OL Oncology Letters 1792-1074 1792-1082 D.A. Spandidos 10.3892/ol.2026.15566 OL-31-6-15566 Articles Clinical predictors of leptomeningeal metastasis from lung adenocarcinoma based on retrospective analysis LuYunyun 1 LuYi 1 ZhangYueyue 2 DuHongyuan 2 CaoXuan 2 JinDexi 2 ChenGuofu 2 YangXiaohui 2 Department of Radiation Oncology, Ningbo Medical Center Lihuili Hospital, Ningbo, Zhejiang 315048, P.R. China Department of General Surgery, School of Medicine, The First People's Hospital of Wenling (Taizhou University Affiliated Wenling Hospital), Taizhou University, Taizhou, Zhejiang 318000, P.R. China Correspondence to: Professor Guofu Chen or Dr Xiaohui Yang, Department of General Surgery, School of Medicine, The First People's Hospital of Wenling (Taizhou University Affiliated Wenling Hospital), Taizhou University, 1139 Shifu Boulevard, Jiaojiang, Taizhou, Zhejiang 318000, P.R. China, E-mail: wlcgf@163.com, E-mail: xiaohuiy@tzc.edu.cn 062026 01042026 31 6 211 10102025 19022026 Copyright: © Lu et al. 2026 This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

Leptomeningeal metastasis (LM) is common in patients with lung adenocarcinoma, leading to high mortality rates. The predictors of systematic survival in patients with LM and lung adenocarcinoma remain poorly understood. The present study retrospectively analyzed 78 lung adenocarcinoma patients with or without LM treated at Ningbo Medical Center, Lihuili Hospital (Ningbo, China) between November 2016 and August 2024. Clinical characteristics and baseline hematological parameters obtained at LM diagnosis were evaluated for their associations with overall survival (OS). Median OS was 9.0 months (range: 0.2–48.9 months) in patients with LM. Univariate analysis identified age <60 years, Eastern Cooperative Oncology Group performance status (ECOG PS) 0–2, receipt of brain radiotherapy after a diagnosis of LM, no extracranial metastasis, epidermal growth factor receptor (EGFR) 19 mutation, receipt of third-generation EGFR tyrosine kinase inhibitor therapy before a diagnosis of LM, neutrophil-to-lymphocyte ratio <7.5755, platelet-to-lymphocyte ratio (PLR) <156.035 and a molecular graded prognostic assessment >1.5 to be significant predictors of superior OS. According to multivariate analysis, extracranial metastasis [hazard ratio (HR)=2.291; 95% CI, 1.074–4.888; P=0.032], PLR <156.035 (HR=0.233; 95% CI, 0.123–0.442; P<0.001) and ECOG PS 0–2 (HR=0.302; 95% CI, 0.152–0.599; P=0.001) remained predictive of OS. In conclusion, extracranial metastasis, PLR and ECOG PS were identified to be prospective independent clinical prognostic indicators of survival in patients with lung adenocarcinoma and LM. Overall, the present study highlighted the potential use of clinical characteristics and hematological variables before treatment to predict the outcomes of patients with lung adenocarcinoma complicated with LM.

 lung adenocarcinoma leptomeningeal metastasis tyrosine kinase inhibitor neutrophil-to-lymphocyte ratio platelet-to-lymphocyte ratio Eastern Cooperative Oncology Group performance status Taizhou University Medicine Special Project 2023XY01 Taizhou Anti-Cancer Association Cancer Special Research Project TACA2025-A02 The present study was supported by Taizhou University Medicine Special Project (grant no. 2023XY01) and Taizhou Anti-Cancer Association Cancer Special Research Project (grant no. TACA2025-A02). Introduction

Leptomeningeal metastasis (LM) is a type of central nervous system metastasis defined by the infiltration and spread of tumor cells in the leptomeninges or spinal leptomeninges. Clinically, it can present with a broad spectrum of symptoms, including headache, vomiting, mental and behavioral changes, seizures, hearing loss, visual disturbances, limb weakness and paresthesia (1,2). LM represents 11–20% of central nervous system metastases, with the highest rates reported in melanoma and lung cancer (2). LM is associated with a notably poor prognosis and short survival time, with median overall survival (OS) ranging from 3.6 weeks to 11 months (1). LM develops in 5–10% all patients with solid tumors and 3–5% patients with non-small cell lung cancer (NSCLC) (2–4). The development of LM is a complex, multi-step process that is driven by tumor cells acquiring capabilities for invasion and migration. These capabilities are supported by dynamic cytoskeletal reorganization and the activation of conserved pro-metastatic signaling pathways, such as the SPP1-CD44 axis and CXCL2/CCL2 chemokines (5). In the absence of treatment, patients with LM can succumb within 4–6 weeks. Conversely, the median duration of survival with treatment ranges from 3.6 to 4.5 months (6,7). Because of the blood-brain barrier (BBB), the majority of drugs cannot enter the meningeal cavity; therefore the meningeal cavity serves as a sanctuary site for tumor cells, allowing them to evade systemic therapies. However, due to advancements in imaging and systemic disease control, LM is increasingly diagnosed in patients with lung adenocarcinoma (8,9). With improvements in therapy, including targeted therapy, radiotherapy and intrathecal therapy, OS in patients with NSCLC and LM has improved, increasing from 1–3 to a current range of 11.0–20.4 months (10–12).

A single-arm phase II clinical study previously revealed that intrathecal therapy was efficient and safe in patients with LM from epidermal growth factor receptor (EGFR) mutation-positive NSCLC. Using the Response Assessment in Neuro-Oncology (RANO) criteria, the response rate was high at ~80.3%, but certain patients (19.7%) did not respond to treatment (10). In the BLOOM trial of patients with EGFR mutation-positive NSCLC and LM, a double dose of osimertinib produced an effective response with an objective response rate (ORR) of 62% and a median duration of response of 15.2 months, according to a neuroradiological blinded central independent review (11). Similarly, in 80 patients with EGFR or anaplastic lymphoma kinase (ALK) mutation-positive lung adenocarcinoma, median OS was 10.4 months, compared with 3.8 months in patients with wild-type lung adenocarcinoma (13). Furthermore, a retrospective study of patients with lung adenocarcinoma and LM revealed that brain radiotherapy could prolong survival, with median OS of 6.2 months (95% CI, 4.4–12.4) (13). This outcome was achieved because brain radiotherapy could damage the BBB, thereby increasing the intracranial antitumor drug concentration (14). Although a growing number of therapeutic methods have led to improvements in OS and ORR among patients with LM from lung adenocarcinoma, the existence of non-responders highlights the need to identify the factors influencing treatment outcomes.

The present retrospective study aimed to identify prognostic factors for advanced lung adenocarcinoma with LM to enhance the efficacy of novel therapies by Kaplan-Meier method.

 Materials and methods <sec> <title>Patients

The present study was approved by the Ethics Committee of the Ningbo Medical Center, Lihuili Hospital (approval no. KY2024SL297-01). Requirement for informed consent was waived by the Ethics Committee of Ningbo Medical Center Lihuili Hospital since this is a retrospective study. All patients were treated between November 2016 and August 2024.

The inclusion criteria were as follows: i) Age of 18–85 years; ii) confirmation of lung adenocarcinoma by pathology; iii) LM diagnosis by cerebrospinal fluid (CSF) cytology or MRI using previous Chinese expert consensus (15); iv) peripheral blood testing within 7 days of an LM diagnosis and before treatment; and v) receipt of third-generation EGFR tyrosine kinase inhibitor (TKI) therapy, intrathecal therapy or brain radiotherapy after the LM diagnosis. The exclusion criteria were as follows: i) Immunodeficiency; ii) diagnosis of other carcinoma; iii) presence of serious co-morbidities (such as heart disease, infection or renal failure); iv) inadequate laboratory information and clinical data; v) pregnancy; vi) received corticosteroid therapy or had an infection within the 7 days preceding enrollment; and vii) gene mutations (including ALK, RET, HER2 and ROS1) other than EGFR exon 19 or 21 mutations, because for other mutations, either the patient population was small or there were no definitively effective targeted drugs established for leptomeningeal metastasis.

Clinical data collected included the following: i) Age; ii) sex; iii) smoking status; iv) ECOG PS (16); v) Karnofsky performance status (KPS) (17); vi) receipt of chemotherapy or brain radiotherapy before a diagnosis of LM; vii) receipt of brain radiotherapy after a diagnosis of LM (clinical data indicated that there were no standard guidelines prior to 2025; the decision to proceed with brain radiotherapy following LM was based on physician clinical experience rather than solely on patient condition); viii) CSF cytology; ix) brain MRI; x) driver gene mutation status; xi) brain metastasis (BM); xii) extracranial metastasis; xiii) receipt of third-generation EGFR TKI therapy before a LM diagnosis; xiv) receipt of intrathecal therapy; and xv) neurological symptoms (such as dizziness and headache). Peripheral blood test variables included the neutrophil (×109/l), platelet (×109/l) and lymphocyte counts (×109/l) within 7 working days until participants received brain radiotherapy, intrathecal therapy or targeted therapy after the LM diagnosis.

 Response evaluation

Typically, the treatment response of LM is evaluated according to the RANO and European Association of Neuro-Oncology-European Society for Medical Oncology guidelines based on the clinical neurological examination, CSF pathology and MRI (18,19). However, the relevant data were characterized by incomplete recording, inconsistent documentation and a lack of standardization among patients. Consequently, LM-specific endpoints (neurological PFS, CSF cytology clearance and intracranial control) were not reliably assessable in the present study. Therefore, OS, defined as the time from the initial diagnosis of LM to death or the last follow-up, was the primary endpoint of this study.

 Statistical analysis

The results were presented as the median, range and proportion. The Kaplan-Meier method was used for survival analysis. Univariate analysis and multivariate Cox regression analysis were performed to identify the factors independently predictive of survival. If predictors exhibited potential multicollinearity, appropriate factors would be selected through association analysis whilst considering data balance. All tests were two-sided, where P<0.05 was considered to indicate a statistically significant difference. The cut-offs of the neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) were determined by receiver operating characteristic (ROC) curve analysis and the Youden index. Based on previous studies, the molecular graded prognostic assessment (molGPA) score was calculated following the established criteria (Table I), which involved summing the individual scores for KPS, extracranial metastasis and EGFR mutation status. This score was then investigated as a predictive factor (20–22). SPSS 20.0 (IBM Corp.) was used to analyze all statistical data.

 Results <sec> <title>Characteristics of patients

From November 2016 to August 2024, LM from lung adenocarcinoma was confirmed by CSF cytology or MRI in 78 patients. The clinical information of patients is summarized in Table II. The median age of the patients was 64 years (range, 37–83). The cohort included 39 patients <60 years old, where 44 patients were male. The majority of patients were non-smokers (73.1%) and most patients had ECOG PS 0–2 (79.5%). A total of 12 patients (15.4%) received brain radiotherapy before the LM diagnosis and 24 patients (30.8%) underwent brain radiotherapy after LM was detected. In addition, 49 patients (62.8%) had BM and 58 patients (74.4%) exhibited extracranial metastasis. In addition, 69 patients (88.5%) had neurological symptoms, whilst 36 patients (46.2%) received intrathecal therapy and 47 patients (60.3%) received third-generation EGFR TKI therapy before LM was detected. EGFR mutational profiling identified exon 19 deletions in 26 patients (33.3%) and the L858R mutation (exon 21) in 30 patients (38.5%), whilst 22 (28.2%) were EGFR wild-type. Furthermore, nine EGFR-wild-type patients underwent next-generation sequencing testing, which confirmed no EGFR mutations but revealed alterations in other genes, including KRAS, TP53 and PIK3CA. The modalities used to diagnose LM were MRI alone (27 patients; 34.6%), CSF cytology alone (20 patients; 25.6%) and MRI + CSF cytology (31 patients; 39.8%). The molGPA score was 0 in 13 patients (16.7%), 0.5–1.0 in 58 patients (74.4%) and 1.5–2.0 in 7 patients (8.9%).

 Optimal cut-offs for NLR and PLR

In the present study, ROC curve analysis was used to select the appropriate cut-offs for NLR and PLR. For NLR, the optimal cut-off was 7.5755, which had an area under curve (AUC) of 0.599. Further, the optimal cut-off for PLR was 156.035 (AUC=0.654).

 Survival outcomes

Median OS for all patients was 9.0 months. OS was improved in patients <60 years old compared with those aged ≥60 years (Fig. 1) as well as in patients with an ECOG PS of 0–2 versus 3–4 (Fig. 2). The data indicated that brain radiotherapy after the diagnosis of LM can improve the OS relative to no brain radiotherapy (Fig. 3). Concerning gene mutations, OS was improved in patients with EGFR 19 exon deletion compared with those with an EGFR exon 21 mutation or EGFR wild-type (Fig. 4). Furthermore, OS was improved among patients who received third-generation EGFR TKI therapy before the diagnosis of LM compared with those who received non-third-generation EGFR TKIs (Fig. 5). By contrast, NLR ≥7.5755 vs. <7.5755 (Fig. 6) and PLR ≥156.035 vs. <156.035 (Fig. 7) were predictive of shorter OS. In addition, extracranial metastasis was associated with shorter OS vs. extracranial metastasis (Fig. 8). According to the molGPA classification, patients with a score of 0.0 had worse OS compared with those with score of 0.5–1.0 and 1.5–2.0 (Fig. 9). However, sex, smoking status, the receipt of chemotherapy before the diagnosis of LM, the receipt of brain radiotherapy before the diagnosis of LM, BM, neurological symptoms, the modality used to diagnose LM diagnosis and the receipt of intrathecal therapy were not significantly associated with OS (Table III).

Since NLR and PLR share lymphocyte counts and may exhibit collinearity, which could lead to unstable estimates if both were included in the multivariable model, only PLR was selected for inclusion in the model. Multivariate analysis identified non-extracranial metastasis [vs. extracranial metastasis; hazard ratio (HR)=2.291; 95% CI, 1.074–4.888; P=0.032], PLR<156.035 (vs. ≥156.035; HR=0.233; 95% CI, 0.123–0.442) and ECOG PS 0–2 (vs. 3–4; HR=0.302; 95% CI, 0.152–0.599) as meaningful clinical characteristics associated with OS (Table III), whereas the remaining factors were not predictive of outcomes.

 Discussion

The selection of treatment for patients with lung adenocarcinoma and LM is a difficult clinical challenge. Although the overall prognosis is grave, significant heterogeneity in survival exists, underscoring the critical need for reliable prognostic biomarkers to guide personalized management. While prior research has established several prognostic factors, including ECOG PS, intrathecal therapy and antiangiogenic therapy following LM, the integrative role of routinely available hematological indices, such as NLR and PLR, in combination with established clinical parameters, within the framework of contemporary multimodal treatment strategies (involving third-generation EGFR-tyrosine kinase inhibitors and radiotherapy), remains to be further validate (10,23). The present study aimed to address this gap by comprehensively evaluating a panel of pre-treatment clinical and hematological variables. The prognostic significance of several established factors in univariate analysis were assessed, where non-extracranial metastasis, a low PLR (<156.035) and a good performance status (ECOG PS 0–2) were identified as independent predictors of longer OS in the multivariate model. This integrative approach provides a practical and robust framework for risk stratification in this patient population.

Neutrophils, platelets and lymphocytes are commonly investigated as peripheral blood indicators in laboratory examination. The independent prognostic value of NLR and PLR in the present cohort underscores the pivotal role of the systemic inflammatory response and tumor immune microenvironment in LM pathogenesis. Consistent with the results of ex post facto studies, NLR and PLR, as projected factors of lung cancer mortality, can influence the tumor inflammatory microenvironment, particularly in patients with BM and LM (18,19,21). NLR >3.83 (HR=2.66; 95% CI 1.17–6.01; P=0.019) indicated poor survival in a consecutive cohort of patients with BM (24). In a previous systematic review and meta-analysis of patients with BM from lung cancer, 11 studies revealed that high NLR (HR=1.84; 95% CI, 1.47–2.31; P<0.001) was significantly associated with shorter OS, where an analysis of five studies indicated that high PLR was predictive of shorter OS (HR=1.53; 95% CI, 1.07–2.20; P=0.020) (25). In addition, low NLR (HR=0.49; 95% CI, 0.28–0.86; P=0.012) was associated with longer OS in patients with lung cancer and LM. Patients with pre-treatment PLR <145.44 (P=0.031) had longer OS according to a univariate analysis (20). Similarly, in a multivariate analysis of the whole cohort, a pre-treatment PLR cut-off of 209.77 was associated with a significant difference in median OS (8.0 vs. 14.5 months; P=0.047). In addition, post-treatment NLR >3.57 (HR=2.203; 95% CI, 1.060–4.578; P=0.034) was a useful prognostic factor for patients with lung cancer and LM according to a multivariate analysis (26).

Previously, Berghoff et al (27) reported that the brain is not a completely independent organ. Inflammatory cells can stimulate the release of certain substances, such as reactive oxygen species, affecting neighboring cancer cells (28). The inflammatory microenvironment mainly consists of neutrophils, platelets and lymphocytes. As key factors in the tumor response, neutrophils are positioned in the glycolytic and hypoxic tumor core, where they can promote angiogenesis to support tumor cell proliferation (29). Neutrophils represent the most common immune cell type in human blood, where they can regulate cancer initiation, growth and metastasis (29). Other cells, such as T cells and macrophages, can interact with neutrophils in the tumor microenvironment (30). By contrast, platelets can promote epithelial-mesenchymal transition to facilitate cancer cell proliferation and invasiveness. Activated platelets can help cancer cells avoid immune system surveillance and anoikis. Therefore, platelets have an important role in tumor metastasis (31). Lymphocytes inhibit tumor cell proliferation and provoke cancer cell killing by T cells (32). Studies have identified NLR and PLR as key predictors of clinical outcomes in various tumors, such as breast and colon cancer (33,34). In the unique sanctuary of the central nervous system, the value of these ratios may be amplified. Inflammatory crosstalk at the blood-CSF barrier can modulate the permeability of this barrier and create a permissive niche for tumor cell colonization and proliferation. Therefore, elevated NLR and PLR likely reflect a heightened systemic and potentially local pro-metastatic milieu that is permissive for the establishment and progression of LM, explaining their potent prognostic power in the present analysis.

The persistence of ECOG PS as an independent prognostic factor reasserts the fundamental importance of the physiological status of the patient. A poor performance status frequently mirrors a higher systemic tumor burden, increased cancer-related symptomatology and diminished physiological capacity to tolerate aggressive therapies (23). The present findings suggest that ECOG PS captures elements of disease aggressiveness and patient frailty that are not fully explained by laboratory values or mutation status alone. This finding aligns with that by Wang et al (26), who reported a significant difference in median OS (4 vs. 15 months) based on ECOG PS. Other studies have illustrated that patients with low ECOG PS have longer survival, consistent with the findings in the present study (10,20). Consequently, performance status remains an indispensable, non-redundant tool for initial prognostication and clinical decision-making in neuro-oncology.

The present univariate analysis suggested a survival benefit associated with brain radiotherapy administered after LM diagnosis. Although this effect was attenuated in the multivariate model, it hints at the potential role of brain radiotherapy in disrupting the BBB (14). Evidence from the literature suggests that brain radiotherapy is beneficial for patients with NSCLC and LM (35). In a previous retrospective study, univariate analysis illustrated that brain radiotherapy can prolong OS compared with non-brain radiotherapy (median OS: 11.4 vs. 5.0 months; P=0.051) in 80 patients with LM from NSCLC (36). Consistent with the present findings, brain radiotherapy after the diagnosis of LM was associated with longer OS (median OS=20.1 months; P=0.012) according to a univariate analysis. Although multivariate analysis did not confirm this finding-likely due to the limited number of patients who received brain radiotherapy-this treatment modality following the diagnosis of LM may still be predictive of OS in patients with NSCLC and LM (HR=0.580; P=0.132).

According to multivariate analysis, extracranial metastasis negatively influenced OS (P=0.032) in patients with LM from NSCLC in the present study. Previous studies identified extracranial metastasis as a survival biomarker. In patients with LM from NSCLC, univariate analysis illustrated that patients with extracranial metastasis had shorter OS (HR=2.09; P<0.01) regardless of the EGFR status (37). Multivariate analysis in another retrospective study also demonstrated that extracranial metastasis was an independent predictor for poorer survival compared with the absence of extracranial metastasis (median OS: 10.0 vs. 20.0 months; P=0.023) (38). In summary, extracranial metastasis portended inferior outcomes among patients with NSCLC and LM.

Different EGFR mutations have been associated with differences in OS in patients with LM. In a study by the European Society of Medical Oncology, patients with NSCLC and LM who harbored EGFR exon 19 mutations had superior OS (37). There are two reasons for the distinctions between exon 19 and 21 mutations. The tyrosine phosphorylation structure of EGFR differs between exon 19 deletion and exon 21 mutation in NSCLC cell lines, leading to dissimilar alterations of downstream signaling pathways (38). The other reason is that exon 21 mutation requires asymmetric dimerization for activation and oncogenic transformation, whereas exon 19 mutation is not associated with dimerization activation (39). Therefore, patients with LM and EGFR 19 deletion had prolonged median OS (16.1 months; P=0.019) according to univariate analysis in the present study.

Based on univariate analysis, the receipt of third-generation EGFR TKI therapy before the diagnosis of LM appeared to improve OS in the present study (OS=12.1 months; P=0.035). Third-generation EGFR TKIs have a high rate of entry into the BBB owing to their macrocyclic amide structure and optimized lipophilicity (40–42). However, no difference in OS associated with this variable was observed in the multivariate analysis. This finding could be due to the fact that individual third-generation EGFR TKIs, such as osimertinib, almonertinib, furmonertinib and befotertinib, were not investigated. Therefore, further research is needed to identify the optimal third-generation EGFR TKIs for patients with NSCLC and LM.

Previous studies identified molGPA as a prognosis risk factor for patients with LM (21,22). Zhang et al (21) reported that median OS in the low-risk molGPA group was 8.02 months (95% CI, 0.98–38.13), which significantly exceeded the survival duration in the medium- and high-risk groups. In addition, another study reported median OS times of 0.9, 5.8 and 17.7 months for the low-, medium-, and high-risk molGPA groups, respectively (P<0.001) (22). Similarly, the median OS times in these molGPA subgroups in the present study were 4.7, 8.8 and 16.1 months, respectively, with significant differences identified by univariate analysis (P=0.008). These results underscore the potential utility of the molGPA model in optimizing patient management. However, this conclusion is tempered by the small subgroup size (n=7) in the highest molGPA category, which may affect the reliability of the findings. Therefore, further studies are warranted for validation.

The current study had some limitations. It was a retrospective study, within which the small sample size raises the possibility of selection bias. The strong univariate association between post-LM brain radiotherapy and survival was attenuated in the multivariable model. This could be explained if patients receiving radiotherapy had more favorable prognostic factors at baseline, such as superior performance status, lower disease burden or greater fitness for further treatment, all of which independently predict longer survival. In addition, LM was not pathologically confirmed by CSF cytology in all patients. There was also no control group to verify the clinical utility of the cut-offs. The sequence and timing of third-generation EGFR TKI treatment or intrathecal therapy in combination with brain radiotherapy was not standardized. PD-L1 status was also not assessed. Comprehensive next-generation sequencing data was not performed for all patients due to the study's retrospective nature, which may have affected the interpretation of survival and treatment outcomes because of the unknown status of EGFR and other gene mutations. The measured values of NLR and PLR may have been confounded by variations in corticosteroid regimens and occurrence of intercurrent infections. Differences in intrathecal drugs (the doses, fractionation and sites of brain radiotherapy) and third-generation EGFR TKIs could lead to different results. No internal validation was performed when selecting the ROC-based cut-offs, where the data-driven selection of cut-offs can cause overfitting and limited generalizability. The analysis was subject to immortal time bias and confounding by indication due to the retrospective design and timing of therapy administration, which may affect the interpretation of survival benefits associated with therapies, such as radiotherapy and pre-LM TKI use. Despite these limitations, the present study provides evidence for predicting outcomes in patients with lung adenocarcinoma and LM.

In conclusion, the present study identified extracranial metastasis, PLR and ECOG PS as independent predictors of OS in patients with lung adenocarcinoma and LM. In addition, brain radiotherapy, EGFR 19 mutation, the receipt of third-generation EGFR TKI therapy before the diagnosis of LM and molGPA were significantly associated with overall survival. Although intrathecal therapy was associated with negative survival outcomes, EGFR mutation had a crucial influence on survival in the majority of patients. Future studies are needed to delineate the optimal sequencing and combination of radiotherapy, intrathecal therapy and specific TKIs.

 Acknowledgements

Not applicable.

 Availability of data and materials

The data generated in the present study may be requested from the corresponding author.

 Authors' contributions

YuL was responsible for performing the experiments and collecting data, formal analysis of data, design of the methodology, data curation (managing, cleaning, annotating, and preserving data) and writing the original draft. YiL and XC analyzed data, designing the methodology, data curation (managing, cleaning, annotating, and preserving data), validation and writing the original draft. YZ and HD were responsible for the software (programming, code development, and tool testing), design of the methodology, and reviewing and editing the manuscript. YuL and XC confirm the authenticity of all the raw data. XC edited the manuscript. DJ was responsible for the formal analysis of data, and review and editing the manuscript. GC and XY were responsible for providing resources, conceptualization of the study, reviewing and editing the manuscript, and project administration. All authors read and approved the final version of the manuscript.

 Ethics approval and consent to participate

The present study was approved by the Ethics Committee of Ningbo Medical Center Lihuili Hospital (approval no. KY2024SL297-01). This study was conducted in accordance with the Declaration of Helsinki. Requirement for informed consent was waived by the Ethics Committee of Ningbo Medical Center Lihuili Hospital since this is a retrospective study.

 Patient consent for publication

Not applicable.

 Competing interests

The authors declare that they have no competing interests.

 References ParkSBaldryRJungHASunJMLeeSHAhnJSKimYJLeeYKimDWKimSWPhase II efficacy and safety of 80 mg Osimertinib in patients with leptomeningeal metastases associated with epidermal growth factor receptor mutation-positive non-small cell lung cancer (BLOSSOM)J Clin Oncol4227472756202410.1200/JCO.24.0070838828959 ThakkarJPKumthekarPDixitKSStuppRLukasRVLeptomeningeal metastasis from solid tumorsJ Neurol Sci411116706202010.1016/j.jns.2020.11670632007755 LukasRVThakkarJPCristofanilliMChandraSSosmanJAPatelJDKumthekarPStuppRLesniakMSLeptomeningeal metastases: The future is nowJ Neurooncol156443452202210.1007/s11060-021-03924-235048267 BeauchesnePIntrathecal chemotherapy for treatment of leptomeningeal dissemination of metastatic tumoursLancet Oncol11871879201010.1016/S1470-2045(10)70034-620598636 ZhouCShanSWenLLiuDShanCJinXZhouZLiHLiJWangLImmunological and pathological characteristics of brain parenchymal and leptomeningeal metastases from non-small cell lung cancerCell Discov1172202510.1038/s41421-025-00828-740883281 LiaoBCLeeJHLinCCChenYFChangCHHoCCShihJYYuCJYangJCEpidermal growth factor receptor tyrosine kinase inhibitors for non-small-cell lung cancer patients with leptomeningeal carcinomatosisJ Thorac Oncol1017541761201510.1097/JTO.000000000000066926334749 UmemuraSTsubouchiKYoshiokaHHottaKTakigawaNFujiwaraKHoritaNSegawaYHamadaNTakataIClinical outcome in patients with leptomeningeal metastasis from non-small cell lung cancer: Okayama lung cancer study groupLung Cancer77134139201210.1016/j.lungcan.2012.03.00222487432 WangYYangXLiNJXueJXLeptomeningeal metastases in non-small cell lung cancer: Diagnosis and treatmentLung Cancer174113202210.1016/j.lungcan.2022.09.01336206679 OzcanGSinghMVredenburghJJLeptomeningeal metastasis from non-small cell lung cancer and current landscape of treatmentsClin Cancer Res291129202310.1158/1078-0432.CCR-22-158535972437 FanCJiangZTengCSongXLiLShenWJiangQHuangDLvYDuLEfficacy and safety of intrathecal pemetrexed for TKI-failed leptomeningeal metastases from EGFR+ NSCLC: An expanded, single-arm, phase II clinical TrialESMO Open9102384202410.1016/j.esmoop.2024.10238438377785 YangJCHKimSWKimDWLeeJSChoBCAhnJSLeeDHKimTMGoldmanJWNataleRBBrownAPOsimertinib in patients with epidermal growth factor receptor mutation-positive non-small-cell lung cancer and leptomeningeal metastases: The BLOOM studyJ Clin Oncol38538547202010.1200/JCO.19.0045731809241 ChengHPerez-SolerRLeptomeningeal metastases in non-small-cell lung cancerLancet Oncol19e43e45201810.1016/S1470-2045(17)30689-729304362 OyoshiHHirataHHiranoYZendaSZhouYTomizawaKFujisawaTNakamuraMHojoHMotegiAPrognostic impact of EGFR/ALK alterations in leptomeningeal metastasis from lung adenocarcinoma treated with whole-brain radiotherapyClin Exp Metastasis40407413202310.1007/s10585-023-10225-737468822 SakaguchiMMaebayashiTAizawaTIshibashiNSaitoTSuccessful treatment of nonsmall cell lung cancer patients with leptomeningeal metastases using whole brain radiotherapy and tyrosine kinase inhibitorsJ Cancer Res Ther16930932202010.4103/jcrt.JCRT_1343_1632930144 LinGWangYXinTZhangDZhangQLiYChiYFanYLiuAXuHChinese expert consensus on leptomeningeal metastases of lung cancerThorac Cancer16e70088202510.1111/1759-7714.7008840484812 OkenMMCreechRHTormeyDCHortonJDavisTEMcFaddenETCarbonePPToxicity and response criteria of the eastern cooperative oncology groupAm J Clin Oncol5649655198210.1097/00000421-198212000-000147165009 KarnofskyDAAbelmannWHCraverLFBurchenalJHThe use of the nitrogen mustards in the palliative treatment of carcinoma-with particular reference to bronchogenic carcinomaCancer1634656194810.1002/1097-0142(194811)1:4<634::AID-CNCR2820010410>3.0.CO;2-L Le RhunEWellerMvan BentMBrandsmaDFurtnerJRudàRSchadendorfDSeoaneJTonnJCWesselingPLeptomeningeal metastasis from solid tumours: EANO-ESMO clinical practice guideline for diagnosis, treatment and follow-upESMO Open8101624202310.1016/j.esmoop.2023.10162437863528 HongYDuanPHeLLiQChenYWangPFuYLiuTDingZSystematic immunological level determined the prognosis of leptomeningeal metastasis in lung cancerCancer Manag Res1411531164202210.2147/CMAR.S34732335321403 Le RhunENayakLLim-FatMJRudàRPentsovaEForsythPO'BrienBJPreusserMKumthekarPBrandsmaDWellerMNANO-LM: An updated scorecard for the clinical assessment of patients with leptomeningeal metastasesNeuro Oncol27455465202510.1093/neuonc/noae17139208403 ZhangMTongJMaWLuoCLiuHJiangYQinLWangXYuanLZhangJPredictors of lung adenocarcinoma with leptomeningeal metastases: A 2022 targeted-therapy-assisted molGPA modelFront Oncol12903851202210.3389/fonc.2022.90385135795063 YinKLiYSZhengMMJiangBYLiWFYangJJTuHYZhouQZhongWZYangXNA molecular graded prognostic assessment (molGPA) model specific for estimating survival in lung cancer patients with leptomeningeal metastasesLung Cancer131134138201910.1016/j.lungcan.2019.03.01531027690 ChenJPanLLiuYFangYLiRLuZLiuAHeYZengZIntrathecal pemetrexed administration and myelosuppression in patients with leptomeningeal metastases from lung adenocarcinoma: A retrospective studyOncol Res3321072121202510.32604/or.2025.06423740746893 PicarelliHYamakiVNSollaDJFNevilleISSantosAGDFreitasBSAGDiepCPaivaWSTeixeiraMJFigueiredoEGThe preoperative neutrophil-to-lymphocyte ratio predictive value for survival in patients with brain metastasisArq Neuropsiquiatr80922928202210.1055/s-0042-175532436261127 LiZZhangWWuSPrognostic value of pre-treatment neutrophil-to-lymphocyte ratio and platelet-to-lymphocyte ratio in patients with brain metastasis from lung cancer: A systematic review and meta-analysisAltern Ther Health Med301081132024 WangJWYuanQLiLCaoKHLiuQWangHLHuKWuXWanJHRole of systemic immunoinflammation landscape in the overall survival of patients with leptomeningeal metastases from lung cancerOnco Targets Ther16179187202310.2147/OTT.S40238936993872 BerghoffASFuchsERichenGMlecnikBBindeaGSpanbergerTHacklMWidhalmGDieckmannKPrayerDDensity of tumor-infiltrating lymphocytes correlates with extent of brain edema and overall survival time in patients with brain metastasesOncoimmunology5e1057388201510.1080/2162402X.2015.105738826942067 HanahanDWeinbergRAHallmarks of cancer: The next generationCell144646674201110.1016/j.cell.2011.02.01321376230 NgMSFKwokITanLShiCCerezo-WallisDTanYLeongKCalvoGFYangKZhangYDeterministic reprogramming of neutrophils within tumorsScience383eadf6493202410.1126/science.adf649338207030 YuXLiCWangZXuYShaoSShaoFWangHLiuJNeutrophils in cancer: Dual roles through intercellular interactionsOncogene4311631177202438472320 ZhouLZhangZTianYLiZLiuZZhuSThe critical role of platelet in cancer progression and metastasisEur J Med Res28385202310.1186/s40001-023-01342-w37770941 DoltonGRiusCWallASzomolayBBianchiVGallowaySAEHasanMSMorinTCaillaudMEThomasHLTargeting of multiple tumor-associated antigens by individual T cell receptors during successful cancer immunotherapyCell18633333349202310.1016/j.cell.2023.06.02037490916 HeYLiuXWangMKeHGeCNeutrophil-to-lymphocyte ratio as a predictor of cardiovascular mortality in cancer survivorsSci Rep1420980202410.1038/s41598-024-72027-539251691 YuHTanLXueBFengLFangPMengXLuoXPlatelet-to-lymphocyte, neutrophil-to-lymphocyte and lymphocyte-to-monocyte ratios are related to cancer-related fatigue and quality of life in patients with cancer: A cross-sectional studyBMJ Open13e075398202310.1136/bmjopen-2023-07539838056937 BarhourABZakiPMcGranahanTMVenurVVellayappanBPalmerJHalaszLMYangJTBlauMTsengYDEmergent radiotherapy for brain and leptomeningeal metastases: A narrative reviewAnn Panlliat Med1214051419202310.21037/apm-22-1276 ZhenJWenLLaiMZhouZShanCLiSLinTWuJWangWXuSWhole brain radiotherapy (WBRT) for leptomeningeal metastasis from NSCLC in the era of targeted therapy: A retrospective studyRadiat Oncol15185202010.1186/s13014-020-01627-y32736566 ChiangCLYangHCLuoYHChenCJWuHMChenYMHuYSLinCJChungWYShiauCYLeptomeningeal metastasis in patients with non-small cell lung cancer after stereotactic radiosurgery for brain metastasisJ Neurosurg139385392202210.3171/2022.11.JNS22188836681978 LiuXLiGZhangHChangQFangMLuCTianPMeiFMolecular characteristics and prognostic factors of leptomeningeal metastasis in non-small cell lung cancerClin Neurol Neurosurg225107572202310.1016/j.clineuro.2022.10757236610238 TamuraKYoshidaTMasudaKMatsumotoYShinnoYOkumaYGotoYHorinouchiHYamamotoNOheYComparison of clinical outcomes of Osimertinib and first-generation EGFR-tyrosine kinase inhibitors (TKIs) in TKI-untreated EGFR-mutated non-small-cell lung cancer with leptomeningeal metastasesESMO Open8101594202310.1016/j.esmoop.2023.10159437517364 OkabeTOkamotoITamuraKTerashimaMYoshidaTSatohTTakadaMFukuokaMNakagawaKDifferential constitutive activation of the epidermal growth factor receptor in non-small cell lung cancer cells bearing EGFR gene mutation and amplificationCancer Res6720462053200710.1158/0008-5472.CAN-06-333917332333 ChoJChenLSangjiNOkabeTYonesakaKFrancisJMFlavinRJJohnsonWKwonJYuSCetuximab response of lung cancer-derived EGF receptor mutants is associated with asymmetric dimerizationCancer Res7367706779201310.1158/0008-5472.CAN-13-114524063894 ChenHYangSWangLWuYWuYMaSHeZZhangCLiuYTangHHigh-dose furmonertinib in patients with EGFR-mutated NSCLC and leptomeningeal metastases: A prospective real-world studyJ Thorac Oncol206575202510.1016/j.jtho.2024.09.138539260521

Kaplan-Meier curve analysis of overall survival among patients with lung adenocarcinoma and leptomeningeal metastasis in relation to age.

Kaplan-Meier curve analysis of overall survival among patients with lung adenocarcinoma and leptomeningeal metastasis in relation...

Kaplan-Meier curve analyses of overall survival among patients with lung adenocarcinoma and leptomeningeal metastasis in relation to Eastern Cooperative Oncology Group performance status.

Kaplan-Meier curve analyses of overall survival among patients with lung adenocarcinoma and leptomeningeal metastasis in relation to Eastern Cooperative Oncology Group performance ...

Kaplan-Meier curve analyses of overall survival among patients with lung adenocarcinoma and LM in relation to the receipt of brain radiotherapy after the LM diagnosis. LM, leptomeningeal metastasis.

Kaplan-Meier curve analyses of overall survival among patients with lung adenocarcinoma and LM in relation to the receipt of brain radiotherapy after the LM diagnosis. LM, leptomen...

Kaplan-Meier curve analyses of overall survival among patients with lung adenocarcinoma and leptomeningeal metastasis in relation to the gene mutation status. EGFR, epidermal growth factor receptor.

Kaplan-Meier curve analyses of overall survival among patients with lung adenocarcinoma and leptomeningeal metastasis in relation to the gene mutation status. EGFR, epidermal growt...

Kaplan-Meier curve analysis of overall survival among patients with lung adenocarcinoma and LM in relation to the receipt of receiving third-generation EGFR TKI therapy before the LM diagnosis. LM, leptomeningeal metastasis; EGFR, epidermal growth factor receptor; TKI, tyrosine kinase inhibitor.

Kaplan-Meier curve analysis of overall survival among patients with lung adenocarcinoma and LM in relation to the receipt of receiving third-generation EGFR TKI therapy before the ...

Kaplan-Meier curve analyses of overall survival among patients with lung adenocarcinoma and leptomeningeal metastasis in relation to NLR. NLR, neutrophil-to-lymphocyte ratio.

Kaplan-Meier curve analyses of overall survival among patients with lung adenocarcinoma and leptomeningeal metastasis in relation to NLR. NLR, neutrophil-to-lymphocyte...

Kaplan-Meier curve analyses of overall survival among patients with lung adenocarcinoma and leptomeningeal metastasis in relation to PLR. PLR, platelet-to-lymphocyte ratio.

Kaplan-Meier curve analyses of overall survival among patients with lung adenocarcinoma and leptomeningeal metastasis in relation to PLR. PLR, platelet-to-lymphocyte...

Kaplan-Meier curve analyses of overall survival among patients with lung adenocarcinoma and leptomeningeal metastasis in relation to extracranial metastasis.

Kaplan-Meier curve analyses of overall survival among patients with lung adenocarcinoma and leptomeningeal metastasis in relation to...

Kaplan-Meier curve analyses of overall survival among patients with lung adenocarcinoma and leptomeningeal metastasis in relation to molGPA. molGPA, molecular graded prognostic assessment.

Kaplan-Meier curve analyses of overall survival among patients with lung adenocarcinoma and leptomeningeal metastasis in relation to molGPA. molGPA, molecular graded prognostic...

Scoring standard of molGPA for LM.

molGPA for LM
Predictor 0 0.5 1
Karnofsky performance score <60 60-70 80-100
Extracranial metastasis Present Absent -
Epidermal growth factor receptor No sensitive mutation Sensitive mutation -

molGPA, molecular graded prognostic assessment; LM, leptomeningeal metastasis.

Clinical characteristics of patients (n=78).

Characteristic Patients
Median age, years (range) 64.0 (37–83)
Age, years, N (%)
  <60 39 (50.0)
  ≥60 39 (50.0)
Sex, N (%)
  Male 44 (56.4)
  Female 34 (43.6)
Smoking, N (%)
  Yes 21 (26.9)
  No 57 (73.1)
Eastern Cooperative Oncology Group performance status, N (%)
  0-2 62 (79.5)
  3-4 16 (20.5)
Chemotherapy before LM, N (%)
  Yes 42 (53.8)
  No 36 (46.2)
Brain radiotherapy before LM, N (%)
  Yes 12 (15.4)
  No 66 (84.6)
Brain radiotherapy after LM, N (%)
  Yes 24 (30.8)
  No 54 (69.2)
Brain metastasis, N (%)
  Yes 49 (62.8)
  No 29 (37.2)
Extracranial metastasis, N (%)
  Yes 58 (74.4)
  No 20 (25.6)
Neurological symptoms, N (%)
  Yes 69 (88.5)
  No 9 (11.5)
Gene mutation, N (%)
  EGFR 19 mutation 26 (33.3)
  EGFR 21 mutation 30 (38.5)
  EGFR wild-type 22 (28.2)
Third-generation EGFR tyrosine kinase inhibitor therapy before LM,
N (%)
  Yes 47 (60.3)
  No 31 (39.7)
Modality of LM diagnosis, N (%)
  MRI 27 (34.6)
  Cerebrospinal fluid cytology 20 (25.6)
  Both 31 (39.8)
Neutrophil-to-lymphocyte ratio (%)
  <7.5755 61 (78.2)
  ≥7.5755 17 (21.8)
Platelet-to-lymphocyte ratio, N (%)
  <156.035 35 (44.9)
  ≥156.035 43 (55.1)
Intrathecal therapy (%)
  Yes 36 (46.2)
  No 42 (53.8)
Molecular graded prognostic assessment (%)
  0.0 13 (16.7)
  0.5–1.0 58 (74.4)
  1.5–2.0 7 (8.9)

LM, leptomeningeal metastasis. EGFR, epidermal growth factor receptor.

Uni- and multivariate analyses for overall survival of patients with LM.

Univariate Multivariate
Characteristic N Median, months 95% CI P-value Hazard ratio 95% CI P-value
Age, years 0.029 0.844 0.474–1.502 0.563
  <60 39 13.0 6.96–19.04
  ≥60 39 6.5 3.00–10.00
Sex 0.742
  Male 44 7.6 1.68–13.52
  Female 34 10.0 5.79–14.21
Smoking 0.687
  Yes 21 9.1 0.00–22.96
  No 57 9.0 5.95–12.05
Eastern Cooperative Oncology <0.001 0.302 0.152–0.599 0.001
Group performance status
  0-2 62 12.1 4.97–19.23
  3-4 16 3.3 0.00–8.79
Chemotherapy before LM 0.314
  Yes 42 11.6 6.61–16.59
  No 36 7.6 3.25–11.95
Brain radiotherapy before LM 0.514
  Yes 12 5.8 4.18–7.42
  No 66 10.0 6.69–13.31
Brain radiotherapy after LM 0.012 0.509 0.253–1.026 0.059
  Yes 24 20.1 5.02–35.18
  No 54 6.5 2.73–10.28
Brain metastasis 0.260
  Yes 49 7.6 3.92–11.28
  No 29 16.0 9.51–22.49
Extracranial metastasis 0.012 2.291 1.074–4.888 0.032
  Yes 58 7.6 3.99–11.21
  No 20 16.0 5.47–26.53
Neurological symptoms 0.110
  Yes 69 7.6 3.58–11.62
  No 9 9.1 /
Gene mutation 0.019 1.050 0.711–1.552 0.806
  EGFR 19 mutation 26 16.1 14.64–17.56
  EGFR 21 mutation 30 5.8 1.34–10.26
  Without mutation 22 4.7 2.51–6.89
Third-generation EGFR tyrosine kinase inhibitor therapy before LM 0.035 0.958 0.478–1.920 0.904
  Yes 47 12.1 4.20–20.00
  No 31 5.9 2.39–9.41
Modality of LM diagnosis 0.666
  MRI 27 9.0 2.18–15.83
  Cerebrospinal fluid cytology 20 9.1 6.18–12.02
  Both 31 10.0 2.08–17.92
Neutrophil-to-lymphocyte ratio <0.001
  <7.5755 61 14.9 10.53–19.27
  ≥7.5755 17 1.4 0.86–1.94
Platelet-to-lymphocyte ratio <0.001 0.233 0.123–0.442 <0.001
  <156.035 35 16.4 10.56–22.24
  ≥156.035 43 4.4 3.01–5.80
Intrathecal therapy 0.880
  Yes 36 10.0 3.70–16.30
  No 42 8.8 4.25–13.35
Molecular graded prognostic assessment 0.008 0.625 0.268–1.458 0.277
  0.0 13 4.7 0.00–9.40
  0.5–1.0 58 8.8 4.61–12.99
  1.5–2.0 7 16.1 15.89–16.31

LM, leptomeningeal metastasis; OS, overall survival; EGFR, epidermal growth factor receptor.