. 10.1038/s41698-022-00292-6 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Dong, Z. Y. et al. Potential predictive value of TP53 and KRAS mutation status for response to PD-1 Blockade immunotherapy in lung adenocarcinoma. Clin. Cancer Res.23, 3012–3024. 10.1158/1078-0432.CCR-16-2554 (2017). [DOI] [PubMed] [Google Scholar]
  • 34.Li, X., Pasche, B., Zhang, W. & Chen, K. Association of MUC16 mutation with tumor mutation load and outcomes in patients with gastric cancer. JAMA Oncol.4, 1691–1698. 10.1001/jamaoncol.2018.2805 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Ma, X., Dong, L., Liu, X., Ou, K. & Yang, L. POLE/POLD1 mutation and tumor immunotherapy. J. Exp. Clin. Cancer Res.41, 216. 10.1186/s13046-022-02422-1 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1 (2.2MB, docx)
Supplementary Material 2 (601.5KB, xlsx)

Data Availability Statement

The datasets analyzed during the current study are available in the cBioPortal repository [http://www.cbioportal.org/study/summary? id=msk_met_2021] and recently published two studies [doi:10.1038/nature14664 and 10.1038/ng.2405].

Articles from Scientific Reports are provided here courtesy of Nature Publishing Group

RESOURCES