Over the past decade, immune checkpoint inhibitors (ICIs) have significantly enhanced outcomes and increased survival for patients with various advanced solid tumors through blocking key immune regulators such as PD-1/PD-L1 and CTLA-4 (1, 2). The finding has had a significant impact on the fields of cancer treatment. Although ICIs increase anti-tumor immunity, they can also disrupt immune tolerance, which can lead to immune-related adverse events (irAEs). According to clinical evidence, patients receiving anti-CTLA-4 therapy may experience up to 90% of irAEs, while patients getting anti-PD-1/PD-L1 treatment had a cumulative frequency of 70% (3). The integumentary, gastrointestinal, respiratory, endocrine, and musculoskeletal systems are just a few of the organ systems in the human body that these irAEs might impact (4–6). Certain adverse effects may endure over time or emerge at a delayed stage, potentially resulting in a chronic disease trajectory (6).

One of the main areas of current research is the link between irAEs and clinical outcomes. According to the widely accepted theory, better clinical results may be associated with the incidence of irAEs. However, existing evidence indicates that this association is not universally applicable, exhibiting considerable heterogeneity across various studies and patient populations. Notably, irAEs affecting the endocrine system (7–9), such as thyroid dysfunction, and cutaneous irAEs demonstrate a higher incidence and a relatively consistent association with favorable prognoses (10). Conversely, certain irAEs, including immune checkpoint inhibitor-associated myocarditis (11) and severe pneumonia (12, 13), are linked to poorer prognoses, treatment interruptions, or diminished therapeutic efficacy. Moreover, the majority of studies have insufficiently accounted for the influence of toxicity severity within a single organ system in their association analyses. Recent investigations have additionally demonstrated that the long-term management of chronic irAEs is significantly correlated with clinical outcomes, indicating that these effects warrant thorough evaluation alongside the characteristics of disease progression (6). Mechanistically, the emergence of irAEs is hypothesized to be related to the induction of, or concurrent systemic, immune-inflammatory states. Importantly, specific peripheral blood markers that reflect the intensity of cancer-associated inflammation, such as the platelet/lymphocyte ratio (PLR), monocyte/lymphocyte ratio (MLR), and neutrophil/lymphocyte ratio (NLR), have been identified as independent predictors of the efficacy of ICIs monotherapy in solid tumors (14, 15). Nevertheless, the predictive value of these inflammatory markers for irAEs and their relationship with treatment efficacy remain inadequately explored and inconclusive. Current research predominantly focuses on cohorts receiving ICI monotherapy. Therefore, the predictive value of these inflammatory markers, along with their intricate associations with irAEs and treatment efficacy in advanced cancer patient populations undergoing ICI therapy, necessitates validation through more comprehensive and extensive large-scale studies.

Consequently, the clinical findings highlight the importance of conducting a thorough examination of the link between irAEs and inflammatory markers in relation to treatment efficacy. This study employed a complementary research design comprising two components: (1) a multicenter retrospective cohort study (N = 870) dynamically tracked the types, onset times, and severity of irAEs, exploring their associations with inflammatory markers, and use the time-dependent models to assess their impact on survival outcomes; and (2) a meta-analysis that integrated individual patient data from 38 studies (n=55,966) to quantify the moderating effects of factors such as tumor type and irAE occurrence through meta-regression and subgroup analysis. This dual validation framework not only enhances statistical efficacy but also elucidates the underlying factors driving the observed associations at the level of clinical characterization.

This investigation combines real-world cohort data with meta-analytic approaches to systematically elucidate the complex and multifaceted relationship between irAEs and patient survival outcomes. The findings demonstrate that the prognostic impact of irAEs is markedly influenced by both their severity and the specific organs affected: high-grade irAEs (greater than Grade 2) serve as independent adverse prognostic indicators. Organ-specific analyses reveal that hepatotoxicity and cardiotoxicity are associated with a significantly elevated risk of mortality, whereas skin toxicity and endocrine toxicity appear to confer a protective effect. Further stratified analyses indicate variability in these associations across distinct cancer types, such as non-small cell lung cancer and gastrointestinal malignancies. Moreover, post-treatment inflammatory biomarkers—including the MLR, PLR, and NLR—emerged as significant independent prognostic factors. Collectively, these results highlight the imperative of nuanced stratification of irAEs alongside systemic inflammatory status in clinical settings to facilitate the development of personalized prognostic evaluation frameworks.

The observed positive correlation between adverse effects and therapeutic efficacy may illustrate the “double-edged sword” nature of the biological effects of pharmacological agents. Utilizing a time-dependent Cox proportional hazards model, this study demonstrated that the incidence of irAEs is significantly associated with enhanced OS in patients. This finding aligns with the majority of existing literature, thereby reinforcing the hypothesis that irAEs may function as potential biomarkers indicative of ICI therapeutic efficacy (19–21). Regarding organ-specific irAEs, the findings of this study contribute significantly to the ongoing discourse within the existing body of evidence. Prior research has indicated that organ-specific irAEs exhibit heterogeneous effects on survival outcomes; specifically, endocrine and musculoskeletal irAEs are associated with survival benefits, while gastrointestinal, pneumonitis, neurologic, hepatic, renal, and other organ-specific irAEs negatively impact survival, suggesting that different organ-specific irAEs correlate with varying prognostic outcomes (22). The subgroup analysis within this study revealed that the incidence of endocrine toxicity was 25.2%, with primary clinical manifestations including hypothyroidism (21.6%), hyperthyroidism (2.4%), and hyperglycemia (0.9%), a distribution consistent with previous literature (23). This investigation, utilizing a time-dependent multivariable Cox regression analysis, substantiated that hepatotoxicity constitutes an independent prognostic factor for patient survival, thereby reinforcing evidence previously reported in the literature (24). Conversely, skin irAEs were linked to a significant extension of both OS and PFS, corroborating findings from earlier studies (25). The multivariate analysis conducted in this study did not reveal a statistically significant negative impact of pneumonia irAEs on patient survival. This outcome contrasts with existing literature, which presents conflicting views on this matter. One potential explanation for this discrepancy is that pneumonia irAEs, while potentially severe or fatal, may have adverse effects that are partially counterbalanced by the therapeutic advantages conferred by ICIs (26). This could elucidate the absence of significant survival detriments observed in the current study. Additionally, the inclusion of patients with advanced NSCLC, who frequently present with pre-existing lung conditions, may have further diminished the efficacy of ICIs. Moreover, the literature exhibits a notable divergence regarding the relationship between gastrointestinal irAEs (such as colitis or diarrhea) and survival benefits. Some studies report no significant survival advantage, while others indicate that these irAEs may be predictive of improved OS and PFS (27). This inconsistency may stem from variations in the baseline intestinal flora among patients (28, 29). However, the present study was unable to adequately evaluate this association due to the limited number of gastrointestinal irAEs documented. Furthermore, irAEs are hypothesized to signify a reactivation of the immune system, prompting an investigation into whether patients experiencing severe irAEs exhibit improved outcomes compared to those with mild irAEs. Consistent with our hypothesis, although the treatment did not significantly improve PFS, patients who developed grade >2 irAEs had shorter OS than those without, in agreement with prior reports (30, 31). This finding is particularly noteworthy as it contrasts with the results of several studies, which indicated that patients with severe irAEs (grades 3-5) had significantly enhanced PFS and OS compared to those with mild irAEs (30). This discrepancy suggests that the relationship between the severity of toxicity and survival benefit may be influenced by various factors, including the intensity of the therapeutic intervention and the underlying disease status (32).

Beyond adverse drug reactions, systemic inflammatory status plays a crucial role in modulating therapeutic responses in cancer treatment. Blood-based inflammatory indices (NLR, PLR, MLR) serve as established prognostic biomarkers for immune checkpoint inhibitors across malignancies including colorectal, NSCLC, and renal carcinoma (33–35). In advanced malignancies receiving anti-PD-1 treatment, inflammatory index dynamics are poorly defined. Our analysis revealed that NLR, PLR, and MLR were significantly correlated with OS. These findings align with observations reported in other oncological patient populations. Research by Ziting Ou et al. (35) further demonstrated that NLR, PLR, and MLR could assess short-term immunotherapy responses in anti-PD-1-treated gastric cancer cases, while Mingyu Wan et al. (36) established NLR as a robust predictor of PFS/OS in advanced GC patients receiving chemoimmunotherapy. Collectively, these findings underscore the cancer-specific dependence of the predictive value of inflammatory markers, necessitating careful individual consideration when selecting optimal biomarkers for personalized treatment. In the current investigation of the relationship between inflammatory markers and irAEs, the findings revealed no significant correlation, which diverges from the conclusions of prior research. Earlier studies have indicated that the NLR is markedly elevated in patients with NSCLC experiencing irAEs, and it has been identified as a potential biomarker due to its accessibility and cost-effectiveness (37). A similar trend has been observed in patients with cardiac adverse events, as evidenced by elevated NLR levels in individuals experiencing Major Adverse Cardiovascular Events, as reported by Moey et al (38). Additionally, elevated NLR and C-reactive protein levels have been associated with immune-related pneumonial (39). However, the present study did not replicate these associations, which may be attributed to the heterogeneity introduced by the multicancer cohort encompassing various subtypes of irAEs.

In interpreting the aforementioned findings, it is imperative to thoroughly consider the treatment context of the studied cohort, wherein the vast majority of patients (88.3%) received immune checkpoint inhibitors in combination with chemotherapy. This prevalent therapeutic approach introduces additional complexity to the elucidation of the relationship between irAEs and treatment efficacy. Specifically, multiple synergistic mechanisms exist between chemotherapy and immunotherapy: chemotherapeutic agents not only amplify and deepen T-cell responses via the induction of immunogenic cell death but also potentiate antitumor immunity by modulating pathways associated with the PD-(L)1 axis, such as the inhibition of STAT signaling and the augmentation of the cytotoxic T cell to regulatory T cell rati (40). Furthermore, evidence indicates that chemotherapeutic drugs such as pemetrexed can elicit systemic immune activation by promoting T-cell stimulation and increasing the susceptibility of cancer cells to immune-mediated cytotoxicity (41). These mechanisms may partially explain the observation that, within the context of combination therapy, certain mild to moderate irAEs (e.g., dermatologic and endocrine toxicities) remain associated with significant survival benefits. Clinical data support the potential advantages of combination regimens. For instance, a retrospective analysis of patients with PD-L1 expression ≥50% revealed no statistically significant difference in overall survival between the pembrolizumab monotherapy and combination therapy groups (13.3 months versus 20.4 months, P = 0.18). However, the combination therapy cohort showed a trend toward improved tumor response rates (42). Nonetheless, the use of combination therapy complicates the attribution of adverse events, as chemotherapy-induced toxicities (such as elevated hepatic enzymes and diarrhea) may overlap with clinical manifestations of irAEs, potentially resulting in underestimation or misclassification of irAE incidence. Consequently, the conclusions drawn from this study should be interpreted within the context of combination therapy as the prevailing clinical paradigm. Future prospective investigations are warranted to differentiate among treatment modalities, thereby elucidating the independent prognostic impact of irAEs in the settings of immune monotherapy versus combination therapy.

Underlying these clinical observations is a potential dual regulatory network involving systemic inflammation and immunotoxicity. Although the precise mechanisms of irAEs remain incompletely understood, there is a prevailing consensus that they represent a bystander effect of activated T cells, aligning with the action mechanism of ICIs: the antigen-specific T cells that are activated inadvertently result in the targeting of host tissues, accompanied by B-cell-mediated autoantibody production and a surge of inflammatory cytokines, ultimately resulting in organ-specific inflammation. Recent advances have further clarified the critical regulatory role of myeloid immune cells in irAEs. Notably, a post-mortem study of a patients with melanoma experiencing fulminant myocarditis following treatment with nivolumab and ipilimumab revealed significant T-cell infiltration of the heart muscle and pronounced M1 macrophage polarization (43). This finding directly demonstrates that PD-1 inhibitors can mediate cardiotoxicity by modulating macrophage polarization states (43). Additionally, the association between the NLR and ICI outcomes may stem from the pivotal role of neutrophils within the tumor microenvironment (44, 45). Neutrophils extensively infiltrate the tumor microenvironment (TME), and their excessive proliferation leads to the secretion of various pro-tumorigenic substances, promoting TME formation and influencing immune responses (46). Furthermore, neutrophils are strongly implicated in tumorigenesis, progression, and early metastasis (47).

This study is not without limitations. Its retrospective design potentially introduces selection and information biases. Although propensity score matching and time-dependent multivariable Cox regression models were utilized to mitigate the impact of known confounding factors, the possibility of residual confounding due to unmeasured variables cannot be entirely excluded. The study population is relatively homogeneous, and certain subgroup sample sizes are limited, which may compromise the stability and generalizability of the findings. Additionally, although the diagnosis of irAEs adhered to established guidelines, challenges persist in achieving consistent assessment of subjective symptoms. Moreover, the prevalent use of combination therapies further complicates the attribution of adverse events. The meta-analysis, encompassing 38 retrospective studies, may have reduced reliability due to heterogeneity in reporting and the lack of comparability in baseline characteristics across studies.

In summary, evidence derived from clinical trials and meta-analyses indicates that the range of organs affected, the severity of irAEs, and systemic inflammatory markers significantly impact the efficacy of immunotherapy. Notably, skin toxicity and endocrine toxicity are generally associated with a favorable prognosis, whereas cardiovascular toxicities, hepatotoxicity and irAEs of grade ≥3 correlate with a poorer prognosis. It is important to highlight that, although inflammatory markers possess independent predictive value regarding efficacy, their influence is not contingent upon the occurrence of irAEs. Given the limitations inherent in the current study, there is a pressing need for large-scale prospective investigations and an exploration of the underlying mechanisms to further elucidate the regulatory pathways governing irAEs and inflammatory responses in relation to immunotherapy efficacy. The objective of this study is to establish a robust evidence-based framework for enhancing risk stratification management of irAEs, ultimately aiming to improve the clinical outcomes for cancer patients.