Analysis of influencing factors and construction of predictive models of immune-related skin adverse events in urologic neoplasms
10.3760/cma.j.cn115396-20251014-00250
- VernacularTitle:泌尿系肿瘤免疫相关皮肤不良事件的影响因素分析与预测模型构建
- Author:
Ran SUN
1
;
Kai DANG
;
Yongan ZHOU
;
Yang YANG
;
Xiangyu WANG
;
Jinhua LIU
;
Jing XIAO
;
Teng CUI
Author Information
1. 首都医科大学附属北京安贞医院皮肤性病科，北京 100029
- Keywords:
Urologic neoplasms;
Immune checkpoint inhibitor;
Skin;
Adverse reaction;
Influencing factors;
Nomogram
- From:
International Journal of Surgery
2025;52(10):665-671
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To investigate the risk factors of skin adverse events associated with immune checkpoint inhibitor (ICI) therapy in patients with urologic neoplasms, and establish a predictive model.Methods:A single-center retrospective case-control study enrolled 91 advanced urologic neoplasms patients who received ICI therapy at the Department of Urology, Beijing Friendship Hospital, Capital Medical University from January 2020 to June 2025. Patients were divided into the skin lesion group ( n=44) and the control group ( n=47). Patients in the skin lesion group experienced related skin adverse events during ICI treatment, while patients in the control group did not experience such events during ICI treatment. The general data and laboratory indicators were compared between the two groups. The normally distributed measurement data were expressed as mean±standard deviation ( ± s), and the independent sample t-test was used for comparison between groups; the non-normally distributed measurement data were expressed as the median (interquartile range) [ M ( Q1, Q3)], and the Kruskal-Wallis test was used for comparison between groups. The count data were expressed as the number of cases and percentages, and the Chi-test was used for comparison between groups. First, a univariate analysis was conducted on the influencing factors of skin adverse events in patients with urologic neoplasms after ICI treatment. Then, the indicators with statistically significant differences in the univariate analysis were further included in the multivariate Logistic regression model to screen the independent risk factors for predicting skin adverse events. The R software was used to incorporate the factors with significant differences from multivariate analysis into the prediction model and construct a Nomogram. The calibration curve was utilized to evaluate the consistency between predicted values and actual observed results. Meanwhile, the discrimination of the model was verified by the receiver operating characteristic (ROC) curve and the area under the curve (AUC), so as to comprehensively verify the reliability and clinical application value of the prediction model. Results:The results of univariate analysis showed that there were statistically significant differences between the skin lesion group and the control group in terms of the proportion of other immune responses, serum albumin level, absolute eosinophil count, and C-reactive protein (CRP) levels ( P<0.05). These factors were included in multivariate Logistic regression, which identified elevated absolute eosinophil count and elevated CRP as the independent risk factors for related skin adverse events in patients with urologic neoplasms after ICI treatment. A predictive nomogram was built based on these factors. The calibration curve showed high consistency between predicted and actual probabilities, and ROC analysis confirmed the combined model had high predictive value (AUC=0.883, P<0.001). Conclusions:Elevated absolute eosinophil count and elevated CRP level are independent predictors of immune-related skin adverse events in urologic neoplasms patients after ICI treatment. The prediction model constructed based on these two factors facilitates early clinical screening and identification of high-risk patients.
