2026 年第 3 期 第 21 卷

基于白细胞计数和血小板计数构建非霍奇金淋巴瘤患者动员CD+34细胞采集效果的列线图模型

Construction of a nomogram model for CD+34 cell collection efficacy during mobilization in non-Hodgkin lymphoma patients based on white blood cell and platelet counts

作者：陈茜1韩好好1翁文秀1陈曦2

英文作者：Chen Qian1 Han Haohao1 Weng Wenxiu1 Chen Xi2

单位：1浙江省人民医院（杭州医学院附属人民医院）血液内科，杭州310000；2浙江省肿瘤医院血液内科，杭州310000

英文单位：1Department of Hematology Zhejiang Provincial People′s Hospital (People′s Hospital of Hangzhou Medical College) Hangzhou 310000 China; 2Department of Hematology Zhejiang Cancer Hospital Hangzhou 310000 China

关键词：非霍奇金淋巴瘤；白细胞计数；血小板计数；列线图模型；外周血造血干细胞采集

英文关键词：Non-Hodgkinlymphoma;Whitebloodcellcount;Plateletcount;Nomogrammodel;Peripheralbloodhematopoieticstemcellcollection

• 摘要：

• 目的 研究基于白细胞计数（WBC）和血小板计数（PLT）构建非霍奇金淋巴瘤（NHL）患者动员CD+34细胞采集效果的预测模型。方法 回顾性选取2023年1月至2025年1月于浙江省人民医院90例接受外周血造血干细胞（PBSC）采集的NHL患者，以单次采集CD+34细胞数量作为PBSC采集不良/失败标准，收集患者的临床资料和血液指标结果，分析NHL患者PBSC采集不良/失败的危险因素，基于多因素Logistic回归分析结果构建预测模型，并对构建的模型进行评估。结果 根据采集情况将90例NHL患者分为PBSC采集不良/失败组（31例）和采集成功组（59例）。采集不良/失败组患者化疗次数多于采集成功组，WBC、PLT、血红蛋白均低于采集成功组（均P＜0.05）。多因素Logistic回归分析显示，WBC、PLT均是PBSC采集不良/失败的独立保护因素，化疗次数为独立危险因素（均P＜0.05）。受试者工作特征曲线显示列线图模型的曲线下面积为0.93，对模型进行Hosmer-Lemeshow拟合优度检验，χ2=3.595，P=0.892。决策曲线结果显示列线图模型能产生更好的临床效益。结论 低WBC、低PLT、化疗次数多的NHL患者PBSC采集不良/失败风险更大，基于以上因素建立的列线图模型对PBSC采集不良/失败风险的预测价值良好。

• Objective To establish a predictive model for CD+34 cell collection efficacy during mobilization in patients with non-Hodgkin lymphoma (NHL) based on white blood cell count (WBC) and platelet count (PLT). Methods A total of 90 NHL patients who underwent peripheral blood stem cell (PBSC) collection at Zhejiang Provincial People′s Hospital from January 2023 to January 2025 were retrospectively selected. The number of CD+34 cells collected in a single apheresis was used as the criterion for poor/failed PBSC collection. Clinical data and blood index results of the patients were collected to analyze the risk factors for poor/failed PBSC collection in NHL patients. A predictive model was constructed based on the results of multivariate Logistic regression analysis, and the constructed model was evaluated. Results A total of 90 NHL patients were divided into the poor/failed PBSC collection group (31 cases) and the successful collection group (59 cases) according to the collection results. Compared with the successful collection group, the poor/failed collection group had a higher number of chemotherapy cycles and lower levels of WBC, PLT, and hemoglobin (all P<0.05). Multivariate Logistic regression analysis showed that WBC, PLT were independent protective factors for poor/failed PBSC collection, and number of chemotherapy was an independent risk factor (all P<0.05). The receiver operating characteristic curve indicated that the area under the curve of the nomogram model was 0.93. The Hosmer-Lemeshow goodness-of-fit test for the model yielded χ2=3.595 and P=0.892. Decision curve analysis demonstrated that the nomogram model could provide better clinical net benefits. Conclusion NHL patients with low WBC, low PLT, and more number of chemotherapy have an increased risk of poor/failed PBSC collection. The nomogram model established based on these factors exhibits good predictive value for the risk of poor/failed PBSC collection.