Lung cancer persists as the foremost contributor to cancer-related mortality worldwide, with non-small cell lung cancer (NSCLC) constituting approximately 85% of all diagnosed cases (1). Globally, an estimated 2.2 million new lung cancer diagnoses and 1.8 million deaths were reported in 2020 alone, highlighting the critical need for innovative and effective therapeutic interventions (2). Despite advancements in early detection methods and molecularly targeted therapies, a significant proportion of NSCLC patients present with advanced or metastatic disease at initial diagnosis, necessitating reliance on systemic chemotherapy as a cornerstone of treatment (3).

Historically, platinum-based doublets, such as cisplatin or carboplatin combined with taxanes like paclitaxel or pemetrexed, have served as the backbone of first-line therapy for NSCLC (4). Among taxanes, paclitaxel has been extensively utilized; however, its solvent-based formulation, which relies on polyethoxylated castor oil as a delivery vehicle, is frequently associated with hypersensitivity reactions and suboptimal tumor penetration due to poor biodistribution (5). The advent of nanotechnology in oncology drug delivery catalyzed the development of nanoparticle albumin-bound paclitaxel (nab-paclitaxel), a formulation designed to circumvent these limitations by leveraging endogenous albumin pathways for enhanced tumor-specific accumulation (6). In a landmark phase III clinical trial, nab-paclitaxel combined with carboplatin demonstrated superior objective response rates and a reduced incidence of neuropathy compared to its solvent-based counterpart, ultimately securing FDA approval as a first-line regimen for NSCLC (7).

The therapeutic paradigm for NSCLC has undergone further transformation with the integration of immune checkpoint inhibitors (ICIs), particularly those targeting the PD-1/PD-L1 axis. These agents disrupt critical immune evasion mechanisms employed by tumors, thereby restoring antitumor immunity and eliciting durable clinical responses in a subset of patients (8). Pivotal clinical trials, including KEYNOTE-189 and IMpower130, established the efficacy of combining ICIs such as pembrolizumab and atezolizumab with platinum-pemetrexed chemotherapy, demonstrating statistically significant improvements in overall survival (OS) and progression-free survival (PFS) compared to chemotherapy alone (9, 10). Despite these advances, the potential synergy between ICIs and nab-paclitaxel-platinum combinations remains inadequately explored. Preclinical studies suggest that nab-paclitaxel may exert immunomodulatory effects, such as enhancing dendritic cell activation and promoting tumor antigen presentation, which could synergize with PD-1/PD-L1 blockade to amplify antitumor responses (11). Early-phase clinical trials, exemplified by IMpower132, have reported promising outcomes with atezolizumab in combination with nab-paclitaxel and carboplatin, achieving a median PFS of 7.0 months compared to 5.7 months with chemotherapy alone (12).

Despite these advances, no comprehensive synthesis has evaluated the efficacy and safety of PD-1/PD-L1 inhibitors combined with nab-paclitaxel and platinum across multiple trials. Existing meta-analyses predominantly focus on solvent-based taxanes or pemetrexed-based regimens (13, 14). Our study addresses this gap by conducting a systematic review and meta-analysis of four randomized controlled trials (RCTs), aiming to quantify the clinical benefits and risks of this triplet therapy. By synthesizing data from 1998 patients, our findings offer valuable evidence to help guide clinicians in optimizing first-line treatment for NSCLC.

This report was designed and executed in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) (15, 16). As the study exclusively utilized aggregated data from publicly available published literature, ethical committee approval and individual patient consent were deemed unnecessary under current research ethics guidelines.

This meta-analysis systematically evaluated the efficacy and safety of PD-1/PD-L1 inhibitors combined with nab-paclitaxel and platinum-based chemotherapy in 1,998 patients with NSCLC across four RCTs, including both advanced and resectable disease settings. The pooled results demonstrated significant improvements in ORR, PFS and OS for the immunotherapy-containing regimen compared to chemotherapy alone. Additionally, while thrombocytopenia and immune-related AEs were more frequent in the experimental arm, other toxicities remained comparable between groups. These findings align with prior evidence supporting the integration of immune checkpoint inhibitors (ICIs) into first-line NSCLC treatment paradigms and extend their relevance to the perioperative setting (22).

The observed improvement in ORR (OR = 1.81) underscores the synergistic potential of combining ICIs with nab-paclitaxel-platinum chemotherapy. This enhanced antitumor activity may be attributed to nab-paclitaxel’s immunomodulatory effects, which are thought to create a more favorable tumor microenvironment for ICIs. Preclinical studies have demonstrated that nab-paclitaxel, more efficiently than solvent-based formulations, can promote the translocation of calreticulin to the cell surface, a key damage-associated molecular pattern (DAMP) signaling immunogenic cell death (ICD) (23). This, in turn, facilitates enhanced antigen uptake and presentation by dendritic cells, leading to a more robust priming of tumor-specific T-cell responses (24–26). Furthermore, clinical studies have observed a reduction in circulating immunosuppressive myeloid-derived suppressor cells (MDSCs) following nab-paclitaxel treatment, providing a plausible biological mechanism for the improved efficacy seen in our analysis (27). This multifaceted immunomodulation potentially synergizes with the mechanism of action of PD-1/PD-L1 inhibitors, which function by reversing T-cell exhaustion (24, 28).

All treatment regimens evaluated in the included trials were administered as first-line therapies for patients with advanced or resectable NSCLC. The PD-1/PD-L1 inhibitors were given concurrently with chemotherapy, not as a sequential or forced subsequent option. This design is pivotal, as concurrent administration maximizes the potential for synergistic immuno-chemotherapeutic effects by leveraging chemotherapy-induced immunogenic cell death to enhance the efficacy of simultaneous immune checkpoint blockade (1, 23, 28).

Notably, the efficacy of this combination extends beyond the metastatic setting. As demonstrated in the TD-FOREKNOW trial by Lei et al. (17), which exclusively enrolled patients with resectable stage IIIA/IIIB disease, the neoadjuvant use of camrelizumab plus nab-paclitaxel/platinum resulted in dramatically higher pathologic complete response (pCR) and major pathologic response (MPR) rates compared to chemotherapy alone (32.6% vs. 8.9% and 65.1% vs. 15.6%, respectively). This aligns with the findings from the landmark CheckMate 816 trial of neoadjuvant nivolumab plus chemotherapy (29), confirming that the synergy between immunotherapy and chemotherapy is potent in the perioperative setting, potentially leading to improved long-term survival outcomes by eradicating micrometastatic disease. The safety profile in this resectable population was manageable and did not preclude successful surgery, which is a critical consideration for neoadjuvant strategies (30).

The PFS benefit (HR = 0.65, 95% CI: 0.58–0.72) and OS advantage (HR = 0.81, 95% CI: 0.72–0.91) in the advanced disease trials further validate the durability of this combination. These findings are consistent with the established survival benefits of ICI-chemotherapy combinations seen in broader populations (3, 21). Our results align with a recent network meta-analysis by Zhao et al. (31), which compared various ICI-chemotherapy regimens and found that nab-paclitaxel-based combinations ranked highly for both PFS and OS in squamous NSCLC. Notably, subgroup analyses revealed stronger PFS benefits in patients with high PD-L1 expression, aligning with its established role as a predictive biomarker. However, the lack of a significant OS benefit in PD-L1-negative subgroups, combined with the considerable response observed in some of these patients, underscores the limitations of PD-L1 as a standalone biomarker and highlights the critical need for a more comprehensive, multi-factorial approach to patient selection (32).

While our analysis was necessarily focused on PD-L1—as it was the primary biomarker reported in the included trials—the field is rapidly evolving to encompass other promising biomarkers. Tumor mutational burden (TMB) serves as a powerful complement, predicting response to immunotherapy independent of PD-L1 status by reflecting neoantigen burden (28, 33, 34). Beyond genomic features, the tumor immune microenvironment (TME) provides crucial contextual information; the density and location of CD8+ tumor-infiltrating lymphocytes (TILs) are strong positive predictors of response (35), whereas upregulation of alternative immune checkpoints (LAG-3, TIM-3) can signify adaptive resistance mechanisms (36). Furthermore, systemic host factors, such as a derived neutrophil-to-lymphocyte ratio (dNLR), offer readily accessible proxies for an immunosuppressive state (37), and compelling evidence now links the composition of the gut microbiome to immunotherapy efficacy through modulation of host immunity (38). Future studies should integrate multifaceted biomarkers—including tumor-intrinsic, microenvironmental, and host-related factors—into unified models to optimally identify patients who will benefit most from this combination regimen.

In terms of safety, the increased risk of thrombocytopenia (RR = 1.83) and immune-related AEs (RR = 2.49) in the experimental group is consistent with prior meta-analyses of ICI-chemotherapy combinations (13, 39). However, the clinical implications of these findings extend beyond statistical significance and warrant careful consideration, as they directly impact treatment feasibility and adoption in real-world settings. Grade ≥3 thrombocytopenia can necessitate treatment delays, dose reductions, or even discontinuation, potentially compromising treatment intensity and efficacy. It also requires vigilant monitoring, interventions such as platelet transfusions, and increases the risk of bleeding events, thereby adding complexity and cost to patient management (40). The near 2.5-fold increased risk of immune-related adverse events (irAEs) presents an even greater challenge. irAEs often demand prompt recognition, specialist involvement, and prolonged courses of corticosteroids or other immunosuppressants, which are associated with their own morbidities and can diminish quality of life (41). The management burden of these toxicities significantly influences the real-world adoption of this regimen. In clinical practice, the successful and safe implementation of this combination is contingent upon institutional preparedness, including established management protocols and multidisciplinary support—resources that may not be uniformly available across all treatment centers (42). Our trial-based findings are further contextualized by real-world pharmacovigilance data. Analyses of freely available databases like the FDA Adverse Event Reporting System (FAERS) provide insights into the “real-life” incidence and nature of these ADRs. For example, a large-scale FAERS analysis not only confirms the signal for immune-mediated thrombocytopenia but also reveals a spectrum of associated clinical manifestations, such as severe bleeding events, and identifies drug-specific risk differences (43). Moreover, such databases are invaluable for detecting rare but fatal toxicities (e.g., ICI-associated myocarditis) that may be under-represented in clinical trials due to their limited sample size (44).

Therefore, while the therapeutic efficacy of this combination regimen is compelling, clinicians must engage in shared decision-making with patients by carefully weighing its significant survival benefits against the increased risks of specific and manageable yet clinically consequential toxicities, such as thrombocytopenia and immune-related adverse events. This highlights the necessity of not only biomarker-driven patient selection but also proactive toxicity monitoring and management protocols. It is important to note, however, that the overall safety profile features favorable aspects: the incidence of severe neutropenia did not differ significantly from that in the control group, and the low rate of severe neurotoxicity confirms the preserved safety advantage of nab-paclitaxel over solvent-based paclitaxel (7, 16, 45).

When considering the broader landscape of immune-combination therapies, our findings invite comparison with other established first-line regimens. The significant survival benefit observed here is comparable to that achieved with pembrolizumab plus pemetrexed/platinum in patients with non-squamous NSCLC in KEYNOTE-189 trail (10). This suggests that nab-paclitaxel-based triplets represent a potent alternative backbone for immunotherapy in NSCLC. Furthermore, the efficacy in squamous NSCLC, as demonstrated in the ASTRUM-004 trial, fills an important niche, as treatment options for this subtype have historically been more limited compared to non-squamous carcinomas (11) The consistency of benefit across different PD-1/PD-L1 inhibitors (atezolizumab, serplulimab, camrelizumab) supports a class-effect phenomenon for this combination strategy, reinforcing the robustness of this therapeutic approach (4, 6). The potential universality and promotion of this combination regimen may extend beyond the current scope. The rationale of combining taxane-based chemotherapy with immunotherapy could be applicable to other solid malignancies. For instance, similar regimens are being explored in aggressive cancers like triple-negative breast cancer (38). Future research should focus on head-to-head comparisons between nab-paclitaxel and other chemotherapy backbones (e.g., pemetrexed, paclitaxel) in combination with the same ICI to definitively establish superiority, if any.

Our study has several limitations that must be acknowledged. First and foremost, the clinical and methodological heterogeneity across the included trials necessitates a nuanced interpretation of our pooled results. The analyses encompassed a spectrum of NSCLC disease stages (from resectable to advanced metastatic disease), histological subtypes (squamous and non-squamous), and different PD-1/PD-L1 inhibitors. While this diversity increases the generalizability of the regimen’s activity, it also introduces complexity. This heterogeneity informs a more cautious and scenario-specific application of our findings. The impressive improvement in pathological complete response (pCR) and major pathological response (MPR) rates, driven exclusively by the neoadjuvant trial (17), robustly supports the efficacy of this regimen for inducing tumor downstaging in operable patients. However, the overall survival (OS) and progression-free survival (PFS) benefits are primarily derived from the metastatic setting and should not be directly extrapolated to predict survival outcomes in the resectable population, where long-term follow-up data are still maturing. Furthermore, although our subgroup analyses and low statistical heterogeneity (I²) for OS and PFS suggest a consistent treatment effect, the limited number of trials may underpower the detection of true differences among subgroups. Similarly, the observed “class effect” across different PD-1/PD-L1 inhibitors is encouraging but does not preclude the possibility of subtle differences in efficacy or safety profiles among specific agents.

Additionally, our work has several other important limitations. First, the number of included studies was relatively small (k=4), which limited the statistical power of some subgroup analyses and increases the uncertainty around our estimates. This limitation is compounded by the fact that the review protocol was not prospectively registered in a public registry (e.g., PROSPERO). Although we adhered to a pre-defined internal protocol and followed PRISMA guidelines rigorously, the lack of public registration introduces a potential for selective reporting bias, as it may raise concerns about the pre-specification of outcomes and analyses (46). Second, as appropriately noted in methodological literature, the assessment of publication bias in a meta-analysis with few studies is inherently underpowered and unreliable (47). While we employed conservative methods, the potential for unpublished negative studies cannot be ruled out. Third, model selection in this meta-analysis was primarily based on quantitative measures of statistical heterogeneity (the I² statistic and the p-value from Cochrane’s Q test). Although this provides an objective framework, we acknowledge that future analyses should, as suggested by the Cochrane Handbook, more comprehensively integrate considerations of clinical and methodological heterogeneity to guide model choice. Finally, we could not perform a detailed analysis based on the specific PD-1/PD-L1 inhibitor used, as the mechanisms, while overlapping, may not be identical. Some research suggests differences in binding epitopes and Fc receptor interactions might lead to varying clinical effects.

This meta-analysis reinforces the clinical value of PD-1/PD-L1 inhibitors combined with nab-paclitaxel-platinum chemotherapy in NSCLC, albeit with increased risks of thrombocytopenia and immune-related toxicities. These results advocate for the adoption of this triplet regimen as a first-line therapeutic option, provided clinicians implement proactive AEs monitoring.