Clinicopathological characteristics and prognosis of pulmonary spindle cell carcinoma: a population-based retrospective study using SEER data

Introduction

Pulmonary spindle cell carcinoma (PSCC) is a rare type of pulmonary sarcomatoid carcinoma (PSC) (1). Its incidence accounts for only 0.17–0.4% of all lung cancers (2,3). PSC includes several histological entities, such as pleomorphic carcinoma, spindle cell carcinoma, giant cell carcinoma, carcinosarcoma, and pulmonary blastoma (4). This study focused on PSCC because it is a pathologically distinct and particularly rare spindle-cell-predominant subtype, and because pooling all PSC subtypes could obscure subtype-specific clinicopathological patterns. The pathological specimens of PSCC consist of only spindle cells, with no giant cells or pleomorphic components observed. The cells adhere in nested, swirled or irregular bundles and are arranged densely, with deep stained nuclei. The growth pattern is sarcomatoid. Focal and scattered lymphoplasmic cells can be seen and infiltrate in or around the tumor (4-6). Due to its low incidence, there are few studies on the pathological characteristics of patients with PSCC and the factors affecting survival and prognosis. This retrospective study used the population-level sample size of the Surveillance, Epidemiology, and End Results (SEER) database to describe the clinicopathological characteristics and survival-associated factors of patients with PSCC. Because the present analysis was designed around PSCC, it did not attempt to provide a comprehensive analysis of other PSC subtypes. We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0924/rc).

Methods

Data source and study population

This population-based retrospective study used data from the SEER database. Access to the SEER data was obtained under the SEER data-use agreement, and case listings were exported using SEER*Stat version 8.3.6. Patients diagnosed between 2004 and 2014 were screened from the restored SEER-derived workbooks.

The original PSCC source cohort contained International Classification of Diseases for Oncology, Third Edition (ICD-O-3) histology codes 8032/3 and 8004/3. Because ICD-O-3 code 8004/3 may include undifferentiated malignant neoplasms and does not consistently confirm epithelial spindle cell carcinoma, the primary reanalysis was restricted to ICD-O-3 code 8032/3. The 15 patients coded as 8004/3 were excluded. Eligible PSCC patients were older than 20 years, had a single primary pulmonary malignancy, had a definitive histopathological diagnosis, and had complete tumor-node-metastasis (TNM) information. For baseline comparison, eligible patients in the processed non-small cell lung cancer (NSCLC) + squamous cell carcinoma (SCC) comparison workbook who were not coded as 8032/3 or 8004/3 were classified as the other NSCLC comparator cohort.

The American Joint Committee on Cancer (AJCC) 8th edition stage group was taken from the processed staging variable supplied in the restored comparison workbook. The accompanying data-cleaning note documented conversion from AJCC 6th/7th edition variables to AJCC 8th edition categories, and a staging traceability summary is provided as Appendix 1. Because the analysis variable used for Cox modeling grouped T2 and T3 together, the final AJCC stage group was not re-derived from the collapsed modeling variables in this revision. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Variables and outcomes

The clinicopathological variables included age at diagnosis, sex, race, tumor differentiation, primary site, histological diagnosis, tumor size, T stage, N stage, M stage, AJCC 8th edition stage group, surgical resection, type of surgery, radiotherapy, and chemotherapy. Treatment variables were analyzed as registry-recorded treatment receipt and were not interpreted as randomized treatment assignments.

The primary outcome was cancer-specific survival (CSS), defined as the interval from diagnosis to death attributed to the tumor in SEER. Overall survival (OS), defined as the interval from diagnosis to death from any cause, was analyzed as a secondary outcome. Because SEER cause-specific survival depends on registry-processed cause-of-death information derived from death certificates, CSS may be affected by outcome misclassification, particularly in elderly or advanced-stage lung cancer populations; therefore, OS was reported alongside CSS.

Statistical analysis

Continuous variables were summarized as medians with interquartile ranges (IQRs), and categorical variables were summarized as frequencies and percentages. Baseline differences between strict ICD-O-3 8032/3 PSCC and other NSCLC patients were assessed using the Mann-Whitney U test for continuous variables and the chi-square test or Fisher’s exact test for categorical variables, as appropriate. CSS and OS were estimated using the Kaplan-Meier method, and survival curves were compared using log-rank tests. All tests were two-sided, and P<0.05 was considered statistically significant.

Associations with CSS and OS were evaluated using Cox proportional hazards regression models and reported as hazard ratios (HRs) with 95% confidence intervals (CIs). The multivariable model was rebuilt using the variables included in the original prognostic model: sex, T stage, N stage, M stage, surgery, and chemotherapy. Model performance was summarized using the concordance index (C-index), calibration by predicted-risk quartiles, and an internally derived point-scale plot.

To address reviewer concerns about other PSC subtypes, PSC-related histology codes were audited in the raw extraction and processed comparison workbook. Descriptive subtype summaries were limited to histology codes retained in the processed comparison workbook; raw-only subtype availability is summarized in Table S1. All analyses were performed using Python with pandas, scipy, matplotlib, and lifelines.

Results

Patient characteristics

A total of 220,953 eligible patients were included in the reanalysis, including 250 strict ICD-O-3 8032/3 PSCC patients and 220,703 other NSCLC patients. The baseline characteristics are summarized in Table 1. Among strict PSCC patients, 134 (53.6%) were male and 116 (46.4%) were female, with a median age of 72 years (IQR, 63–80 years). The median tumor diameter was 5.7 cm in strict PSCC and 4.0 cm in other NSCLC patients (P<0.001). The most common primary site was the upper lobe (111 cases, 50.5% among patients with known primary site), followed by the lower lobe (77 cases, 35.0%) and middle lobe (18 cases, 8.2%). Compared with other NSCLC patients, strict PSCC patients had less differentiated tumors (P<0.001), larger tumor size (P<0.001), and a higher proportion of T4 disease (53.2% vs. 41.1%; P<0.001 for T-stage distribution). AJCC 8th edition stage distribution did not differ significantly between strict PSCC and other NSCLC patients (P=0.223).

Treatment patterns and survival outcomes

Radiotherapy was not associated with a significant CSS or OS difference among patients with strict PSCC. In stage-specific analyses, chemotherapy was not associated with a significant CSS difference in stage I or II patients (Figure 1A,1B), whereas chemotherapy was associated with longer CSS in stage III and IV patients (Figure 1C,1D). A similar pattern was observed for OS, with significant differences in stage III and IV patients but not in stage I or II patients. Surgical treatment was recorded in 68 strict PSCC patients (27.2%).

Figure 1 Kaplan-Meier estimates of cancer-specific survival (CSS) according to chemotherapy status across AJCC stages in the strict ICD-O-3 8032/3 cohort. (A) Stage I patients (P=0.47). (B) Stage II patients (P=0.78). (C) Stage III patients (P=0.01). (D) Stage IV patients (P<0.001). P values were calculated using the log-rank test. AJCC, American Joint Committee on Cancer; CSS, cancer-specific survival; ICD-O-3, International Classification of Diseases for Oncology, Third Edition.

In the strict ICD-O-3 8032/3 cohort, median CSS was 5 months, and the 1-year, 3-year, 5-year, and 10-year CSS rates were 29.2%, 17.6%, 16.3%, and 14.7%, respectively. Median OS was 4 months, and the 1-year, 3-year, 5-year, and 10-year OS rates were 27.1%, 14.8%, 13.2%, and 8.5%, respectively. In other NSCLC patients, median CSS was 12 months and median OS was 10 months. The corresponding 1-year, 3-year, 5-year, and 10-year CSS rates for other NSCLC patients were 48.6%, 27.6%, 21.4%, and 16.2%, respectively. During follow-up, 221 of 250 strict PSCC patients died from any cause, and 198 deaths were classified as attributable to this cancer diagnosis. The Kaplan-Meier curves are shown for stage-specific chemotherapy groups in Figure 1, PSCC versus other NSCLC in Figure 2, AJCC stage in Figure 3, surgery status in Figure 4, and chemotherapy status in Figure 5.

Figure 2 Kaplan-Meier estimates of (A) CSS and (B) OS comparing strict ICD-O-3 8032/3 pulmonary spindle cell carcinoma patients with other NSCLC patients. CSS, cancer-specific survival; ICD-O-3, International Classification of Diseases for Oncology, Third Edition; NSCLC, non-small cell lung cancer; OS, overall survival; PSCC, pulmonary spindle cell carcinoma.

Figure 3 Kaplan-Meier estimates of (A) CSS and (B) OS in strict ICD-O-3 8032/3 pulmonary spindle cell carcinoma patients stratified by AJCC 8^th edition stage. AJCC, American Joint Committee on Cancer; CSS, cancer-specific survival; ICD-O-3, International Classification of Diseases for Oncology, Third Edition; OS, overall survival; PSCC, pulmonary spindle cell carcinoma.

Figure 4 Kaplan-Meier estimates of (A) CSS and (B) OS in strict ICD-O-3 8032/3 pulmonary spindle cell carcinoma patients according to surgery status. CSS, cancer-specific survival; ICD-O-3, International Classification of Diseases for Oncology, Third Edition; OS, overall survival; PSCC, pulmonary spindle cell carcinoma.

Figure 5 Kaplan-Meier estimates of (A) CSS and (B) OS in strict ICD-O-3 8032/3 pulmonary spindle cell carcinoma patients according to chemotherapy status. CSS, cancer-specific survival; ICD-O-3, International Classification of Diseases for Oncology, Third Edition; OS, overall survival; PSCC, pulmonary spindle cell carcinoma.

CSS and OS differed significantly by AJCC stage (both P<0.001). For CSS, median survival was not reached in stage I patients, 22 months in stage II patients, 5 months in stage III patients, and 2 months in stage IV patients. Patients who underwent surgery had longer CSS and OS than those who did not undergo surgery (both P<0.001), with median CSS of 21 versus 3 months and median OS of 19 versus 3 months, respectively. Overall chemotherapy groups did not differ significantly by unadjusted CSS log-rank testing (P=0.170), although stage-specific analyses showed significant CSS differences in stage III (P=0.011) and stage IV (P<0.001) patients. The stage-specific chemotherapy analyses were underpowered for stage I and II patients, with only 4 stage I and 10 stage II chemotherapy-treated cases. Stage-specific surgery subgroup analyses were not emphasized because only 68 strict PSCC patients underwent surgery and because surgery assignment is strongly affected by resectability, physiologic reserve, and selection bias.

Prognostic factors and model performance

The Cox proportional hazards regression results are shown in Table 2 for CSS and Table 3 for OS. In the multivariable CSS model, female sex (HR 0.739, 95% CI: 0.551–0.992; P=0.044), T2–T3 stage (HR 1.969, 95% CI: 1.042–3.718; P=0.037), T4 stage (HR 3.132, 95% CI: 1.644–5.968; P<0.001), N2 stage (HR 1.694, 95% CI: 1.177–2.437; P=0.005), M1 stage (HR 2.586, 95% CI: 1.840–3.636; P<0.001), surgery (HR 0.505, 95% CI: 0.343–0.744; P<0.001), and chemotherapy (HR 0.375, 95% CI: 0.269–0.522; P<0.001) were associated with CSS. The CSS model C-index was 0.757. In the multivariable OS model, female sex, higher T stage, N2 stage, M1 stage, surgery, and chemotherapy were also associated with OS, and the OS model C-index was 0.761. These treatment-related associations should not be interpreted as causal treatment effects because treatment allocation was not randomized. The internally derived CSS point-scale model and calibration plots are shown in Figures 6,7.

Figure 6 Calibration plot for the internally derived CSS prognostic model in the strict ICD-O-3 8032/3 cohort. CSS, cancer-specific survival; ICD-O-3, International Classification of Diseases for Oncology, Third Edition.

Figure 7 Point-scale visualization of the internally derived CSS Cox prognostic model in the strict ICD-O-3 8032/3 cohort. CSS, cancer-specific survival; ICD-O-3, International Classification of Diseases for Oncology, Third Edition; M, metastasis; N, node; T, tumor.

Context for other PSC subtypes

To provide context for other PSC subtypes, PSC-related histology codes were audited in the restored data. In the processed comparison workbook, pleomorphic carcinoma (ICD-O-3 8022/3, n=237) and giant cell carcinoma (ICD-O-3 8031/3, n=216) were retained alongside strict PSCC (ICD-O-3 8032/3, n=250). Carcinosarcoma (ICD-O-3 8980/3) and pulmonary blastoma (ICD-O-3 8972/3) were present in the raw extraction but were not retained in the processed NSCLC + SCC comparison workbook used for the current survival analyses. Therefore, Table 4 and Table S1 provide the available raw and processed subtype audit, while inferential analyses remain focused on strict PSCC versus other NSCLC.

Discussion

The present reanalysis restricted the PSCC cohort to ICD-O-3 code 8032/3 and excluded 15 patients coded as 8004/3. Even under this stricter disease definition, PSCC remained rare and aggressive. The strict cohort represented approximately 0.11% of eligible NSCLC patients in the processed comparison dataset. The median age at diagnosis was 72 years, 53.6% of patients were male, and nearly half of patients had stage IV disease. These findings are broadly consistent with prior reports that PSCC usually occurs in older adults and is often diagnosed at an advanced stage, although sex distribution may vary across small series and registry-based studies. Female sex was associated with longer CSS and OS in multivariable analysis. Prior studies have suggested that PSC may be linked to molecular abnormalities such as TP53 alterations (7-11), but the SEER database does not contain molecular data; therefore, the biological explanation for the observed sex association remains speculative.

The strict PSCC cohort had larger tumors than other NSCLC patients, with a median tumor diameter of 5.7 versus 4.0 cm. Tumors were most commonly located in the upper lobe and were usually poorly differentiated or undifferentiated. These clinicopathological features may reflect delayed detection, rapid growth, and the aggressive biology of pulmonary sarcomatoid tumors (12-17). Radiotherapy was not associated with a significant CSS or OS difference in this strict cohort. Surgery was associated with longer CSS and OS, but this association should be interpreted cautiously. Surgical selection is strongly influenced by resectability, comorbidity, cardiopulmonary reserve, performance status (18), and clinician judgment, none of which can be fully captured in SEER. Therefore, the observed association should not be interpreted as proof that surgery caused improved survival.

Chemotherapy was associated with longer CSS and OS in the multivariable models, and stage-specific analyses showed significant CSS and OS differences among stage III and IV patients. However, these findings also require caution. The stage I and II chemotherapy subgroups were very small, and treatment assignment in this retrospective cohort was not randomized. Patients who received chemotherapy likely differed from untreated patients in baseline health, tumor burden, treatment tolerance, and clinician assessment. In addition, SEER does not record chemotherapy regimen, dose, line of therapy, response, targeted therapy, or immunotherapy details. Recent PSC evidence also underscores that clinicogenomic features, immune checkpoint inhibitor exposure, and TP53 mutation status are now clinically relevant in PSC (19,20), but those variables were unavailable in this historical SEER cohort. The present results therefore support only an association between chemotherapy receipt and survival in selected patients, not a definitive treatment recommendation.

This study has several limitations. First, although the main analysis was rebuilt using the strict ICD-O-3 8032/3 definition, the analysis remains dependent on registry coding and processed SEER variables. Second, CSS in SEER is derived from cause-of-death information and may be misclassified, especially in elderly or advanced-stage lung cancer patients; for this reason, OS was added alongside CSS. Third, key clinical confounders were unavailable, including smoking history, comorbidities, performance status (18), pulmonary function, imaging findings, recurrence, metastatic burden or sites, detailed radiotherapy and chemotherapy regimens, molecular alterations, and gene mutations. Fourth, the cohort spanned 2004 to 2014 and therefore largely predates contemporary targeted therapy and immune checkpoint inhibitors (19,20). Finally, the prognostic model was internally derived and internally evaluated only. The C-index and calibration results describe apparent performance within the available dataset, but external validation using independent clinical data is required before the model can be considered reliable for individualized clinical prediction.

Conclusions

Strictly defined ICD-O-3 8032/3 PSCC has distinctive clinicopathological characteristics and poor CSS and OS. Sex, T stage, N stage, M stage, surgery, and chemotherapy were associated with survival in this retrospective SEER cohort. Treatment-related findings should be interpreted as associations rather than causal evidence of treatment benefit. The internally derived CSS prognostic model showed exploratory predictive performance but requires external validation. Multicenter prospective registry studies with detailed clinical, pathological, therapeutic, and molecular information are needed to validate these findings and guide more precise treatment strategies for patients with PSCC.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0924/rc

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0924/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0924/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Maggiore C, Mulè A, Fadda G, et al. Histological classification of lung cancer. Rays 2004;29:353-5.
2. Rossi G, Cavazza A, Sturm N, et al. Pulmonary carcinomas with pleomorphic, sarcomatoid, or sarcomatous elements: a clinicopathologic and immunohistochemical study of 75 cases. Am J Surg Pathol 2003;27:311-24. [Crossref] [PubMed]
3. Terada T. Spindle cell carcinoma of the lung: frequency, clinical features, and immunohistochemical studies of three cases. Respir Med CME 2010;3:241-5.
4. Travis WD, Brambilla E, Nicholson AG, et al. The 2015 World Health Organization Classification of Lung Tumors: Impact of Genetic, Clinical and Radiologic Advances Since the 2004 Classification. J Thorac Oncol 2015;10:1243-60. [Crossref] [PubMed]
5. Pelosi G, Melotti F, Cavazza A, et al. A modified vimentin histological score helps recognize pulmonary sarcomatoid carcinoma in small biopsy samples. Anticancer Res 2012;32:1463-73.
6. Tamaki T, Shimizu T, Niki M, et al. Immunohistochemical analysis of NANOG expression and epithelial-mesenchymal transition in pulmonary sarcomatoid carcinoma. Oncol Lett 2017;13:3695-702. [Crossref] [PubMed]
7. Mainwaring MG, Poor C, Zander DS, et al. Complete remission of pulmonary spindle cell carcinoma after treatment with oral germanium sesquioxide. Chest 2000;117:591-3. [Crossref] [PubMed]
8. Kontic M, Stojsic J, Stevic R, et al. Could spindle cell lung carcinoma be considered and treated as sarcoma, according to its clinical course, morphology, immunophenotype and genetic finding? Pathol Oncol Res 2013;19:129-33. [Crossref] [PubMed]
9. Wang Y, Wang Z, Piha-Paul S, et al. Outcome analysis of Phase I trial patients with metastatic KRAS and/or TP53 mutant non-small cell lung cancer. Oncotarget 2018;9:33258-70. [Crossref] [PubMed]
10. Canale M, Petracci E, Delmonte A, et al. Impact of TP53 Mutations on Outcome in EGFR-Mutated Patients Treated with First-Line Tyrosine Kinase Inhibitors. Clin Cancer Res 2017;23:2195-202. [Crossref] [PubMed]
11. Labbé C, Cabanero M, Korpanty GJ, et al. Prognostic and predictive effects of TP53 co-mutation in patients with EGFR-mutated non-small cell lung cancer (NSCLC). Lung Cancer 2017;111:23-9. [Crossref] [PubMed]
12. Kim TH, Kim SJ, Ryu YH, et al. Pleomorphic carcinoma of lung: comparison of CT features and pathologic findings. Radiology 2004;232:554-9. [Crossref] [PubMed]
13. Venissac N, Pop D, Lassalle S, et al. Sarcomatoid lung cancer (spindle/giant cells): an aggressive disease? J Thorac Cardiovasc Surg 2007;134:619-23. [Crossref] [PubMed]
14. Choi HS, Seol H, Heo IY, et al. Fine-needle aspiration cytology of pleomorphic carcinomas of the lung. Korean J Pathol 2012;46:576-82. [Crossref] [PubMed]
15. Terra SB, Aubry MC, Yi ES, et al. Immunohistochemical study of 36 cases of pulmonary sarcomatoid carcinoma--sensitivity of TTF-1 is superior to napsin. Hum Pathol 2014;45:294-302. [Crossref] [PubMed]
16. Chang YL, Lee YC, Shih JY, et al. Pulmonary pleomorphic (spindle) cell carcinoma: peculiar clinicopathologic manifestations different from ordinary non-small cell carcinoma. Lung Cancer 2001;34:91-7. [Crossref] [PubMed]
17. Karim NA, Schuster J, Eldessouki I, et al. Pulmonary sarcomatoid carcinoma: University of Cincinnati experience. Oncotarget 2018;9:4102-8. [Crossref] [PubMed]
18. Gajra A, Marr AS, Ganti AK. Management of patients with lung cancer and poor performance status. J Natl Compr Canc Netw 2014;12:1015-25. [Crossref] [PubMed]
19. Hong L, Di Federico A, Liu B, et al. Distinct Clinicogenomic Features and Immunotherapy Associations in Pulmonary Sarcomatoid Carcinoma: A Multicenter Retrospective Study. J Thorac Oncol 2025;20:1763-77. [Crossref] [PubMed]
20. Wu M, Zhou N, Guan M, et al. Efficacy of immune checkpoint inhibitors in pulmonary sarcomatoid carcinoma and predictive potential of mutated TP53. Lung Cancer 2025;199:108068. [Crossref] [PubMed]