A recurrence risk prediction model and recurrence patterns in pulmonary lymphoepithelial carcinoma based on clinical and dynamic hematologic parameters

Introduction

Pulmonary lymphoepithelial carcinoma (pLELC) is a rare subtype of non-small cell lung cancer (NSCLC), accounting for approximately 1% of newly diagnosed lung cancers (1). Closely associated with Epstein-Barr virus (EBV) infection, it predominantly affects young, non-smoking individuals, with higher incidence in East and Southeast Asian populations (2). Classified as squamous cell carcinoma in the 2021 World Health Organization (WHO) classification of lung tumors (3), pLELC is characterized by poorly differentiated or undifferentiated tumor cells with nest-like or diffuse syncytial growth, oval to round vesicular nuclei with prominent nucleoli, and variable lymphocytic infiltration (Figure S1). Positivity for EBV-encoded RNA (EBER) by in situ hybridization is a key diagnostic criterion (4,5). Cytokeratin staining (CK5/6, p40, and p63 positive) helps differentiate pLELC from lymphomas (non-epithelial tumors), which are cytokeratin-negative and CD45 (leukocyte common antigen)-positive. Radiologically, pLELC typically presents on chest computed tomography (CT) as a well-defined central or perihilar solid mass (6), with relatively high 18F-fluorodeoxyglucose (FDG) uptake on positron emission tomography-computed tomography (PET-CT), indicating moderate-to-high glycolytic activity (7).

For patients with resectable pLELC, radical surgical resection remains the cornerstone treatment. However, postoperative recurrence continues to significantly impact long-term survival. Overall, pLELC carries a more favorable prognosis compared to conventional lung adenocarcinoma or squamous cell carcinoma. Long-term disease control is achieved in nearly all patients with early-stage (stage I/II) disease and in approximately 85% of those with stage III disease, while even patients with metastatic disease maintain a 2-year remission rate exceeding 50% (8). In light of the excellent outcomes observed in early-stage pLELC, the accurate prediction and mitigation of recurrence risk are of paramount importance. Therefore, the development of a reliable model to predict postoperative recurrence is essential for implementing risk stratification and further improving therapeutic outcomes.

Prognostic models for predicting recurrence in resectable pLELC remain limited, as they predominantly rely on conventional clinicopathological variables such as tumor-node-metastasis (TNM) stage, lymph node status, and tumor size. The distinct pathogenesis of pLELC, characterized by persistent EBV infection—a feature absent in typical NSCLC—suggests a potentially unique tumor microenvironment and biological behavior. Unlike most NSCLCs, pLELC harbors few classic driver mutations but exhibits prominent lymphocytic infiltration and high programmed death-ligand 1 (PD-L1) expression, mirroring nasopharyngeal carcinoma and indicating a virally driven oncogenic and immune escape process (2,9). Consequently, recurrence models derived from NSCLC cohorts are likely of limited value for individualized risk assessment in resected pLELC.

Increasing evidence indicates that recurrence risk is collectively determined by the primary tumor’s characteristics, the host’s systemic status, and perioperative factors (10,11). Beyond tumor pathology, systemic parameters such as immune-inflammatory and coagulation markers provide critical prognostic insights. For example, elevated D-dimer, indicative of a hypercoagulable and hyperfibrinolytic state, is associated with adverse outcomes across various cancers (12). Concurrently, dynamic fluctuations in inflammatory and immune biomarkers—including the platelet count and neutrophil-to-lymphocyte ratio (NLR)—may reflect ongoing immune remodeling and tumor-mediated evasion, positioning them as potential early indicators of recurrence (13,14). Furthermore, perioperative variables—such as prolonged operative time, significant blood loss, and the presence of vascular tumor emboli—can signal heightened tumor invasiveness and a pronounced surgical stress response, which are likewise implicated in promoting postoperative recurrence (15).

This study aimed to evaluate postoperative recurrence risk and prognostic outcomes in patients with resected pLELC. By incorporating dynamic hematologic parameters—such as D-dimer, platelet count, and NLR—into the analysis, we sought to develop an individualized recurrence prediction model and to characterize recurrence patterns and prognostic implications. These efforts are intended to improve postoperative surveillance and guide timely secondary interventions. We present this article in accordance with the TRIPOD reporting checklist (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-1-0187/rc).

Methods

Study population

This retrospective cohort study reviewed electronic medical records from The First Affiliated Hospital of Guangzhou Medical University between August 2013 and September 2021. Patients were enrolled according to the following criteria: (I) postoperative pathological diagnosis of pLELC; (II) age between 18 and 80 years; (III) absence of locally advanced or distant metastatic disease at diagnosis; (IV) receipt of curative-intent surgical resection (R0 resection), including systematic lymph node dissection); (V) no history of other primary malignancies; and (VI) availability of complete clinical and follow-up data. Exclusion criteria were as follows: (I) a prior or synchronous primary malignancy; (II) advanced-stage pLELC; (III) receipt of palliative surgery, biopsy only, or R1/R2 resection; (IV) death within 30 days after surgery; and (V) missing essential clinical or follow-up information.

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of The First Affiliated Hospital of Guangzhou Medical University (approval No. ES-2023-077). Informed consent was waived in this retrospective study.

Clinical and laboratory data collection

Comprehensive clinical and pathological data were systematically collected, including demographic variables (age, sex, smoking history, and comorbidities) and tumor-related characteristics (gross type: central or peripheral; TNM stage; and presence of vascular tumor thrombus). Treatment-related information included whether patients had received neoadjuvant therapy. Hematologic parameters included preoperative and postoperative platelet counts, D-dimer levels, tumor biomarkers, and inflammation-related indices such as the NLR, platelet-to-lymphocyte ratio (PLR), and lymphocyte-to-monocyte ratio (LMR). Postoperative laboratory data were obtained within 14 days after surgery, and changes from baseline were calculated to reflect dynamic perioperative variations. All laboratory measurements were performed in the same institutional laboratory using standardized assays, with quality control and verification conducted by trained laboratory personnel to ensure consistency and reliability.

Follow-up and outcome definitions

Patient survival data were obtained through the hospital’s electronic medical record system and supplemented by telephone follow-up. Locoregional recurrence was defined as recurrence confined to the ipsilateral lung, bronchial stump, pleura, or regional lymph nodes (hilar, mediastinal, or supraclavicular). Distant metastasis was defined as recurrence involving the contralateral lung or any extrathoracic organs (e.g., bone, liver), and this category also included patients who presented with both locoregional and distant lesions. The primary endpoints were disease-free survival (DFS) and overall survival (OS). DFS was defined as the interval from the date of surgery to the first documented recurrence or metastasis, with censoring at the date of the last follow-up for patients without recurrence. OS was defined as the interval from diagnosis to death from any cause, or censored at the last follow-up for surviving patients. A secondary endpoint, post-recurrence survival (PRS), was defined as the time from disease recurrence to death or last follow-up.

Statistical analysis

Continuous variables were compared using the independent samples t-test or Mann-Whitney U test, depending on data distribution, while categorical variables were analyzed using the χ² test or Fisher’s exact test. Survival curves for OS and DFS were generated using the Kaplan-Meier method, and differences between groups were evaluated with the log-rank test and Breslow test. Stratified analyses were further performed according to TNM stage and recurrence pattern. Prognostic factors were first evaluated by univariate Cox proportional hazards regression. Variables with statistical significance were subsequently included in a multivariate Cox regression model. A stepwise selection approach (entry/removal criterion α=0.05) was applied to avoid overfitting, and hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated. Based on the independent prognostic factors identified through stepwise regression, a recurrence risk prediction model was developed. A forest plot was used to visualize the relative effects of the included variables, and a nomogram was constructed to provide individualized predictions of 1-, 3-, and 5-year recurrence probabilities. Model discrimination was assessed using receiver operating characteristic (ROC) curves and the area under the curve (AUC), and compared with the traditional TNM staging system. Internal validation was performed using bootstrap resampling (1,000 iterations) to evaluate model robustness and calibration. For recurrence pattern analysis, the specific sites of recurrence were recorded in detail, and survival outcomes were compared across subgroups to explore the prognostic implications of different recurrence characteristics. All statistical analyses were conducted using SPSS software (version 26.0) and R software (version 4.2.1). All tests were two-sided, and P<0.05 was considered statistically significant.

Results

Patient characteristics

Between August 2013 and September 2021, 497 patients with pathologically confirmed pLELC were initially screened. All patients were EBER‑positive by in situ hybridization. After excluding patients without follow-up at our institution (n=114), those with metastatic nasopharyngeal carcinoma (n=23), and one patient lost to follow-up, a total of 359 patients were identified as having primary pLELC. Among them, 238 patients who underwent curative surgery were included in the final analysis, while 121 patients who received non-surgical treatment were excluded (Figure 1). The follow-up was conducted until July 1, 2023, with a median duration of 52.9 months (95% CI: 34.0–69.6). During this period, 57 patients (23.9%) experienced recurrence, while 181 (76.1%) remained disease-free. The median age of the cohort was 54 years [interquartile range (IQR): 46–62 years], with no significant difference observed between the recurrent and non-recurrent groups (52.8±10.3 vs. 53.8±11.4 years, P=0.44). Similarly, no significant intergroup differences were found in sex distribution (female: 49.2%), smoking history (21.8%), or comorbidity rates (19.3%) (all P>0.05). In contrast, central-type tumors were significantly more prevalent in the recurrence group (63.2%, 36/57) than in the recurrence-free group (35.9%, 65/181; P<0.001). The recurrence group was also associated with more advanced tumor (T) stages (P=0.03) and a significantly higher incidence of N2 lymph node metastasis (61.4% vs. 23.8%; P<0.001). Consequently, stage III disease was far more common in patients with recurrence (64.9% vs. 28.7% for stage I/II combined; P<0.001) (Table 1).

Figure 1 Flowchart of patient selection and study design for pLELC. pLELC, pulmonary lymphoepithelial carcinoma.

Survival analysis

Subsequent analysis of survival outcomes revealed a marked disparity between patients with and without recurrence. Patients with recurrence (n=57) had a median OS (mOS) of 77.4 months (95% CI: 63.9–91.0), markedly inferior to those without recurrence (n=181), who did not reach mOS (mean: 99.0 months; 95% CI: 96.0–102.1, P<0.001) (Figure 2). To determine whether tumor recurrence was an independent prognostic factor, we performed a Cox regression analysis. Univariate Cox proportional hazards analysis identified several significant predictors of shorter OS, including older age (P=0.007), stage III disease (vs. stage I, P=0.01), presence of comorbidities (P=0.02), central-type gross morphology (P=0.007), and tumor recurrence (P<0.001) (Table 2). After multivariate adjustment, age (HR =1.07, 95% CI: 1.03–1.12, P=0.002), central-type gross morphology (HR =2.51, 95% CI: 1.01–6.22, P=0.047), and tumor recurrence (HR =9.24, 95% CI: 5.76–14.82, P<0.001) remained independent prognostic factors for OS (Table 2).

Figure 2 Kaplan-Meier curves for OS in patients with resected pLELC, stratified by recurrence status. OS, overall survival; pLELC, pulmonary lymphoepithelial carcinoma.

Recurrence risk analysis and model construction

Given that recurrence was identified as the most critical determinant of OS in patients with pLELC, we further investigated factors associated with recurrence risk to construct a predictive model for DFS. Univariate Cox regression analysis identified several factors significantly associated with shorter DFS. These included advanced TNM stage (stage II: HR =3.99, 95% CI: 1.53–10.39, P=0.005; stage III: HR =7.31, 95% CI: 3.08–17.32, P<0.001), central-type tumor morphology (HR =2.57, 95% CI: 1.50–4.41, P=0.001, the presence of tumor thrombus (HR =2.44, 95% CI: 1.30–4.57, P=0.006), elevated initial D-dimer level (HR =1.01, 95% CI: 1.00–1.01, P<0.001), elevated PLR (HR =1.01, 95% CI: 1.00–1.01, P=0.03), elevated postoperative NLR (HR =1.24, 95% CI: 1.05–1.46, P=0.01), cancer antigen 125 (CA125) (HR =1.02, 95% CI: 1.01–1.04, P=0.001), CA153 (HR =1.03, 95% CI: 1.01–1.07, P=0.04), and longer operative duration (HR =1.01, 95% CI: 1.01–1.01, P=0.005). In contrast, no significant association with DFS was observed for age, smoking history, comorbidities, family history, tracheal infiltration, pleural invasion, carcinoembryonic antigen (CEA), initial NLR, postoperative platelet count (all P>0.05) (Table 3).

In the multivariate Cox regression (Table S1), 13 variables were entered with only 57 recurrence events, yielding the events-per-variable (EPV) ratio below 10 and suggesting a possible risk of overfitting. To mitigate this risk, a stepwise selection procedure (α=0.05) was employed. This approach identified five independent prognostic factors for recurrence: central-type tumor (HR =2.41, 95% CI: 1.40–4.16, P=0.002), preoperative D-dimer level (HR =1.001, 95% CI: 1.0001–1.001, P=0.01), the difference between postoperative and preoperative platelet (HR =1.005, 95% CI: 1.001–1.009, P=0.008), N stage (HR =1.66, 95% CI: 1.33–2.07, P<0.001), and elevated NLR (HR =1.198, 95% CI: 1.02–1.41, P=0.03). Based on the factors retained through stepwise selection, a nomogram was subsequently constructed to facilitate individualized recurrence-risk estimation (Figure 3A). The recurrence risk prediction model, constructed with selected variables, demonstrated superior discriminatory performance at all evaluated time points in the training cohort compared to the conventional TNM staging system. Specifically, the model achieved AUCs of 0.822, 0.792, and 0.744 for 1-, 3-, and 5-year predictions, respectively, outperforming the TNM system (AUCs: 0.644, 0.705, and 0.665). Internal validation with 1,000 bootstrap resamples confirmed model robustness. The bootstrap-corrected concordance index (C-index) was 0.752 (original C-index: 0.771), with an estimated optimism of 0.038. Despite a slight decrease, the corrected C-index remained above 0.7, indicating a robust discriminative ability for recurrence risk prediction (Figure 3B-3D).

Figure 3 Development and performance of the recurrence risk prediction model in patients with resected pLELC. (A) Nomogram for predicting the risk of postoperative recurrence in patients with resected pLELC. (B-D) Time-dependent ROC curves comparing the predictive performance of the recurrence risk model and the TNM staging system for 1- (B), 3- (C), and 5-year (D) DFS. The analysis is based on independent prognostic factors identified in patients with resected pLELC. AUC, area under the curve; DFS, disease-free survival; NLR, neutrophil-to-lymphocyte ratio; pLELC, pulmonary lymphoepithelial carcinoma; ROC, receiver operating characteristic; TNM, tumor-node-metastasis.

Recurrence pattern and survival analysis

Following the establishment of the recurrence-prediction model, we further explored recurrence patterns and their prognostic implications, with a focus on comparing locoregional and distant relapse. Among all patients with recurrent pLELC, the median DFS was 21.8 months (IQR: 11.1–36.2 months) (Figure 4A). Of these, 50 patients had complete data on recurrence sites and were included in the pattern analysis. The distribution of recurrence sites revealed that regional lymph node recurrence was the most common (76.0%), comprising single-node (34.0%) and multiple-node (42.0%) involvement. Other recurrence sites included ipsilateral pulmonary (34.0%), bone (26.0%), pleural (18.0%), liver (16.0%), and contralateral pulmonary (12.0%) (Figure 4B). It should be noted that the percentages for distant sites sum to over 100% as some patients had multiple metastatic sites. The median DFS was comparable between the locoregional (21.2 months, IQR: 11.0–32 months) and distant recurrence groups (20.5 months, IQR: 8.6–41.4 months) (Figure 4A). Kaplan-Meier analysis confirmed no significant difference in DFS between the two groups (log-rank test, P=0.32) (Figure 4C). Similarly, OS also showed no significant difference between the groups (P=0.07), although a trend toward shorter OS was observed in the distant metastasis group (Figure 4D). In contrast, a significant difference was observed in PRS. The median PRS for patients with distant metastasis was 53.5 months (95% CI: 28.3–78.7 months), which was significantly shorter than that of the locoregional recurrence group, in which the median PRS was not reached (P=0.04) (Figure 4E). However, comparisons of DFS, PRS, or OS among these different recurrence sites (e.g., lymph node vs. pulmonary vs. bone) revealed no statistically significant differences.

Figure 4 Survival outcomes and recurrence patterns in patients with resected pLELC. (A) Swimmer plot depicting DFS and OS among patients following pLELC resection. (B) Distribution of recurrence sites in the pLELC cohort. Human organ and bone illustrations in (B) were created by brgfx and obtained from Freepik (human organ diagram: https://www.freepik.com/free-vector/human-body-internal-organs-diagram_258584397.htm; bone illustration: https://www.freepik.com/free-vector/skeleton-lower-back-suffering-from-ankylosing-spondylitis_24781235.htm). (C-E) Kaplan-Meier curves for DFS (C), OS (D), and PRS (E) stratified by recurrence site in pLELC patients. DFS, disease-free survival; OS, overall survival; pLELC, pulmonary lymphoepithelial carcinoma; PRS, post-recurrence survival; TNM, tumor-node-metastasis.

Discussion

In this study, recurrence was identified as an independent prognostic factor for OS in patients with pLELC who underwent radical resection, emphasizing the critical importance of identifying and managing patients at high risk of recurrence to achieve better long-term outcomes. We further developed a recurrence prediction model that integrated clinical characteristics with dynamic hematological markers, demonstrating that N stage, central tumor location, elevated preoperative D-dimer levels, postoperative increases in platelet count, and postoperative NLR were independent predictors of recurrence. The proposed model showed superior predictive performance compared with the conventional TNM staging system, supporting its potential clinical utility. In addition, we delineated the recurrence landscape of pLELC: locoregional relapse and distant metastasis occurred with similar frequency and comparable DFS, whereas distant metastasis was associated with significantly shorter PRS, indicating a markedly worse prognosis. Lymph node relapse and intrathoracic spread were the predominant recurrence patterns observed in this cohort. These findings not only refine risk stratification in postoperative pLELC, but also provide guidance for tailoring surveillance strategies and prioritizing patients who may benefit from intensified systemic monitoring.

Our findings indicate that while TNM staging was significantly associated with OS in univariate analysis, only recurrence retained independent prognostic significance in multivariate Cox regression. This suggests that the impact of TNM stage on OS may be mediated predominantly via the risk of postoperative recurrence. Patients with higher stage disease often carry greater tumor burden and increased nodal metastasis, which predispose to recurrence after surgery and consequently shorten OS (16). Previous studies have shown that greater tumor burden and the presence of lymph-node micrometastases are significantly correlated with both DFS and OS in lung cancer patients (17,18). Such evidence not only supplements prior research in squamous cell lung carcinoma and pLELC, but also highlights the critical role of recurrence in this tumor subtype (19). Recurrence, as an independent adverse prognostic factor in pLELC, not only signals disease progression but also reflects incomplete eradication of tumor clones by initial definitive treatment. Prospective studies should incorporate more sensitive measures—such as minimal residual disease (MRD) or circulating tumor DNA (ctDNA)—or more detailed pathological assessments to improve postoperative risk stratification.

The hematologic indicators included in our prediction model are biologically plausible (20,21). Elevated D-dimer levels reflect activation of the coagulation-fibrinolysis axis, which facilitates tumor dissemination (12,22). Perioperative increases in platelet count may indicate a pro-metastatic microenvironment, as platelets facilitate epithelial-mesenchymal transition and protect circulating tumor cells from immune clearance (13). An elevated postoperative NLR suggests a systemic inflammatory state and impaired anti-tumor immunity, which have been repeatedly associated with inferior outcomes in lung cancer (14,23,24). Classical pathological features such as advanced nodal stage and central tumor location may also reflect higher tumor burden or technical constraints in surgical clearance, increasing the likelihood of residual microscopic disease that subsequently leads to recurrence (25). Unlike models relying solely on static clinical or radiologic variables, our approach integrates perioperative hematologic parameters reflecting real-time alterations, thereby enabling more individualized recurrence-risk stratification (26,27). The model demonstrated strong discrimination outperforming the conventional TNM staging system. Importantly, these hematologic indicators are inexpensive, widely available, and feasible for longitudinal monitoring, supporting their utility in postoperative surveillance. Traditional pathological variables such as tumor differentiation were not included in the final model. However, this is unlikely to compromise performance given the limited heterogeneity of differentiation in pLELC and the inconsistent evidence supporting its prognostic value in prior literature (28). Overall, these findings underscore the value of incorporating dynamic hematologic markers into recurrence prediction for pLELC.

In this study, the median DFS among 57 patients who experienced postoperative recurrence of pLELC was 21.8 months (IQR 11.1–36.2 months), indicating that recurrence tends to occur later and more heterogeneously than in typical NSCLC (29,30). The lymphocyte-rich stroma and favorable immune microenvironment of pLELC may delay tumor immune escape and metastatic dissemination (28). Late recurrence in some patients highlights the need for extended and individualized follow-up. The recurrence of pLELC was predominantly locoregional, whereas conventional squamous cell carcinoma usually relapses distantly, predominantly to bone and liver (31). Meanwhile, nasopharyngeal carcinoma—another EBV-related malignancy—mainly exhibits delayed locoregional relapse (32). Thus, pLELC appears to exhibit an intermediate recurrence pattern between nasopharyngeal carcinoma and NSCLC, reflecting its unique EBV-driven biology (28,33). Variations in stage and treatment across studies may contribute to inconsistent findings, emphasizing the need for multicenter standardized analyses. In our cohort, DFS did not significantly differ between local and distant recurrence, and OS showed only a nonsignificant trend toward poorer outcomes in patients with distant metastasis (P=0.07). However, PRS was significantly shorter in the distant metastasis group (P=0.04). This finding suggests that distant recurrence reflects a more aggressive tumor biology with limited responsiveness to local salvage therapy, underscoring the importance of early identification and surveillance of patients at risk for distant spread.

This study has several limitations. First, it was a single-center retrospective analysis with a limited sample size and relatively few recurrence events, which may have led to an underestimation of the effects of certain prognostic variables. Although stepwise regression and internal validation enhanced the robustness of our model, external validation and prospective studies are still required to confirm its generalizability. Second, hematologic biomarkers are influenced by multiple perioperative and individual factors, such as infection or surgical stress, and the causal relationship between these biomarkers and recurrence risk remains to be clarified through mechanistic and functional studies. Subsequently, although the Cox proportional hazards model was adopted in this study due to its widespread use in prognostic modeling and its ability to generate individualized risk predictions through estimation of baseline survival functions, it does not assume a specific underlying survival distribution. Compared with parametric survival models, this may limit its performance in scenarios requiring precise estimation of absolute event times or long-term extrapolation. Future studies may incorporate parametric approaches, such as exponential or Weibull models, to further improve the accuracy of absolute survival time prediction. Finally, although EBV DNA has been established as an important prognostic biomarker for pLELC, its detection rate among surgical patients was low, and longitudinal postoperative monitoring was lacking. Future studies integrating MRD indicators such as EBV DNA and ctDNA, and validating the model in large, multicenter prospective cohorts, are warranted to further enhance its clinical applicability.

Conclusions

In summary, this study innovatively developed a postoperative recurrence-prediction model for pLELC by integrating clinical characteristics with dynamic hematologic parameters. The model demonstrated superior predictive performance compared with the conventional TNM staging system, enabling more accurate identification of patients at high risk of recurrence and providing a practical tool for individualized surveillance and early intervention. Owing to its strong clinical interpretability and reliance on readily available laboratory indicators, the model offers excellent accessibility and translational potential in real-world settings. Moreover, our analysis of recurrence patterns revealed that patients with distant metastasis had worse survival outcomes than those with locoregional relapse, highlighting the biological heterogeneity of recurrence in pLELC. Future multicenter prospective studies incorporating molecular residual disease markers such as EBV DNA and ctDNA are warranted to further enhance the robustness and clinical applicability of the model.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-1-0187/rc

Data Sharing Statement: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-1-0187/dss

Peer Review File: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-1-0187/prf

Funding: This work was supported by grants from the Natural Science Fund of Guangdong Basic and Applied Basic Research Foundation (No. 2026A1515012452 to H.C.), the National Natural Science Foundation of China (No. 82203640 to H.C.), Guangzhou Science and Technology Plan Project (No. 2024B03J0862 to C.Z.), the Guangzhou Basic and Applied Basic Research Foundation (No. SL2023A04J00549 to H.C.), China Postdoctoral Science Foundation (No. 2022M710893 to H.C.), and Postdoctoral Startup Foundation of Guangzhou City (to H.C.). The funding bodies were not involved in the design of this study, in the collection, analysis, and interpretation of the data, or in writing of the manuscript.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-1-0187/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. This study was approved by the Ethics Committee of The First Affiliated Hospital of Guangzhou Medical University (approval No. ES-2023-077). Informed consent was waived in this retrospective study.

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. Fan Y, Li C, Qin J, et al. Primary pulmonary lymphoepithelioma-like carcinoma. Med Oncol 2020;37:20. [Crossref] [PubMed]
2. Hong S, Liu D, Luo S, et al. The genomic landscape of Epstein-Barr virus-associated pulmonary lymphoepithelioma-like carcinoma. Nat Commun 2019;10:3108. [Crossref] [PubMed]
3. Sasaki T, Kuno H, Hiyama T, et al. 2021 WHO Classification of Lung Cancer: Molecular Biology Research and Radiologic-Pathologic Correlation. Radiographics 2024;44:e230136. [Crossref] [PubMed]
4. Lin L, Lin T, Zeng B. Primary lymphoepithelioma-like carcinoma of the lung: An unusual cancer and clinical outcomes of 14 patients. Oncol Lett 2017;14:3110-6. [Crossref] [PubMed]
5. Chen FF, Yan JJ, Lai WW, et al. Epstein-Barr virus-associated nonsmall cell lung carcinoma: undifferentiated "lymphoepithelioma-like" carcinoma as a distinct entity with better prognosis. Cancer 1998;82:2334-42. [Crossref] [PubMed]
6. Ooi GC, Ho JC, Khong PL, et al. Computed tomography characteristics of advanced primary pulmonary lymphoepithelioma-like carcinoma. Eur Radiol 2003;13:522-6. [Crossref] [PubMed]
7. Hu Y, Ren S, Liu Y, et al. Pulmonary Lymphoepithelioma-Like Carcinoma: A Mini-Review. Onco Targets Ther 2020;13:3921-9. [Crossref] [PubMed]
8. Tay CK, Chua YC, Takano A, et al. Primary pulmonary lymphoepithelioma-like carcinoma in Singapore. Ann Thorac Med 2018;13:30-5. [Crossref] [PubMed]
9. Chau SL, Tong JH, Chow C, et al. Distinct Molecular Landscape of Epstein-Barr Virus Associated Pulmonary Lymphoepithelioma-Like Carcinoma Revealed by Genomic Sequencing. Cancers (Basel) 2020;12:2065. [Crossref] [PubMed]
10. Kitano Y, Okabe H, Yamashita YI, et al. Tumour-infiltrating inflammatory and immune cells in patients with extrahepatic cholangiocarcinoma. Br J Cancer 2018;118:171-80. [Crossref] [PubMed]
11. She S, Shi J, Zhu J, et al. Impact of inflammation and the immune system on hepatocellular carcinoma recurrence after hepatectomy. Cancer Med 2024;13:e7018. [Crossref] [PubMed]
12. Ay C, Dunkler D, Pirker R, et al. High D-dimer levels are associated with poor prognosis in cancer patients. Haematologica 2012;97:1158-64. [Crossref] [PubMed]
13. Gay LJ, Felding-Habermann B. Contribution of platelets to tumour metastasis. Nat Rev Cancer 2011;11:123-34. [Crossref] [PubMed]
14. Yin Y, Wang J, Wang X, et al. Prognostic value of the neutrophil to lymphocyte ratio in lung cancer: A meta-analysis. Clinics (Sao Paulo) 2015;70:524-30. [Crossref] [PubMed]
15. Hiller JG, Perry NJ, Poulogiannis G, et al. Perioperative events influence cancer recurrence risk after surgery. Nat Rev Clin Oncol 2018;15:205-18. [Crossref] [PubMed]
16. Ji H, Hu C, Yang X, et al. Lymph node metastasis in cancer progression: molecular mechanisms, clinical significance and therapeutic interventions. Signal Transduct Target Ther 2023;8:367. [Crossref] [PubMed]
17. Jonnalagadda S, Smith C, Mhango G, et al. The number of lymph node metastases as a prognostic factor in patients with N1 non-small cell lung cancer. Chest 2011;140:433-40. [Crossref] [PubMed]
18. Bazan JG, Duan F, Snyder BS, et al. Metabolic tumor volume predicts overall survival and local control in patients with stage III non-small cell lung cancer treated in ACRIN 6668/RTOG 0235. Eur J Nucl Med Mol Imaging 2017;44:17-24. [Crossref] [PubMed]
19. Qiang G, Liang C, Yu Q, et al. Risk factors for recurrence after complete resection of pathological stage N2 non-small cell lung cancer. Thorac Cancer 2015;6:166-71. [Crossref] [PubMed]
20. Pang B, Wang J, Wang J, et al. Association of baseline plasma D-dimer and platelets with progression-free survival in patients with stage IV non-small cell lung cancer treated with anti-PD-1 antibody. Transl Lung Cancer Res 2024;13:3106-21. [Crossref] [PubMed]
21. Wang J, Li H, Xu R, et al. The MLR, NLR, PLR and D-dimer are associated with clinical outcome in lung cancer patients treated with surgery. BMC Pulm Med 2022;22:104. [Crossref] [PubMed]
22. Siddiqui NA, Malik M, Wijeratne Fernando R, et al. D-Dimer: A Potential Solution to Problems of Cancer Screening, Surveillance, and Prognosis Assessment. Cureus 2021;13:e15064. [Crossref] [PubMed]
23. Wculek SK, Malanchi I. Neutrophils support lung colonization of metastasis-initiating breast cancer cells. Nature 2015;528:413-7. [Crossref] [PubMed]
24. Ménétrier-Caux C, Ray-Coquard I, Blay JY, et al. Lymphopenia in Cancer Patients and its Effects on Response to Immunotherapy: an opportunity for combination with Cytokines? J Immunother Cancer 2019;7:85. [Crossref] [PubMed]
25. Coello MC, Luketich JD, Litle VR, et al. Prognostic significance of micrometastasis in non-small-cell lung cancer. Clin Lung Cancer 2004;5:214-25. [Crossref] [PubMed]
26. Kim PJ, Hwang HS, Choi G, et al. A new model using deep learning to predict recurrence after surgical resection of lung adenocarcinoma. Sci Rep 2024;14:6366. [Crossref] [PubMed]
27. Liang Z, Lin W, Fang S, et al. Development and Validation of a Predictive Nomogram for Early Postoperative Recurrence in Non-Small Cell Lung Cancer. Med Sci Monit 2025;31:e948441. [Crossref] [PubMed]
28. Chen B, Zhang Y, Dai S, et al. Molecular characteristics of primary pulmonary lymphoepithelioma-like carcinoma based on integrated genomic analyses. Signal Transduct Target Ther 2021;6:6. [Crossref] [PubMed]
29. Qiu B, Guo W, Zhang F, et al. Dynamic recurrence risk and adjuvant chemotherapy benefit prediction by ctDNA in resected NSCLC. Nat Commun 2021;12:6770. [Crossref] [PubMed]
30. Karacz CM, Yan J, Zhu H, et al. Timing, Sites, and Correlates of Lung Cancer Recurrence. Clin Lung Cancer 2020;21:127-135.e3. [Crossref] [PubMed]
31. Feld R, Rubinstein LV, Weisenberger TH. Sites of recurrence in resected stage I non-small-cell lung cancer: a guide for future studies. J Clin Oncol 1984;2:1352-8. [Crossref] [PubMed]
32. Chen YH, Luo SD, Wu SC, et al. Clinical Characteristics and Predictive Outcomes of Recurrent Nasopharyngeal Carcinoma-A Lingering Pitfall of the Long Latency. Cancers (Basel) 2022;14:3795. [Crossref] [PubMed]
33. Lee TH, Shin H, Ahn YC, et al. Regional lymph node recurrence after stereotactic body radiation therapy for lung cancer: Patterns of recurrence, treatment approaches, and clinical outcomes (KROG 21-09). Radiother Oncol 2023;183:109572. [Crossref] [PubMed]