Maintenance therapy for metastatic non-squamous non-small cell lung cancer: efficacy and biomarker analysis of pemetrexed discontinuation

Introduction

Immune checkpoint inhibitors (ICIs) have established a new paradigm for lung cancer treatment (1). Currently, a common standard first-line treatment for driver-gene-negative metastatic non-squamous non-small cell lung cancer (nsqNSCLC) is a combination of ICIs plus pemetrexed and platinum, followed by maintenance therapy with pemetrexed and ICIs, irrespective of programmed cell death ligand 1 (PD-L1) expression (1). Besides immunochemotherapy combinations, chemo-free ICIs immunotherapy has also demonstrated its efficacy in metastatic nsqNSCLC, especially for patients with PD-L1 tumor proportion score (TPS) ≥50% and age ≥75 years (2-4). However, in general, immunochemotherapy combination showed higher objective response rates (ORRs) and numerically longer survival, partially because of potential benefit of rapid disease control during the first few weeks, and synergistic effects between ICIs and chemotherapy (3). Another mode of immunochemotherapy combination as demonstrated in the CHECKMATE-9LA study is ICIs combined with short-term chemotherapy followed by ICIs maintenance (5). The idea is to provide early rapid disease control achieved by immunochemotherapy combination and durable survival benefit achieved by ICIs, while minimize the side effects of long-term chemotherapy (5).

As regard to maintenance therapy, in pre-immunotherapy era, pemetrexed has demonstrated its efficacy as a single agent or in combination with bevacizumab for metastatic nsqNSCLC (6). In current immunotherapy era, programmed death-1 (PD-1) or PD-L1 antibodies plus pemetrexed as maintenance therapy have also demonstrated superior efficacy compared to pemetrexed singlet maintenance in clinical trials, and also real-world patients not receiving maintenance therapy (7). However, long-term maintenance with pemetrexed and ICIs brings controversial issues. On one hand, long-term pemetrexed maintenance is associated with increased adverse effects including anemia, neutropenia, fatigue, and also increased medical costs (6). On the other hand, basic research has shown that pemetrexed activates T-cell immune responses and induces immunogenic cell death, and shows the synergistic effect with ICIs (8). Therefore, the question of whether dual drug maintenance therapy is necessary or excessive remains to be investigated.

Previously, a few retrospective studies have explored the impact of pemetrexed cessation in maintenance following induction immunochemotherapy on the prognosis and safety of metastatic nsqNSCLC (9-15). Results were inconsistent, but several studies showed that cessation of pemetrexed in the maintenance therapy had slight impact on the survival of metastatic nsqNSCLC (9,10,12,14,15). More surprisingly, in several studies, the median progression-free survival (PFS) and median overall survival (OS) for patients who stopped pemetrexed maintenance were numerically much longer than those remained on ICIs plus pemetrexed maintenance, although not statistically significant (9,14,15). The notable discrepancy in survival raises the concern of whether pemetrexed long-term maintenance in immunotherapy era has incremental benefit or worse effect.

However, the above studies showed several inherent differences in study designs, which partially explained the inconsistent results. Firstly, Shah et al. included patients treated with ICIs plus platinum-based chemotherapy as induction therapy, regardless of whether pemetrexed was included (9). This means their cohorts included patients who switched cytotoxic drugs during first-line treatment. Secondly, some studies did not exclude patients with concomitant bevacizumab in induction or maintenance phase (10,11). Thirdly, some studies enrolled patients with squamous subtype, for whom the standard first-line maintenance treatment did not involve pemetrexed plus ICIs, and had different natural course compared to nsqNSCLC (10,11). Moreover, several studies included patients with stage IIIB or IIIC disease or patients with driver mutations, the proportion of whom were incomparable between ICIs vs. ICIs plus pemetrexed maintenance group (9-11,15). All the above differences in study designs may confound the results and limit the interpretability of the conclusions for clinical practice.

Among the above studies, few investigated predictive biomarkers for selection of patients with benefit from pemetrexed cessation. Two studies respectively showed that patients with PD-L1 TPS <1% and patients without response to induction immunochemotherapy were more likely to benefit from ICIs plus pemetrexed rather than ICIs as maintenance (P=0.039 and P=0.0069, respectively) (10,12). Biomarkers to guide the decision making of pemetrexed discontinuation are lacking.

Therefore, this study aims to evaluate the efficacy of ICIs versus ICIs plus pemetrexed maintenance therapy for driver-gene-negative stage IV nsqNSCLC, focusing on identifying predictive biomarkers for differential benefits from continuation or discontinuation of pemetrexed maintenance. This study aims to optimize therapeutic efficacy while avoiding side effects and overmedication. We present this article in accordance with the STROBE reporting checklist (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-aw-1244/rc).

Methods

Patient selection and treatment protocol

Medical records of Fudan University Shanghai Cancer Center (FUSCC) from September 2020 to April 2024 were retrospectively searched. Patients diagnosed with stage IV nsqNSCLC treated with ICIs plus pemetrexed and platinum as first-line regimen in the Department of Medical Oncology were identified. Patients carrying EGFR, ALK or ROS1 driver mutations were excluded. A total of 167 eligible patients were identified, and 141 patients received maintenance therapy. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by Fudan University Shanghai Cancer Center Ethics Committee (approval No. 2025-Exp201). As a retrospective non-interventional study, individual consent for this analysis was waived in accordance with guidelines of FUSCC Ethics Committee.

Study design and endpoints

Patients were divided according to their maintenance status. Patients in ICIs + pemetrexed (n=101) group continued standard doublet maintenance, and patients in ICIs (n=40) group discontinued pemetrexed in any time during treatment courses permanently due to any reasons and continued ICIs maintenance. The primary endpoint was PFS, which is defined as the time from treatment initiation to the date of disease progression or death of any cause, regardless of which occurred first. Secondary endpoints included OS, ORR, and safety. OS was defined as the time period from treatment initiation to the date of death of any cause. The ORR is defined as the proportion of patients with complete response (CR) or partial response (PR).

Efficacy and safety assessments

Tumor response assessments were based on RECIST criteria (version 1.1). Adverse events (AEs) were documented according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE v5.0). AEs were collected based on a patient self-reporting system and by reviewing biochemical test results and medical records.

Biomarker analysis

A total of 33 patients in our cohort (11 from ICI maintenance group and 22 from ICI plus pemetrexed group) were previously enrolled in a prospective clinical trial in FUSCC to investigate the efficacy of extracellular vesicle (EV) -derived RNAs as biomarkers to predict antitumor efficacy of immunochemotherapy in nsqNSCLC (ClinicalTrials.gov identifier: NCT04427475). Briefly, plasma before immunochemotherapy was obtained after consent. EVs were isolated from 1 mL plasma using the exoRNeasy Serum/Plasma Kit (Qiagen, Cat. No. 77144, Hilden, Germany). For EV RNA extraction, the EVs were lysed directly using QIAzol (Qiagen) to extract EV-derived RNA, and total RNA was eluted with RNase-free water. RNA sequencing libraries were prepared using the SMARTer Stranded Total RNA sequencing Kit (Clontech, Palo Alto, California, USA) and sequenced on an Illumina sequencing platform (San Diego, California, USA). Raw sequencing reads were filtered using FastQC (version 0.11.8) and aligned to the Gencode human genome (GRCh38) with the STAR aligner (version 2.7.1a). Gene expression levels were quantified using FeatureCounts (Version 1.6.3) and converted to transcripts per kilobase million. Annotation data for mRNA, lncRNA, and pseudogene genes were obtained from the GENCODE database (Human, version 29). Differential expression analysis was performed using the limma R package, followed by KEGG pathway.

Statistical analysis

Propensity score matching (PSM) comparison was conducted using a 1:1 nearest-neighbor approach with a caliper width set to 0.02 of the standard deviation using R software (version 4.2.3). Matching was based on factors of age, smoking status, sex and PD-L1 status. Covariate balance was evaluated by comparing standardized mean differences (SMD) before and after matching, with an SMD of ≤0.2 considered indicative of adequate balance. Two-group comparisons were conducted using unpaired t-test or chi-square test. Survival analysis was performed using the Kaplan-Meier method, and P values were calculated by log-rank test. Hazard ratios (HRs) calculation and multivariate analysis were performed by COX regression analysis. All statistical analyses were conducted with SPSS version 19 or MATLAB. A two-tailed approach was used for all statistical tests, and a P value of less than 0.05 was considered statistically significant.

Results

Patient characteristics

Medical records of a total of 167 Stage IV nsqNSCLC patients treated with a combination of ICIs and pemetrexed and platinum as induction therapy in the Department of Medical Oncology, FUSCC were retrieved (Figure S1A). 26 out of 167 patients did not receive maintenance therapy, 25 were because of progression of disease, and 1 was because of patient’s preference. The remaining 141 patients received maintenance therapy, who were further divided into ICIs + pemetrexed group (n=101) and ICIs group (n=40) based on whether they discontinued pemetrexed permanently at any time during maintenance therapy due to any reasons other than progression disease. Baseline characteristics were showed in Table 1. The proportion of patients with brain metastasis were significantly higher in ICIs group than in ICIs + pemetrexed group (P=0.01). The remaining characteristics were balanced between two groups. The reasons of pemetrexed discontinuation of ICIs group were intolerable AEs (12/40, 30%) and patients’ or physicians’ choices owing to non-toxicity-related reasons (28/40, 70%). Because of the retrospective design, the factors contributing to non-toxicity-related patients’ or physicians’ choices were not fully documented. Reasons for patient-led discontinuation included refusal of further cytotoxic agents and difficulties in pemetrexed accessibility. Reasons for physicians-led discontinuation were based on the clinical opinion that ICI single-agent maintenance was sufficient to achieve satisfactory results.

Efficacy

Median PFS was 11.5 months for all patients and 14.4 months for patients receiving maintenance therapy. Further analysis showed no significant difference in median PFS between ICIs and ICIs + pemetrexed maintenance group (26.9 vs. 12.4 months, Log-Rank P=0.052) (Figure 1A). In addition, a PSM comparison was performed to adjust for potential biases namely age, smoking status, sex and PD-L1 status. A cohort of 74 patients was generated after 1:1 matching (37 in ICIs group and 37 in ICIs + pemetrexed group). Baseline characteristics of PSM-adjusted cohort was showed in Table S1. Besides smoking status, other characters between two maintenance groups were perfectly matched (Table S1). Consistently, no significant difference in median PFS between ICIs and ICIs + pemetrexed maintenance group was observed after PSM (26.9 vs. 20.2 months, Log-Rank P=0.41) (Figure 1B). In both original and PSM-adjusted cohort, the tendency of superior survival benefits in ICIs maintenance group were notable.

Figure 1 Progression-free survival. (A) Kaplan-Meier curves of PFS in the patients of ICIs maintenance group vs. ICIs + pemetrexed maintenance group before PSM. No statistically significant difference in median PFS was found between the two groups. (B) Kaplan-Meier curves of PFS in the patients of ICIs maintenance group vs. ICIs + pemetrexed maintenance group in PSM-adjusted cohort. No statistically difference in median PFS was found between the two groups. (C) Multivariate Cox regression analysis of factors associated with PFS. CI, confidence interval; HR, hazard ratio; ICIs, immune checkpoint inhibitors; PD-L1, programmed cell death ligand 1; PFS, progression-free survival; PR, partial response; PSM, propensity score matching; SD, stable disease.

The ORR was 61% (86/141) in patients receiving maintenance therapy, and comparable between ICIs and ICIs + pemetrexed maintenance group (P=0.09, Table 2). Multivariate Cox regression analysis showed number of metastatic organs and PD-L1 expression were independent biomarkers of PFS, while cessation of pemetrexed in maintenance not an independent biomarker (Figure 1C). The OS data were immature due to retrospective nature, but no significant differences were found between ICIs vs. ICIs + pemetrexed maintenance group (Figure S1B).

Safety

The detailed toxic profiles were provided in Table 3. AEs that occurred in more than 5 patients in either group were included (Table 3). In the ICIs maintenance group, the most frequent AEs are anemia, lymphopenia, thrombocytopenia and neutropenia. In ICIs + pemetrexed maintenance group, the most frequent AEs are anemia, lymphopenia, and liver injury. As for AEs of Grade 3 to 5, the most frequent AEs are neutropenia, anemia, and leukopenia in ICIs maintenance group, compared to neutropenia and lymphopenia in ICIs + pemetrexed maintenance group (Table 3). Safety profiles were comparative between the two maintenance groups, showing no notable advantage of pemetrexed discontinuation. However, considering the retrospective nature of this study, the higher occurrences of certain AEs in ICIs group may be the reasons for discontinuation of pemetrexed by physicians’ or patients’ choices, rather than simple demonstration of toxic profiles. Therefore, conclusions regarding the safety issue of pemetrexed discontinuation should be drawn cautiously. No new safety signals were observed.

Biomarker explorations

Further subgroup analysis of potential biomarkers for the cessation of pemetrexed was performed (Figure 2A). Results showed that patients under 65 years old (median PFS: 26.9 vs. 12.4 months, P=0.02), with smoking history (median PFS: 26.9 vs. 12.9 months, P=0.043) and with bone metastasis (median PFS: 42.4 vs. 12.1 months, P=0.03) showed significantly longer PFS with ICIs maintenance compared to ICIs + pemetrexed maintenance (Figure 2A). As PD-L1 ≥50% was a biomarker for first-line chemo-free pembrolizumab monotherapy, patients were divided into PD-L1 <49% and PD-L1 ≥50% subgroups. Interestingly, for both subgroups of different PD-L1 status, no significant difference in median PFS was observed between the two maintenance groups, suggesting PD-L1 low expression is not a negative biomarker for cessation of pemetrexed in maintenance (Figure 2A).

Figure 2 Subgroups analysis. (A) Forest plot of PFS hazard ratios. (B,C) Severe lymphopenia as a biomarker for superior survival benefits of pemetrexed cessation. Kaplan-Meier curves of PFS for patients with ≥ Grade 2 (B) or ≤ Grade 1 (C) lymphopenia before or during treatments. Patients experiencing ≥ Grade 2 lymphopenia before or during treatments showed significantly longer PFS with pemetrexed discontinuation. CI, confidence interval; HR, hazard ratio; ICIs, immune checkpoint inhibitors; PD-L1, programmed cell death ligand 1; PFS, progression-free survival; PR, partial response; SD, stable disease.

Because long-term pemetrexed was reported to cause lymphopenia (16), and lymphopenia was reported to be negatively associated with prognosis in NSCLC (17,18), we further explored peripheral blood index as biomarkers for pemetrexed cessation. Results showed that patients who experienced Grade 2 or higher lymphopenia before or anytime during treatment had significantly longer median PFS in ICIs maintenance group compared to ICIs + pemetrexed maintenance group (median PFS: 42.4 vs. 15.0 months, P=0.048, HR =0.30), while the survival difference was not significant between the two maintenance groups in patients experiencing Grade 1 or no lymphopenia (Figure 2A-2C). Subgroup analysis using the PSM-adjusted cohort was also performed. However, no significant biomarker for the cessation of pemetrexed was identified, partially because of the small sample size of the PSM-adjusted cohort (Figure S1C).

To identify plasma biomarkers to help decision making of pemetrexed discontinuation, RNA sequencing data of EV-derived RNA of 33 patients (11 from ICIs maintenance group and 22 from ICIs plus pemetrexed group) who were previously enrolled in other clinical trials were analyzed. We divided 11 patients of ICIs maintenance arm into two groups according to the duration of response (7 patients with PFS longer than median as long-survival group and others as short-survival group). Heatmap and volcano plots showed the differential expression genes between long- and short-survival group in patients with ICIs maintenance (Figure 3A,3B). Gene set enrichment analysis (GSEA) analysis showed differential genes enriched in pathways involving glycan biosynthesis, cell metabolism and cytokine signaling (Figure 3C). We identified low expression of EV-derived EHF gene as a biomarker for significantly prolonged PFS with pemetrexed discontinuation treatment (42.4 vs. 10.4 months, P=0.007, Figure 3D). Contradictory, in patients with ICIs plus pemetrexed maintenance, low expression of EHF gene has the tendency to correlate with worse survival, although analysis was not statistically significant (Figure S1D).

Figure 3 Exploratory analysis identified EV-derived EHF expression as a biomarker for pemetrexed discontinuation. (A) Heatmap showing differential expression genes between the long-survival group (n=7) and short-survival group (n=4) in the ICIs maintenance group. (B) Volcano plots of differentially expressed genes. (C) GSEA analysis showed differential genes enriched in glycan biosynthesis, cell metabolism and cytokine signaling pathways. (D) Patients with low expression of EV-derived EHF gene had significantly prolonged PFS with pemetrexed discontinuation (42.4 vs. 10.4 months, P=0.007). BgRatio, background ratio; EV, extracellular vesicle; FC, fold change; GSEA, gene set enrichment analysis; ICIs, immune checkpoint inhibitors; PPAR, peroxisome proliferator-activated receptor; PFS, progression-free survival.

Discussion

Immunotherapy-based regimens have become the cornerstone of NSCLC treatment. However, question about maintenance treatments remains controversial, especially for nsqNSCLC. Although pemetrexed has demonstrated its efficacy as maintenance therapy in pre-immunotherapy era, the role of its long-term maintenance remains unclear in immunotherapy era. Therefore, this study aimed to assess the impact of pemetrexed cessation during maintenance therapy on efficacy and safety in driver-negative metastatic nsqNSCLC. Furthermore, our study explored potential peripheral blood biomarkers for benefits of pemetrexed cessation for the first time to our knowledge.

Our study showed real-world maintenance treatment patterns for driver-negative metastatic nsqNSCLC in our center. Majority of the cohort complied with the protocols in clinical trial setting and received ICIs plus pemetrexed maintenance. The reasons of pemetrexed discontinuation for the remaining patients include patients’ or physicians’ choices and intolerable AEs, which were consistent with other studies (9,12,13). Previous reports showed that long-term pemetrexed treatments were associated with increased occurrences of fatigue, anemia, and renal dysfunction (9,19), but the incidence rates of anemia and renal dysfunction between ICIs and ICIs + pemetrexed maintenance groups in our cohort were comparable. The occurrence rate of all grade anemia was slightly higher in the ICIs + pemetrexed group, but the occurrence rate of severe anemia was much higher in ICIs group. We also observed numerically notably higher rates of all grade thrombocytopenia in the ICIs group, and numerically notably higher rates of all grade hypoalbuminemia in the ICIs + pemetrexed group. Pemetrexed discontinuation seems not advantageous considering the toxicity profile. However, we should not neglect the retrospective nature of this study, which implies that the occurrence of AEs may be the reasons for discontinuation of pemetrexed in the ICIs group.

Regarding efficacy, although the survival difference between two groups was not statistically significant in both original and PSM-adjusted cohort, the tendency of superior survival benefits in ICIs maintenance group were notable. Although majority studies including ours showed comparable efficacy for pemetrexed discontinuation, the superior survival tendency of pemetrexed maintenance in our study was consistent with some but contradictory with other previous reports (9-15). The variations in the results may be attributed to clinicopathological differences, sample sizes and treatment regimens of patients included. As stated before, our study strictly recruited stage IV driver-negative nsqNSCLC patients who received ICIs plus platinum and pemetrexed as induction therapy, followed by maintenance therapy, others varied in the aspects of induction therapy regimens, proportions of stage III patients, squamous subtypes and driver gene status (9-11,15). Therefore, our study represented the standard scenario of driver-gene-negative Stage IV nsqNSCLC treated with immunochemotherapy in the first-line, making our conclusions more precise and applicable.

Moreover, we explored biomarkers to facilitate physician’s decision on pemetrexed discontinuation in maintenance. Previously, Nakamura et al. reported that patients with stable disease to induction immunochemotherapy benefited more from ICIs plus pemetrexed compared to ICIs alone as maintenance (P=0.0069), while no significant differences in survival were observed between two maintenance groups for patients with PR or CR to induction immunochemotherapy. However, as previously stated, Nakamura et al.’s cohort included patients with squamous subtype (10). Gu et al. reported that patients with PD-L1 TPS <1% were more likely to benefit from ICIs plus pemetrexed as maintenance (P=0.039), while there were no differences in survival between the two maintenance groups for patients with PD-L1 TPS 1–49% or ≥50% (12). Because of the small sample size of PSM-adjusted cohort in our study, subgroup analysis was mainly performed in the original cohort.

Our study found no significant differences between ICIs and ICIs + pemetrexed maintenance groups in patients with different PD-L1 status and tumor response degrees. Interestingly, our results showed that patients aged under 65 years old, with smoking history, or with bone metastases demonstrated significantly improved survival with ICIs maintenance compared to ICIs + pemetrexed maintenance. Limited sample size may partially explain the inconsistency in subgroup analysis. The survival benefits gained from pemetrexed cessation in these subgroups may be partially explained by stronger immune competence in younger patients, higher mutation burdens in smokers, and potential bone marrow immune recovery from cytotoxic drugs as indicated in previous studies (20,21). The trend toward improved survival in patients without brain metastasis may be partially explained by the limited efficacy of ICIs due to immunosuppressed environments in brain (22). Collectively, these findings indicate that host factors and metastatic patterns may offer stratification value in guiding maintenance therapy decisions.

Besides the above clinicopathological biomarkers, we further explored other convenient biomarkers. Peripheral blood cell counts were postulated to represent changes in the tumor microenvironment (16-18). Peripheral lymphopenia, pre-existing or occurred during treatment, was found in more than 20% of patients with metastatic solid tumors (18). Lymphopenia during treatments was primarily caused by cytotoxic drugs or radiotherapy (18). Lymphopenia was shown to be related to poor response to treatment and poor prognosis in various cancers, and in patients with lymphopenia, the recovery to normal lymphocyte count is associated with responses to ICIs (18,23). Regarding NSCLC, KEYNOTE-189 study showed that the occurrence rates of Grade 3 to 4 lymphopenia were comparable between pembrolizumab plus chemotherapy group and placebo plus chemotherapy group (22% vs. 25%) (24). In KEYNOTE-042 and KEYNOTE-010 studies, the occurrence rates of Grade 3 to 4 lymphopenia in pembrolizumab monotherapy arm were notably much lower, 7% and 9% respectively (24). Although direct comparison of the absolute numbers between different studies was imprecise, the discrepancy of occurrence rates of lymphopenia among the above studies was too obvious to ignore, which indicated that chemotherapy attributes to the AE of lymphopenia more than ICIs. Meta-analysis showed that in patients treated with ICIs, lymphopenia at baseline or during treatment was associated with significantly worse survival (17). More specifically, Sumiyoshi et al. retrospectively investigated advanced nsqNSCLC patients who received pemetrexed-based therapy followed by maintenance without concomitant ICIs, and found that lymphopenia was a negative biomarker of survival (16). The above studies indicated the promising prognostic characters of lymphopenia in cancer treatment.

We focused on exploring lymphopenia as a biomarker for pemetrexed maintenance (16). Our results showed that patients who experienced Grade 2 or higher lymphopenia before or anytime during treatment had significantly longer median PFS in ICIs than ICIs + pemetrexed maintenance group (median PFS: 42.4 vs. 15.0 months, P=0.048, HR =0.30), while no significant survival difference was found between two maintenance groups in patients with < Grade 1 lymphopenia. Our results indicated the potential benefits of monitoring peripheral lymphocyte count.

Additionally, using baseline plasma to explore EV-derived gene expressions, we identified EV-derived EHF expression as another biomarker to guide decision-making of pemetrexed discontinuation. Emerging evidence indicates that EHF plays a critical role in innate immune system and the pathogenesis of cancers (25,26). The role of EHF in cancers appears context-dependent. In thyroid tumors, EHF is associated with aggressive clinical features, BRAF V600E mutation, and TERT promoter mutations or expression (27). Similarly, in cholangiocarcinoma, EHF remodels the tumor microenvironment by recruiting and activating tumor-associated macrophages, with its co-expression with GLI1 or CCL2 shows a worse prognosis (28). Conversely, loss of EHF during the progression of head and neck squamous cell carcinoma is observed, and patients with preserved EHF expression exhibit significantly better prognosis, suggesting that EHF may possess tumor-suppressive properties (29). In pancreatic cancer, EHF downregulation increases chemoresistance and immunotherapy resistance (30). We observed low EV-derived expression of EHF as a biomarker of more benefits from pemetrexed discontinuation in our study. Although mechanism remains unknown, this result may indicate the involvement of EHF in immune response in nsqNSCLC.

However, questions not solved by our study includes whether recovery from lymphopenia could be achieved by pemetrexed discontinuation, whether this recovery could generate improved prognosis and the optimal duration of pemetrexed maintenance. Due to the retrospective nature, it is difficult to completely trace the alterations in peripheral lymphocyte counts, further prospective validation is warranted. In addition, as a retrospective study, our research is limited by patient selection bias, omittance to report or document AEs by patients or physicians, and limited sample size. Nevertheless, the substantial absolute differences in survival time in our cohort suggests discontinuation of pemetrexed in maintenance is a promising strategy, especially when utilizing lymphopenia and plasma EV-derived EHF expression as biomarkers, to optimize efficacy, and minimize side effects and medical costs.

Conclusions

In conclusion, this study demonstrates that ICIs maintenance is comparably as effective as ICIs + pemetrexed maintenance in patients with metastatic nsqNSCLC without driver gene mutations who received first-line ICIs plus pemetrexed and platinum induction. We also proposed promising biomarkers to facilitate decision to discontinue pemetrexed in maintenance. Validation in larger and prospective clinical trials are needed, and the best timing for pemetrexed cessation awaits to be found.

Acknowledgments

The authors would like to thank the patients and their families for giving their time and commitment to participate in this study.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-aw-1244/rc

Data Sharing Statement: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-aw-1244/dss

Peer Review File: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-aw-1244/prf

Funding: This study was supported by Shanghai Anticancer Association EYAS Project (grant No. SACA-CY1B03).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-aw-1244/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 FUSCC Ethics Committee (approval No. 2025-Exp201). As a retrospective non-interventional study, individual consent for this analysis was waived in accordance with guidelines of FUSCC Ethics Committee.

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. Meyer ML, Fitzgerald BG, Paz-Ares L, et al. New promises and challenges in the treatment of advanced non-small-cell lung cancer. Lancet 2024;404:803-22. [Crossref] [PubMed]
2. Reck M, Rodríguez-Abreu D, Robinson AG, et al. Five-Year Outcomes With Pembrolizumab Versus Chemotherapy for Metastatic Non-Small-Cell Lung Cancer With PD-L1 Tumor Proportion Score ≥ 50. J Clin Oncol 2021;39:2339-49. [Crossref] [PubMed]
3. Akinboro O, Vallejo JJ, Nakajima EC, Ren Y, Mishra-Kalyani PS, Larkins EA, et al. Outcomes of anti–PD-(L)1 therapy with or without chemotherapy (chemo) for first-line (1L) treatment of advanced non–small cell lung cancer (NSCLC) with PD-L1 score ≥ 50%: FDA pooled analysis. J Clin Oncol 2022;40:9000.
4. Akinboro O, Larkins E, Pai-Scherf LH, et al. FDA Approval Summary: Pembrolizumab, Atezolizumab, and Cemiplimab-rwlc as Single Agents for First-Line Treatment of Advanced/Metastatic PD-L1-High NSCLC. Clin Cancer Res 2022;28:2221-8. [Crossref] [PubMed]
5. Paz-Ares L, Ciuleanu TE, Cobo M, et al. First-line nivolumab plus ipilimumab combined with two cycles of chemotherapy in patients with non-small-cell lung cancer (CheckMate 9LA): an international, randomised, open-label, phase 3 trial. Lancet Oncol 2021;22:198-211. [Crossref] [PubMed]
6. Paz-Ares L, de Marinis F, Dediu M, et al. Maintenance therapy with pemetrexed plus best supportive care versus placebo plus best supportive care after induction therapy with pemetrexed plus cisplatin for advanced non-squamous non-small-cell lung cancer (PARAMOUNT): a double-blind, phase 3, randomised controlled trial. Lancet Oncol 2012;13:247-55. [Crossref] [PubMed]
7. Aggarwal H, Bayo K, Han Y, et al. Real-world maintenance therapy and survival outcomes for pembrolizumab plus pemetrexed and platinum for non-small-cell lung cancer in USA. Immunotherapy 2023;15:267-81. [Crossref] [PubMed]
8. Schaer DA, Geeganage S, Amaladas N, et al. The Folate Pathway Inhibitor Pemetrexed Pleiotropically Enhances Effects of Cancer Immunotherapy. Clin Cancer Res 2019;25:7175-88. [Crossref] [PubMed]
9. Shah MV, Solem CT, Bell KF, et al. Real-World Maintenance Treatment Patterns Among Patients with Advanced Non-Small Cell Lung Cancer. Adv Ther 2025;42:2782-96. [Crossref] [PubMed]
10. Nakamura R, Yamada T, Morimoto K, et al. Impact of maintenance therapy following induction immunochemotherapy for untreated advanced non-small cell lung cancer patients. J Cancer Res Clin Oncol 2022;148:2985-94. [Crossref] [PubMed]
11. Li F, Wang H, Xiang Y, et al. Comparison of different maintenance regimens following first-line immunochemotherapy for advanced non-small cell lung cancer. Transl Lung Cancer Res 2023;12:2381-91. [Crossref] [PubMed]
12. Gu X, Shi Z, Shao L, et al. Efficacy and safety of maintenance immune checkpoint inhibitors with or without pemetrexed in advanced non-squamous non-small cell lung cancer: a retrospective study. BMC Cancer 2022;22:576. [Crossref] [PubMed]
13. Goldschmidt JH, Annavarapu S, Venkatasetty D, et al. Outcomes for pembrolizumab stratified by pemetrexed maintenance post pembrolizumab-platinum-pemetrexed induction in metastatic non-small-cell lung cancer. Immunotherapy 2024;16:453-64. [Crossref] [PubMed]
14. Gao M, Nie Y, Wang T, et al. Immune Checkpoint Inhibitor Without Pemetrexed for First-line Maintenance Therapy in Advanced Lung Adenocarcinoma: A Real-World Retrospective Study. Clin Lung Cancer 2025;26:e735-42. [Crossref] [PubMed]
15. Bhatia S, Lu M, Lessans S, et al. Brief Report: Impact of Maintenance Pemetrexed Cessation on Clinical Outcomes of Patients With Metastatic Nonsquamous NSCLC. JTO Clin Res Rep 2024;5:100717. [Crossref] [PubMed]
16. Sumiyoshi I, Okabe T, Togo S, et al. High lymphocyte population-related predictive factors for a long-term response in non-small cell lung cancer patients treated with pemetrexed: a retrospective observational study. J Transl Med 2021;19:92. [Crossref] [PubMed]
17. Zhang Y, Huang C, Li S. Influence of treatment-related lymphopenia on the efficacy of immune checkpoint inhibitors in lung cancer: a meta-analysis. Front Oncol 2023;13:1287555. [Crossref] [PubMed]
18. 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]
19. Dumoulin DW, Visser S, Cornelissen R, et al. Renal Toxicity From Pemetrexed and Pembrolizumab in the Era of Combination Therapy in Patients With Metastatic Nonsquamous Cell NSCLC. J Thorac Oncol 2020;15:1472-83. [Crossref] [PubMed]
20. Shah Y, Verma A, Marderstein AR, et al. Pan-cancer analysis reveals molecular patterns associated with age. Cell Rep 2021;37:110100. [Crossref] [PubMed]
21. Rizvi NA, Hellmann MD, Snyder A, et al. Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer. Science 2015;348:124-8. [Crossref] [PubMed]
22. Patel L, Kolundzic N, Abedalthagafi M. Progress in personalized immunotherapy for patients with brain metastasis. NPJ Precis Oncol 2025;9:31. [Crossref] [PubMed]
23. Damen PJJ, Kroese TE, van Hillegersberg R, et al. The Influence of Severe Radiation-Induced Lymphopenia on Overall Survival in Solid Tumors: A Systematic Review and Meta-Analysis. Int J Radiat Oncol Biol Phys 2021;111:936-48. [Crossref] [PubMed]
24. KEYTRUDA® (pembrolizumab) [package insert] Merck & Co., Inc. [cited Dec. 28th, 2025]. Available online: https://www.accessdata.fda.gov/drugsatfda_docs/label/2014/125514lbl.pdf
25. Yamazaki S, Nakano N, Honjo A, et al. The Transcription Factor Ehf Is Involved in TGF-β-Induced Suppression of FcεRI and c-Kit Expression and FcεRI-Mediated Activation in Mast Cells. J Immunol 2015;195:3427-35. [Crossref] [PubMed]
26. Appel S, Bringmann A, Grünebach F, et al. Epithelial-specific transcription factor ESE-3 is involved in the development of monocyte-derived DCs. Blood 2006;107:3265-70. [Crossref] [PubMed]
27. Xu Y, Gao J, Wang N, et al. BRAF-induced EHF Expression Affects TERT in Aggressive Papillary Thyroid Cancer. J Clin Endocrinol Metab 2025;110:693-705. [Crossref] [PubMed]
28. Luo Y, Li Z, Zhu H, et al. Transcription factor EHF drives cholangiocarcinoma development through transcriptional activation of glioma-associated oncogene homolog 1 and chemokine CCL2. MedComm (2020) 2024;5:e535.
29. Oyelakin A, Nayak KB, Glathar AR, et al. EHF is a novel regulator of cellular redox metabolism and predicts patient prognosis in HNSCC. NAR Cancer 2022;4:zcac017. [Crossref] [PubMed]
30. Xie Y, Zhou T, Li X, et al. Targeting ESE3/EHF With Nifurtimox Inhibits CXCR2(+) Neutrophil Infiltration and Overcomes Pancreatic Cancer Resistance to Chemotherapy and Immunotherapy. Gastroenterology 2024;167:281-97. [Crossref] [PubMed]