本文已被:浏览 191次 下载 144次
Received:May 07, 2026 Published Online:October 23, 2026
Received:May 07, 2026 Published Online:October 23, 2026
中文摘要: 目的 分析血液肿瘤患者化疗所致味觉改变(CITA)的影响因素并构建风险预测模型,以期为该类患者提供化疗所致味觉改变的风险筛查工具。方法 选取2025 年4 月至2025 年12 月天津市某三级甲等医院血液科化疗患者301 例,收集其临床资料,通过单因素分析和多因素logistic 回归分析筛选危险因素,构建风险预测模型,并绘制列线图。结果 本研究共纳入301 例患者,其中有191 例发生CITA,发生率为63.46%。根据多因素logistic 回归分析结果,建立血液肿瘤患者发生CITA 的预测模型:logit(P)=-3.839+1.852×年龄29~49 岁+2.006×年龄50~70 岁+2.443×年龄≥71 岁+2.933×化疗2 周期+2.767×化疗3 周期+3.397×化疗4 周期+2.759×化疗5 周期+2.655×化疗≥6 周期+1.467×糖尿病+1.733×口干或口渴+1.294×中度口腔衰弱+2.605×重度口腔衰弱。模型曲线下面积(AUC)为0.906;Hosmer-Lemeshow 拟合优度检验结果提示模型拟合良好(χ2=2.073,P=0.979);Bootstrap 重复抽样1 000次后,乐观校正AUC 为0.884。结论 本研究构建的血液肿瘤化疗患者CITA 风险预测模型的预测效能良好,能够准确地识别出高风险人群,为临床预防和干预提供实用依据。
Abstract:Objective To analyze the influencing factors of chemotherapy-induced taste alteration(CITA)in patients with hematological malignancies,and construct a risk prediction model,so as to develop a risk screening tool for this adverse reaction. Methods A total of 301 patients receiving chemotherapy in the Department of Hematology of a tertiary A hospital in Tianjin from April 2025 to December 2025 were enrolled,and their clinical data were collected.Univariate analysis and multivariate logistic regression analysis were performed to screen independent risk factors,followed by the establishment of a prediction model and nomogram plotting. Results Among the 301 enrolled patients,191 cases developed CITA with an incidence rate of 63.46%. Based on the results of multivariate logistic regression analysis,a prediction model of CITA in patients with hematological tumors was established: logit(P)=-3.839+1.852×age 29-49 years+2.006×age 50-70 years+2.443×age≥71 years+2.933×2 cycles of chemotherapy+2.767×3 cycles of chemotherapy+3.397×4 cycles of chemotherapy+2.759×5 cycles of chemotherapy+2.655×6 cycles of chemotherapy+1.467×diabetes+1.733×xerostomia or thirst+1.294× moderate oral frailty+2.605×severe oral frailty. The area under the curve(AUC)of the model was 0.906. The Hosmer-Lemeshow test(χ2=2.073,P=0.979)indicated good model calibration.After 1 000 times of Bootstrap resampling,the bias-corrected AUC was 0.884. Conclusion The established risk prediction model exhibits favorable predictive performance for CITA in patients with hematological malignancies,which can accurately identify high-risk individuals and provide an evidence for clinical prevention and targeted intervention.
文章编号: 中图分类号:R733 文献标志码:A
基金项目:天津市医学重点学科建设项目(TJYXZDXK-3-003A);天津医科大学医院管理创新项目(2025YG-11)
引用文本:
