Lung cancer remains a leading cause of cancer-related mortality worldwide, with China experiencing persistently high incidence and mortality rates (1). Over the past few decades, immune checkpoint inhibitors (ICIs), which are antibodies that target inhibitory immune checkpoint molecules such as programmed cell death-1 (PD-1), programmed cell death ligand-1 (PD-L1), and cytotoxic T-lymphocyte antigen-4 (CTLA-4), have revolutionized cancer treatment. ICI therapies have demonstrated substantial clinical benefits, significantly improving overall survival in patients with advanced lung cancer and transforming the standard of care (2, 3). ICIs have largely replaced or are used in combination with conventional treatment modalities such as chemotherapy and radiotherapy. However, the expanding utilization of ICIs has led to a concomitant increase in the reported incidence of immune-related adverse events (irAEs) (4). IrAEs can manifest in any organ system, with the skin and gastrointestinal tract being among the most frequently affected sites. Among these irAEs, immune-mediated colitis (IMC) is highly prevalent (5).

Accurately determining the precise incidence and prevalence of IMCs is challenging, as these values vary considerably depending on the specific ICI treatment regimen and individual patient characteristics. Specifically, the highest incidence rates are reported in patients receiving CTLA-4 inhibitor monotherapy (3.4–15.5%), followed by those receiving combination therapy with CTLA-4 and PD-1 inhibitors (0.7–12.8%). The lowest incidence is observed in patients treated with PD-1/PD-L1 inhibitors (0.7–2.6%) (6). IMC onset can occur as early as one week after initial ICI administration, with a median time to onset of 4–8 weeks (7). Characterized by colonic mucosal inflammation, IMC presents with a heterogeneous spectrum of nonspecific clinical manifestations, including diarrhea (92%), abdominal pain (82%), hematochezia (64%), fever (46%), and vomiting (8). The severity of IMC is highly variable, ranging from mild diarrhea to severe complications such as bowel perforation and even mortality. Notably, a meta-analysis reported a 5% overall mortality rate associated with IMC (9). The heterogeneous clinical presentation and potential for rapid disease progression underscore the critical importance of early detection and timely intervention in cases of IMC.

Traditionally, a presumptive diagnosis of IMC is considered when a patient receiving ICI therapy experiences diarrhea and abdominal pain, supported by endoscopic and/or histopathological findings from gastrointestinal biopsies (10). However, the clinical, endoscopic, and histological features of IMC are nonspecific. Consequently, the traditional diagnostic approach for IMC relies on a process of elimination, requiring the exclusion of all other competing etiologies, including infectious colitis, drug-induced colitis, and graft-versus-host disease (GvHD) (11). This lengthy process can delay diagnosis, potentially exacerbating patient symptoms.

Therefore, a clinical model capable of rapidly identifying patients at risk of developing IMC is urgently needed. Nomograms are powerful graphical tools or algorithms that integrate biological and clinical variables, demonstrating wide applicability in oncology for risk factor identification and risk prediction (12). For example, Tong et al. utilized a nomogram model to investigate the incidence and predictive factors of prolonged chest tube drainage after video-assisted thoracoscopic surgery for the rapid prediction of this complication in clinical practice (13).

This study aims to leverage a nomogram algorithm to identify risk factors associated with IMC development in patients with lung cancer and to develop a robust risk prediction model. Ultimately, this model has the potential to complement traditional diagnostic approaches by enabling rapid risk stratification and informing personalized treatment and care decisions.

Cancer has emerged as the second leading cause of mortality worldwide. Concurrently, ICIs have revolutionized anticancer therapy, significantly improving survival rates in patients with lung cancer (19). However, the expanding indications and increased utilization of ICIs have been accompanied by a substantial increase in the incidence of immune-related adverse events (irAEs), with immune-mediated colitis (IMC) emerging as a prominent concern (20, 21). Identifying the risk factors associated with IMC and developing accurate risk prediction models are crucial steps toward optimizing treatment strategies and improving patient outcomes (22). This study integrated data on tumor characteristics, sociodemographic factors, and relevant hematological indices from a large cohort of patients with lung cancer receiving ICI therapy. This comprehensive analysis aimed to identify independent risk factors for IMC and develop a novel risk prediction model. The developed nomogram model demonstrated excellent discrimination and calibration, suggesting its potential for the rapid and accurate identification of patients at high risk of developing IMC. Importantly, this model relies solely on readily available clinical variables, eliminating the need for invasive procedures such as endoscopic or histopathological examinations. The simplicity, interpretability, and practicality of this model underscore its potential utility in routine clinical practice, facilitating personalized treatment decisions and potentially enhancing the safety of ICI therapy. Additionally, we have developed an online calculator based on the Nomogram model, which is more tailored to the practical circumstances in clinical practice. Clinicians can use the calculation results to provide personalized treatment options for patients, thereby minimizing their IMC risk.

Nomograms have gained increasing popularity in clinical practice because of their ability to transform complex regression equations into intuitive graphical representations, enhancing the interpretability of prediction models. For example, Wang et al. developed a nomogram based on readily available clinicopathological parameters to accurately predict lymph node metastasis in patients with endometrial cancer (23). Similarly, Zhang et al. constructed a user-friendly and cost-effective nomogram utilizing sociodemographic characteristics and routine blood parameters for the early prediction of gestational diabetes mellitus risk with promising accuracy (24). Our study builds upon these advancements by developing a nomogram model for predicting IMC risk in lung cancer patients receiving ICI therapy. On the basis of a multivariate logistic regression analysis, the nomogram integrates multiple predictors via scaled line segments plotted on a single plane, visually representing the interplay between these variables. This user-friendly tool empowers clinicians to rapidly estimate IMC risk in their patients. Furthermore, our nomogram demonstrated excellent discrimination, calibration, and clinical utility in internal validation, suggesting its potential to significantly improve upon traditional diagnostic approaches for IMC. By providing a more accessible and practical tool for clinicians and patients alike, this model has the potential to significantly enhance the management of this important ICI-related complication.

The mechanisms underlying the observed sex disparity in immune-mediated diseases remain an active area of scientific inquiry. Previous studies have revealed significant sex differences in immune responses, with females exhibiting greater susceptibility to certain conditions, including rheumatoid arthritis (RA) and inflammatory bowel disease (IBD) (25). Compared with their male counterparts, women with IBD experience greater X chromosome-related genetic predispositions, a more pronounced hormonal influence on disease activity, and diminished quality of life (26). Recent research has implicated sex-specific variations in the gut microbiome as a contributing factor to the increased incidence of various immune-mediated diseases, including IBD, in females (27). These findings suggest that the gut microbiome may play a role in the observed sex disparity. This heightened immune dysregulation in females may extend to IMCs, as evidenced by the present study’s finding of a 2.79-fold increased risk in women compared with men, a finding that is consistent with previous research (28). Furthermore, the present study revealed a significant association between SCLC and a 2.48-fold increased risk of developing IMC. In clinical practice, PD-L1 inhibitors are the preferred ICI regimen for SCLC, whereas PD-L1 inhibitors are more commonly used in NSCLC (29). Importantly, prior studies have demonstrated that the incidence of IMC varies depending on the specific ICI used (30). One review highlighted that patients receiving PD-L1 inhibitors exhibited a greater likelihood of diarrhea and other gastrointestinal adverse events than did those receiving PD-1 inhibitors (31). The heightened risk of IMC observed in patients with SCLC underscores the critical importance of tailored monitoring and preventive strategies within this patient population.

β2-MG, a low-molecular-weight protein released by activated T and B lymphocytes, has shown diagnostic and discriminatory value in immune-mediated diseases and serves as a potential marker of disease activity in IBD (32). Elevated β2-MG levels are widely recognized as risk factors for various immune-mediated diseases, including IMC (33, 34). Our findings align with these studies, highlighting a positive association between elevated β2-MG and increased IMC risk. Similarly, GLB, an important immune-related marker, plays a significant role in various immune-mediated diseases, including IBD. Previous research has established the GLB as a valuable prognostic biomarker for IBD, accurately reflecting disease severity and progression (35). Given the pathological similarities between IBD and IMC, we hypothesized that the GLB might also have predictive value for IMC risk. Our results support this notion, demonstrating a positive correlation between elevated GLB levels and increased IMC risk. The NLR, another potential predictor identified in our study, has been previously investigated for its diagnostic value in IBD. Marchesi et al. conducted a retrospective study and demonstrated that the NLR can serve as an independent predictor of IBD diagnosis (OR=1.80, P=0.016) (36). Similarly, in the context of IMC, an elevated baseline NLR has been significantly associated with IMC development and severity (2), which aligns with our findings.

This study identified several potential protective factors against IMC, including the WBC count, the CD4/CD8 ratio, and the PLR, which have not been extensively reported in previous studies. Leukocytes, essential components of the innate and adaptive immune systems, play crucial roles in combating infections and diseases. While elevated WBC counts are frequently observed in patients with IMC as a marker of systemic inflammation (37), the present findings suggest that higher WBC counts at baseline might be associated with a reduced risk of developing IMC. This potential protective effect could reflect a more robust baseline immune response capable of mitigating the inflammatory cascade associated with IMC development. The relationship between the CD4/CD8 ratio and the IMC remains less clear. Alterations in CD4/CD8 ratios have been linked to immune dysregulation (38) and potentially reflect the prognosis of IMCs (31). However, further large-scale prospective studies are needed to elucidate the complex and interconnected roles of the CD4/CD8 ratio in the context of IMC. PLR, an emerging inflammatory marker, has been associated with various disease states. Previous studies have suggested that PLR fluctuations can aid in assessing treatment response in IMCs (39). Additionally, Poenariu et al. demonstrated the predictive value of the PLR in ulcerative colitis (40), highlighting its potential as a biomarker in inflammatory bowel diseases. Our finding that a high PLR is associated with a reduced risk of IMC warrants further investigation to elucidate the underlying mechanisms involved.

This study has several strengths. First, it represents a large-scale investigation, the first of its kind in Southwest China, to explore risk factors and develop a prediction model for IMCs in lung cancer patients treated with ICIs. Our findings provide valuable insights and a framework for future research in this area. Second, the implementation of rigorous inclusion and exclusion criteria minimized potential biases arising from data heterogeneity.

However, it is important to acknowledge several limitations inherent to this study design. First, the single-center design of this study limits the generalizability and robustness of the prediction model because of the absence of external validation. Future multicenter studies are warranted to validate and refine the developed prediction model. Second, the model did not incorporate endoscopic findings, histopathological data from gastrointestinal biopsies, or genetic information. The inclusion of these variables could provide a more comprehensive assessment of patient characteristics and potentially enhance the model’s predictive performance. Future iterations of the model should prioritize the incorporation of these variables. Finally, the inherent limitations of retrospective studies, such as recall bias and information bias, cannot be completely mitigated. In the future, we will design a prospective cohort study to address these inherent limitations and validate the clinical applicability of the model.

This study identified eight independent risk factors associated with IMC development in patients with lung cancer receiving ICI therapy. We developed and validated a nomogram prediction model for IMC risk stratification on the basis of these factors. The nomogram model demonstrated excellent discrimination, calibration, and clinical utility, suggesting its potential as a readily applicable tool for the rapid and accurate assessment of IMC risk in routine clinical practice. The identification of high-risk individuals via this model could facilitate personalized treatment decisions, potentially enabling timely interventions to mitigate IMC development and enhance the safety profile of ICI therapy in this patient population.