Development of a prediction model using preoperative immune-inflammatory cell features for persistent central diabetes insipidus 2 weeks after surgery in patients with craniopharyngioma

Introduction

Craniopharyngioma (CP) is an intracranial tumor originating from the sellar and parasellar regions, accounting for 1–3% of all primary brain tumors in adults and 5–10% in children (1). Although histologically benign, CP exhibits strong local invasiveness due to its proximity to the hypothalamic-pituitary axis. As a result, it frequently causes visual impairment, endocrine dysfunction, hydrocephalus, and elevated intracranial pressure, significantly compromising the patient’s quality of life (2-4). Surgical resection remains the primary treatment approach for CP, aiming for maximal tumor removal (2,5). However, the tumor’s anatomical location poses substantial surgical challenges and increases the risk of intraoperative injury to adjacent neuroendocrine structures (6). Consequently, surgical treatment is often associated with a high incidence of postoperative endocrine, metabolic, and behavioral complications (5,7). Among them, central diabetes insipidus (CDI) is one of the most common postoperative complications following CP surgeries (8).

CDI primarily results from damage to the arginine vasopressin secretion pathway. It is clinically characterized by polyuria, hypernatremia, and polydipsia, and in severe cases, can lead to life-threatening disturbances in water and electrolyte balance (9-12). Clinical data suggest that the incidence of postoperative CDI in CP patients exceeds 90%, and a considerable proportion experience transient CDI that resolves within approximately 2 weeks, while others progress to persistent CDI requiring long-term hormone replacement (8). Early detection of patients unlikely to recover is crucial for postoperative care and outcomes, as postoperative CDI has been associated with a 3.9-fold increase in patient mortality (13). Therefore, the preoperative identification of patients at high risk of developing persistent CDI is of great importance for optimizing perioperative management and improving patient outcomes.

Recent studies have shown that the occurrence of CP-related CDI is not only associated with mechanical injury but is also closely linked to the tumor-induced inflammatory microenvironment (14). In particular, CP cyst fluid contains high levels of pro-inflammatory cytokines, such as IL-6, TNF-α, and IL-1β. These inflammatory mediators can induce local hypothalamic neuroinflammation, thereby increasing the vulnerability of vasopressin-secreting neurons during surgical manipulation (15-18). In addition, systemic inflammatory status may further influence the recovery of the neuroendocrine axis after surgery. Systemic inflammation can disrupt the blood-brain barrier, promote microglial activation, and amplify neuroinflammatory signaling, thereby affecting neuronal survival and repair following surgical stress (19-21).

Although several indicators have been used in clinical practice to evaluate hypothalamic involvement in patients with CP, such as the Puget grading system, which assesses the anatomical relationship between the tumor and the hypothalamus based on preoperative imaging (22), these structural assessments mainly reflect anatomical risk and cannot capture the biological heterogeneity of host inflammatory and immune responses. Consequently, even among patients with similar imaging-based grades, postoperative endocrine outcomes may still vary substantially. In recent years, with growing research attention, the influence of systemic immune status on CP surgical outcomes is becoming increasingly recognized (23,24). Preoperative peripheral blood features reflecting immune activation and inflammatory responses may not only predispose patients to CDI but also influence the likelihood of timely recovery (25). These immune traits are readily obtainable and cost-effective, making them promising candidates for preoperative risk assessment. However, current studies in this field remain limited, especially regarding systematic exploration of immune-based predictors for CDI in CP patients.

This study aimed to investigate whether preoperative peripheral immune cell signatures can serve as predictive biomarkers for persistent CDI at 2 weeks after surgery in CP patients. We constructed a logistic regression model integrated with SHapley Additive exPlanations (SHAP) to enhance interpretability. By incorporating immune-related features into the clinical prediction model, this study aims to complement existing anatomy-based risk assessment systems and provide a biologically informed strategy for preoperative risk stratification. This approach not only contributes to understanding the immune underpinnings of postoperative endocrine dysfunction but also provides a practical tool and theoretical basis for identifying high-risk patients before surgery. This could facilitate personalized perioperative management with significant clinical value. We present this article in accordance with the TRIPOD reporting checklist (available at https://gs.amegroups.com/article/view/10.21037/gs-2025-1-562/rc).

Methods

Study population

In this study, clinical and preoperative laboratory data were collected from 252 patients with pathologically confirmed CP who received treatment at Xiangya Hospital of Central South University between 2016 and 2025. All patients included in the study underwent tumor removal through the same surgical route (subfrontal translamina terminalis approach), and the pituitary stalk was transected intraoperatively in each case. The inclusion criteria were as follows: (I) complete preoperative blood routine test results were available; (II) urine volume information, urine specific gravity, and serum sodium were documented; and (III) postoperative pathology confirmed the diagnosis of CP. Patients were excluded if they met any of the following criteria: (I) presence of polyuria before surgery; (II) missing essential clinical information; (III) lack of preoperative immune cell parameters; or (IV) presence of severe infection, immune system disorders, or other conditions that may influence peripheral blood cell counts before surgery. After applying these criteria, a total of 108 patients were included in the final analysis.

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Medical Ethics Committee of Xiangya Hospital, Central South University. Informed consent was waived in this retrospective study.

Variable definition and grouping

CDI was diagnosed based on the appearance of large volumes of dilute urine (urine output >3,000 mL/day, specific gravity <1.005) within 24 hours, accompanied by elevated serum sodium levels. Persistence of CDI was defined as the sustained presence of clinical diagnostic features for more than 2 weeks postoperatively. Recovery of CDI was defined as the resolution of polyuria and normalization of serum sodium and urine specific gravity within 2 weeks after surgery without the need for desmopressin. The demographic and clinical characteristics analyzed included gender, age (≤18 vs. >18 years), extent of tumor resection [gross total resection (GTR) vs. subtotal resection (STR)], tumor volume, and tumor subtype (Q-type, S-type, T-type). Preoperative inflammation-related hematological indices included white blood cells (WBCs), neutrophils, monocytes, platelets (PLTs), red blood cells (RBCs), hemoglobin, hematocrit, lymphocytes, and eosinophils.

Multivariate logistic regression analysis and predictive model construction

To evaluate the predictive value of preoperative inflammation-related parameters for persistent CDI at 2 weeks postoperatively, multivariate logistic regression analysis was first performed on variables that showed significant differences between the recovery and persistent CDI groups (P<0.05). Based on these significant variables, logistic regression models with different variable combinations were subsequently constructed using the “glm” function. The predictive performance of each model was assessed by plotting receiver operating characteristic (ROC) curves and calculating the area under the curve (AUC) using the pROC package. To validate the robustness of the model, we performed five-fold cross-validation on the entire dataset, repeated three times, resulting in 15 rounds to evaluate the model’s predictive performance. Additionally, decision curve analysis and calibration curve analysis were conducted using the rmda and rms R packages. Model stability and generalizability were further validated in gender- and age-stratified subgroups.

SHAP interpretation analysis

To further elucidate the relative importance and directional impact of inflammatory markers in predicting persistent CDI at 2 weeks postoperatively, SHAP analysis was performed using the kernelshap and shapviz packages. SHAP summary bar plots were used to assess the contribution of each variable to the model, while SHAP bee swarm plots were generated to visualize the direction and magnitude of individual feature values on the predicted probability of postoperative persistent CDI.

Mendelian randomization (MR) analysis

GWAS summary statistics for exposure variables were obtained from the IEU OpenGWAS project, including: WBC (ID: ieu-b-30), PLT (ID: ieu-a-1008), neutrophil (ID: ieu-b-34), and monocyte (ID: ieu-b-31). The outcome variable was CDI, based on data from the FinnGen cohort (ID: finn-b-E4_DIABINSIPIDUS). The analysis was performed using the TwoSampleMR R package, and instrumental variables were selected based on the following assumptions: relevance (significantly associated with exposure), independence (not confounded), and exclusion restriction (no direct effect on outcome). Five commonly used MR methods were employed: inverse-variance weighted (IVW), MR-Egger, weighted median, weighted mode, and maximum likelihood. Scatter plots of single-nucleotide polymorphism (SNP) effects were generated to aid in visualization and assess the stability of the results.

Statistical analysis

All statistical analyses were conducted using R version 4.5.0. Descriptive statistics, group comparisons, and regression analyses were performed using the autoReg package. Continuous variables were expressed as mean ± standard deviation, and categorical variables were presented as n (%). Between-group comparisons were conducted using Student’s t-test for continuous variables and the Chi-squared test for categorical variables. A two-tailed P<0.05 was considered indicative of statistical significance.

Results

Comparison of clinical and laboratory variables between the recovery and persistent CDI groups

A total of 108 patients were included in this study, comprising 59 patients (54.6%) in the recovery group and 49 patients (45.4%) in the persistent CDI group. Comparative analyses were conducted on baseline clinical characteristics and preoperative laboratory parameters between the two groups (Table 1). The proportion of female patients was higher in the persistent CDI group (53.1%) compared to the recovery group (35.6%). Individuals aged ≤18 years accounted for 40.8% of the persistent CDI group and 32.2% of the recovery group. However, differences in gender and age distribution between the groups were not statistically significant (both P>0.05). Regarding the extent of tumor resection, 65.3% of patients in the persistent CDI group underwent GTR, compared to 54.2% in the recovery group. STR was performed in 34.7% of the persistent CDI group and 45.8% of the recovery group. No significant differences were observed between the groups in terms of these variables. The mean tumor volume in the persistent CDI group was 48.5±54.5 mm³, slightly higher than that in the recovery group (42.6±34.3 mm³), but the difference was not statistically significant (P=0.51). In terms of tumor subtype, T-type tumors predominated in both groups, accounting for 57.1% in the persistent CDI group and 42.4% in the recovery group. For preoperative laboratory indicators, the WBC count was significantly higher in the persistent CDI group than in the recovery group (P=0.003). PLT count was also significantly elevated in the persistent CDI group (P=0.044). Furthermore, both neutrophil and monocyte counts were higher in the persistent CDI group compared to the recovery group (both P<0.05). No statistically significant differences were found between the two groups in other hematological parameters, including RBC count, hemoglobin, hematocrit, lymphocytes, and eosinophils (all P>0.05). These findings suggest that patients in the persistent CDI group exhibited significantly elevated levels of several inflammation-related biomarkers, such as WBC, PLT, neutrophils, and monocytes, indicating a potential association between preoperative inflammatory status and delayed recovery after CP surgery.

Predictive value of inflammatory indicators for postoperative persistent CDI in patients with CP

To evaluate the clinical utility of preoperative inflammation-related hematological parameters in predicting persistent CDI at 2 weeks after CP surgery, multivariate logistic regression analysis was first performed on WBC, PLT, neutrophil, and monocyte (Table 2). The results indicated that all four variables were independent risk factors for persistent CDI: WBC [odds ratio (OR) =1.759; 95% confidence interval (CI): 1.036–3.239; P=0.048], PLT (OR =1.850; 95% CI: 1.121–3.150; P=0.040), neutrophils (OR =2.301; 95% CI: 1.450–3.652; P=0.008), and monocytes (OR =1.621; 95% CI: 1.083–2.435; P=0.03). Based on these variables, a total of 15 logistic regression models incorporating between one and four predictors were constructed to evaluate their performance. The AUC was calculated for each model. Model performance was assessed in the overall cohort as well as in age- and gender-stratified subgroups, as shown in Table 3. Among the single-variable models, the WBC-based model demonstrated the highest predictive performance, with an AUC of 0.799 in the overall population. Its performance remained robust across subgroups, particularly among patients aged ≤18 years (AUC =0.792) and male patients (AUC =0.788). In contrast, PLT (AUC =0.681) and monocyte-based models (AUC =0.613) showed relatively lower predictive efficacy. Among the combined models, the four-variable model incorporating WBC, PLT, neutrophils, and monocytes exhibited the best overall performance, achieving an AUC of 0.944 in the total cohort (Figure 1A). Internal validation using repeated five-fold cross-validation further confirmed the robustness of the model, with consistently high AUC values across repetitions (Figure 1B). Decision curve analysis suggested that the model provided a favorable net clinical benefit across a wide range of threshold probabilities (Figure 1C). In addition, the calibration curve showed good agreement between predicted and observed probabilities, indicating satisfactory calibration of the model (Figure 1D). Subgroup analyses demonstrated that the four-variable model maintained strong predictive performance across different age and sex groups. The AUC was 0.840 (95% CI: 0.748–0.931) in patients aged ≤18 years and 0.820 (95% CI: 0.726–0.914) in those older than 18 years (Figure 1E). Similarly, the model achieved AUC values of 0.857 (95% CI: 0.745–0.969) in females and 0.882 (95% CI: 0.788–0.976) in males (Figure 1F), suggesting good generalizability and stability of the model across different patient subgroups.

Figure 1 ROC curves illustrate the superior predictive performance of the four-variable model. (A) Diagnostic performance of the four-variable model in the total cohort. (B) Internal validation using repeated five-fold cross-validation. (C) DCA indicated a favorable net clinical benefit of the model over a wide range of threshold probabilities. (D) Calibration curve showing good agreement between predicted and observed probabilities. (E) Subgroup analysis by age. (F) Subgroup analysis by gender. AUC, area under the curve; CI, confidence interval; DCA, decision curve analysis; ROC, receiver operating characteristic.

SHAP analysis reveals the relative importance of four inflammatory indicators in predicting postoperative persistent CDI

To further evaluate the relative contribution of inflammatory markers in predicting postoperative persistent CDI among CP patients, SHAP analysis was performed based on the previously constructed logistic regression model. Four key variables—WBC, PLT, neutrophil, and monocyte—were analyzed for feature importance. As shown in Figure 2A, SHAP summary bar plots ranked WBC as the most important predictor among all features, followed by PLT, while neutrophils and monocytes exhibited relatively lower importance. This ranking was consistent with the prior findings from logistic regression analysis, where WBC demonstrated superior predictive performance in both single-variable and combined models. Moreover, the SHAP bee swarm plot provided insights into the directional impact of each variable on model predictions (Figure 2B). The SHAP values for WBC and PLT increased significantly with rising original values, indicating that elevated levels of these markers contributed positively to persistent CDI prediction. These results suggest that a heightened preoperative systemic inflammatory state is more likely to be associated with the development of postoperative persistent CDI in patients with CP.

Figure 2 SHAP analysis reveals the feature importance of four inflammatory indicators in predicting persistent CDI. (A) Feature importance of WBC, PLT, monocyte, and neutrophil. (B) SHAP beeswarm plot for WBC, PLT, monocyte, and neutrophil. CDI, central diabetes insipidus; PLT, platelet; SHAP, SHapley Additive exPlanations; WBC, white blood cell.

Genetic evidence supporting the association between inflammatory traits and CDI susceptibility based on MR analysis

To explore whether systemic inflammatory traits may have a potential causal relationship with susceptibility to CDI, a two-sample MR analysis was conducted. The results indicated that at least two of the five MR methods demonstrated a significant positive association between elevated levels of the four immune cell traits (WBC, PLT, neutrophils, and monocytes) and an increased risk of CDI (P<0.05, Figure 3A). Scatter plots illustrating the SNP-specific effects on exposure and outcome are shown in Figure 3B, demonstrating consistent effect directions across instrumental variables and analytical approaches. These results provide genetic evidence suggesting that systemic inflammatory traits may contribute to an increased susceptibility to CDI-related neuroendocrine dysregulation.

Figure 3 MR analysis reveals the causal relationship of four inflammatory indicators with persistent CDI. (A) Relationship of WBC, PLT, monocyte, and neutrophil with persistent CDI using five MR methods. (B) SNP effect plots. CDI, central diabetes insipidus; CI, confidence interval; MR, Mendelian randomization; nSNP, number of SNP; OR, odds ratio; PLT, platelet; SNP, single-nucleotide polymorphism; WBC, white blood cell.

Discussion

CDI is one of the most common postoperative complications following CP and significantly affects patients’ postoperative recovery and long-term prognosis. Its occurrence is primarily attributed to damage to the hypothalamic-neurohypophyseal axis (26,27). Despite advances in surgical techniques in recent years, effective preoperative prediction and prevention of CDI remain major clinical challenges (28). This study systematically evaluated the potential value of preoperative peripheral blood inflammation-related immune cell profiles in predicting persistent CDI 2 weeks after CP surgery. The results showed that patients who developed postoperative persistent CDI had significantly higher preoperative levels of WBC, PLT, neutrophils, and monocytes compared to those who achieved recovery within 2 weeks. Further multivariate logistic regression analysis confirmed that all four indicators were independent risk factors for postoperative persistent CDI. A combined predictive model constructed based on these variables demonstrated excellent performance in the entire cohort (AUC =0.944) and exhibited good stability and generalizability across multiple subgroups. SHAP interpretation analysis further identified WBC as the most influential variable in the model. More importantly, MR analysis provided supportive genetic evidence suggesting that systemic inflammatory traits may be associated with CDI susceptibility.

In recent years, increasing attention has been paid to the interplay between systemic inflammation and neuroendocrine dysfunction (29). Studies have shown that systemic inflammatory responses can disrupt the blood-brain barrier, activate immune responses within the central nervous system, and interfere with hypothalamic-pituitary signaling, ultimately leading to disorders of water and electrolyte homeostasis (30,31). The CP tumor itself often contains cholesterol crystals and various inflammatory cytokines (e.g., IL-6, TNF-α, and IL-1β) in its cystic fluid, contributing to local neuroinflammation and increasing the vulnerability of critical neural structures during surgery (17,23,32). Previous studies have also reported cases of CDI induced by inflammatory infections in the pituitary region (33-35). Therefore, elevated preoperative inflammatory status may reflect both tumor-induced immune activation and the patient’s intrinsic immune background, which together may contribute to the development of postoperative persistent CDI.

In this study, patients in the persistent CDI group exhibited significantly higher levels of WBC, PLT, neutrophils, and monocytes compared to those in the recovery group. These parameters are recognized as key indicators of systemic inflammatory status. Elevated counts of WBC and neutrophils are classical markers of systemic inflammation and immune activation (36,37). Both neutrophils and monocytes can infiltrate the brain parenchyma during central nervous system injury (38). Neutrophils contribute to local tissue damage through the release of reactive oxygen species, proteases, and pro-inflammatory cytokines (39). Moreover, studies have indicated that neutrophil levels are associated with the long-term prognosis of CP patients (40). Monocytes, on the other hand, play a central role in orchestrating immune responses by secreting cytokines and presenting antigens (41). In addition to their traditional role in hemostasis, PLTs are also involved in inflammatory modulation. They promote leukocyte recruitment and endothelial activation by releasing platelet factor 4 (PF4) and expressing adhesion molecules such as P-selectin (42). Previous studies have shown that preoperative inflammatory markers such as WBC, neutrophils, and monocytes can aid in the differential diagnosis of CP (43). Our study further expands upon these findings by demonstrating their predictive value for the persistence of postoperative CDI. Logistic regression analysis confirmed that WBC, PLT, neutrophils, and monocytes were all independent risk factors for postoperative persistent CDI. Notably, WBC exhibited the highest predictive performance among individual markers (AUC =0.799) and was identified by SHAP analysis as the most influential variable in the combined prediction model. These findings suggest that WBC may serve as a valuable preoperative biomarker for detecting underlying neuroinflammation or central nervous system vulnerability that predisposes to persistent CDI.

More importantly, MR analysis, which leverages genetic variants as instrumental variables, provided genetic evidence supporting a potential association between these inflammatory traits and susceptibility to CDI. This approach effectively mitigates confounding and reverse causation inherent in traditional observational studies (44). The results not only enhance the causal inference strength of our study but also support a potentially direct pathogenic role of preoperative immune activation in the development of postoperative persistent CDI in CP patients. We hypothesize that systemic immune activation prior to surgery may render glial cells in the hypothalamic-pituitary region into a “pre-activated” state, where even mild mechanical irritation during surgery could trigger an exaggerated neuroinflammatory response. Additionally, sustained systemic inflammation may compromise the integrity of the blood-brain barrier, facilitating the entry of peripheral pro-inflammatory mediators into the central nervous system and disrupting the function of neurons responsible for antidiuretic hormone synthesis and secretion. Postoperatively, monocytes may infiltrate the hypothalamic region, release cytotoxic molecules, and participate in phagocytic processes, thereby exacerbating neuronal loss and dysfunction and ultimately initiating or aggravating the persistence of CDI. However, it should be noted that the MR analysis in this study was conducted using genome-wide association studies (GWAS) data from the general population rather than from patients with postoperative persistent CDI following CP surgery. Therefore, the results should be interpreted with caution and should not be considered as direct evidence of a causal mechanism underlying postoperative persistent CDI in CP patients. Nevertheless, these findings suggest that systemic inflammatory traits may increase susceptibility to CDI-related neuroendocrine dysfunction, which may partially explain the predictive value of inflammatory biomarkers observed in our clinical prediction model.

Advantages and limitations

This study is the first to integrate statistical modeling, machine learning analysis, and causal inference approaches to systematically evaluate potential immune predictors of persistent CDI after CP surgery. The study design is rigorous and offers high clinical translational value. The constructed predictive model provides a scientific basis for preoperative risk stratification, enabling early postoperative monitoring, timely intervention, and potentially reducing the risk of persistent CDI by modulating the inflammatory status preoperatively, thereby improving patient outcomes. Notably, all four indicators used in the model—WBC, PLT, neutrophils, and monocytes—are derived from routine preoperative complete blood counts, making the model simple, cost-effective, non-invasive, and highly suitable for widespread clinical implementation.

However, several limitations should be acknowledged. First, this is a single-center retrospective study, and the representativeness and external generalizability of the findings require further validation through multi-center prospective studies. Second, although patients with overt infections were excluded, the influence of subclinical inflammation on peripheral blood cell counts cannot be entirely ruled out, which may have introduced some degree of confounding. Finally, the potential causal mechanisms linking preoperative immune activation to intraoperative injury of the hypothalamic-neurohypophyseal axis remain to be elucidated. Further mechanistic studies at the molecular and cellular levels are needed to clarify this relationship.

Conclusions

This study demonstrates that elevated preoperative levels of WBC, PLT, neutrophils, and monocytes are significantly associated with an increased risk of persistent CDI at 2 weeks after CP surgery. The four-variable logistic regression model constructed from these parameters exhibits excellent predictive accuracy and robustness. Furthermore, MR analysis provides genetic evidence supporting a potential association between systemic inflammatory traits and susceptibility to CDI. These findings provide a novel perspective for preoperative risk assessment of postoperative persistent CDI and may contribute to advancing individualized and precision-based perioperative management strategies.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://gs.amegroups.com/article/view/10.21037/gs-2025-1-562/rc

Data Sharing Statement: Available at https://gs.amegroups.com/article/view/10.21037/gs-2025-1-562/dss

Peer Review File: Available at https://gs.amegroups.com/article/view/10.21037/gs-2025-1-562/prf

Funding: This study was supported by the National Natural Science Foundation of China (No. 82270825).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://gs.amegroups.com/article/view/10.21037/gs-2025-1-562/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 the Medical Ethics Committee of Xiangya Hospital, Central South University. Informed consent was waived in this retrospective study.

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. Alboqami MN, Khalid S, Albaiahy A, Bukhari BH, et al. Craniopharyngioma: A comprehensive review of the clinical presentation, radiological findings, management, and future Perspective. Heliyon 2024;10:e32112. [Crossref] [PubMed]
2. Karsonovich T, Shafiq I, Mesfin FB. Craniopharyngioma. In: StatPearls. Treasure Island: StatPearls Publishing; 2025.
3. Qin C, Hu W, Wang X, et al. Application of Artificial Intelligence in Diagnosis of Craniopharyngioma. Front Neurol 2021;12:752119. [Crossref] [PubMed]
4. Asha MJ, Oswari S, Takami H, et al. Craniopharyngiomas: Challenges and Controversies. World Neurosurg 2020;142:593-600. [Crossref] [PubMed]
5. Fujio S, Hanada T, Yonenaga M, et al. Surgical aspects in craniopharyngioma treatment. Innov Surg Sci 2021;6:25-33. [Crossref] [PubMed]
6. Diaz MJ, Kwak SH, Root KT, et al. Current Approaches to Craniopharyngioma Management. Front Biosci (Landmark Ed) 2022;27:328. [Crossref] [PubMed]
7. Lin AL, Tabar V, Young RJ, et al. Dabrafenib as a diagnostic and therapeutic strategy for the non-surgical management of papillary craniopharyngioma. Pituitary 2023;26:482-7. [Crossref] [PubMed]
8. Ghirardello S, Hopper N, Albanese A, et al. Diabetes insipidus in craniopharyngioma: postoperative management of water and electrolyte disorders. J Pediatr Endocrinol Metab 2006;19:413-21.
9. Weiner A, Vuguin P. Diabetes Insipidus. Pediatr Rev 2020;41:96-9. [Crossref] [PubMed]
10. Jameel PZ, Lohiya S, Vagha K, et al. Concurrent central diabetes insipidus and cerebral salt wasting disease in a post-operative case of craniopharyngioma: a case report. BMC Pediatr 2021;21:502. [Crossref] [PubMed]
11. Kavya S, Rajanna AH, Kishore N, et al. Central diabetes insipidus: A rare primary manifestation of small-cell lung carcinoma. J Family Med Prim Care 2024;13:4107-10. [Crossref] [PubMed]
12. Arima H, Azuma Y, Morishita Y, et al. Central diabetes insipidus. Nagoya J Med Sci 2016;78:349-58. [Crossref] [PubMed]
13. Almalki MH, Ahmad MM, Brema I, et al. Management of Diabetes Insipidus following Surgery for Pituitary and Suprasellar Tumours. Sultan Qaboos Univ Med J 2021;21:354-64. [Crossref] [PubMed]
14. Lozovanu V, Georgescu CE, Florescu LM, et al. Xanthogranuloma of the Sellar Region: A Comprehensive Review of Neuroimaging in a Rare Inflammatory Entity. J Pers Med 2022;12:943. [Crossref] [PubMed]
15. Grob S, Mirsky DM, Donson AM, et al. Targeting IL-6 Is a Potential Treatment for Primary Cystic Craniopharyngioma. Front Oncol 2019;9:791. [Crossref] [PubMed]
16. Webb LM, Okuno SH, Ransom RC, et al. Recurrent adamantinomatous craniopharyngioma stabilized with tocilizumab and bevacizumab: illustrative case. J Neurosurg Case Lessons 2025;9:CASE24410. [Crossref] [PubMed]
17. Zhao J, Yang Y, Pan Y, et al. Transcription Factor GLI1 Induces IL-6-Mediated Inflammatory Response and Facilitates the Progression of Adamantinomatous Craniopharyngioma. ACS Chem Neurosci 2023;14:3347-56. [Crossref] [PubMed]
18. Dixon BN, Daley RJ, Buie LW, et al. Correlation of IL-6 secretion and hyponatremia with the use of CD19+ chimeric antigen receptor T-cells. Clin Nephrol 2020;93:42-6. [Crossref] [PubMed]
19. Haruwaka K, Ikegami A, Tachibana Y, et al. Dual microglia effects on blood brain barrier permeability induced by systemic inflammation. Nat Commun 2019;10:5816. [Crossref] [PubMed]
20. Xingi E, Koutsoudaki PN, Thanou I, et al. LPS-Induced Systemic Inflammation Affects the Dynamic Interactions of Astrocytes and Microglia with the Vasculature of the Mouse Brain Cortex. Cells 2023;12:1418. [Crossref] [PubMed]
21. Schulz M, Salamero-Boix A, Niesel K, et al. Microenvironmental Regulation of Tumor Progression and Therapeutic Response in Brain Metastasis. Front Immunol 2019;10:1713. [Crossref] [PubMed]
22. Lin B, Xiang S, Chen J, et al. Assessment of quality of life in patients with craniopharyngioma and identification of risk factors for compromised overall wellness. Arch Endocrinol Metab 2023;68:e230001. [Crossref] [PubMed]
23. Jia Y, Ma L, Cai K, et al. Immune infiltration in aggressive papillary craniopharyngioma: High infiltration but low action. Front Immunol 2022;13:995655. [Crossref] [PubMed]
24. Gao Q, Luo J, Pan J, et al. Integrative analyses identify HIF-1α as a potential protective role with immune cell infiltration in adamantinomatous craniopharyngioma. Front Immunol 2022;13:949509. [Crossref] [PubMed]
25. Yuan F, Cai X, Zhu J, et al. A Novel Immune Classification for Predicting Immunotherapy Responsiveness in Patients With Adamantinomatous Craniopharyngioma. Front Neurol 2021;12:704130. [Crossref] [PubMed]
26. Di Iorgi N, Napoli F, Allegri AE, et al. Diabetes insipidus--diagnosis and management. Horm Res Paediatr 2012;77:69-84. [Crossref] [PubMed]
27. Angelousi A, Mytareli C, Xekouki P, et al. Diabetes insipidus secondary to sellar/parasellar lesions. J Neuroendocrinol 2021;33:e12954. [Crossref] [PubMed]
28. Devuyst F, Kazakou P, Balériaux D, et al. Central diabetes insipidus and pituitary stalk thickening in adults: distinction of neoplastic from non-neoplastic lesions. Eur J Endocrinol 2020;181:95-105. [Crossref] [PubMed]
29. Ouakinin SRS, Barreira DP, Gois CJ. Depression and Obesity: Integrating the Role of Stress, Neuroendocrine Dysfunction and Inflammatory Pathways. Front Endocrinol (Lausanne) 2018;9:431. [Crossref] [PubMed]
30. Bando H, Urai S, Kanie K, et al. Hypopituitarism: genetic, developmental, and acquired etiologies with a focus on the emerging concept of autoimmune hypophysitis. Endocr J 2025;72:649-62. [Crossref] [PubMed]
31. Mateska I, Witt A, Hagag E, et al. Succinate mediates inflammation-induced adrenocortical dysfunction. Elife 2023;12:e83064. [Crossref] [PubMed]
32. Zhou J, Zhang C, Pan J, et al. Interleukin-6 induces an epithelial-mesenchymal transition phenotype in human adamantinomatous craniopharyngioma cells and promotes tumor cell migration. Mol Med Rep 2017;15:4123-31. [Crossref] [PubMed]
33. Iwamoto Y, Mori S, Tatsumi F, et al. Central Diabetes Insipidus Due to IgG4-related Hypophysitis That Required over One Year to Reach the Final Diagnosis Due to Symptoms Being Masked by Sialadenitis. Intern Med 2022;61:3541-5. [Crossref] [PubMed]
34. Kawamoto S, Hatanaka K, Imakita M, et al. Central diabetes insipidus in an HHV6 encephalitis patient with a posterior pituitary lesion that developed after tandem cord blood transplantation. Intern Med 2013;52:1107-10. [Crossref] [PubMed]
35. Yamazaki M, Sato A, Nishio S, et al. Transient polyuria related to central diabetes insipidus caused by lymphocytic infundibulo-neurohypophysitis in a patient treated for Graves’ disease. Intern Med 2010;49:1885-90. [Crossref] [PubMed]
36. Mihlan M, Wissmann S, Gavrilov A, et al. Neutrophil trapping and nexocytosis, mast cell-mediated processes for inflammatory signal relay. Cell 2024;187:5316-5335.e28. [Crossref] [PubMed]
37. Arafa A, Kokubo Y, Kashima R, et al. Association Between White Blood Cell Count and Atrial Fibrillation Risk - A Population-Based Prospective Cohort Study. Circ J 2022;87:41-9. [Crossref] [PubMed]
38. Alam A, Thelin EP, Tajsic T, et al. Cellular infiltration in traumatic brain injury. J Neuroinflammation 2020;17:328. [Crossref] [PubMed]
39. Nguyen JN, Mohan EC, Pandya G, et al. CD13 facilitates immune cell migration and aggravates acute injury but promotes chronic post-stroke recovery. J Neuroinflammation 2023;20:232. [Crossref] [PubMed]
40. Zhang J, He M, Liu Z, et al. Impact of neutrophil-lymphocyte ratio on long-term outcome in patients with craniopharyngioma. Medicine (Baltimore) 2018;97:e12375. [Crossref] [PubMed]
41. Corps KN, Roth TL, McGavern DB. Inflammation and neuroprotection in traumatic brain injury. JAMA Neurol 2015;72:355-62. [Crossref] [PubMed]
42. Nagy Z, Vögtle T, Geer MJ, et al. The Gp1ba-Cre transgenic mouse: a new model to delineate platelet and leukocyte functions. Blood 2019;133:331-43. [Crossref] [PubMed]
43. Chen M, Zheng SH, Yang M, et al. The diagnostic value of preoperative inflammatory markers in craniopharyngioma: a multicenter cohort study. J Neurooncol 2018;138:113-22. [Crossref] [PubMed]
44. Sanderson E. Multivariable Mendelian Randomization and Mediation. Cold Spring Harb Perspect Med 2021;11:a038984. [Crossref] [PubMed]