构建基于血脂联合血常规相关参数的列线图诊断模型及其对尘肺病的诊断价值分析  

Diagnostic value of blood lipids combined with blood routine parameters for pneumoconiosis and the construction of nomogram prediction model

作　　者：周取 王维 王子萌 毛隆春 胡娟 李媛媛 于军丽 许商成 刘文彬 ZHOU Qu;WANG Wei;WANG Zimeng;MAO Longchun;HU Juan;LI Yuanyuan;YU Junli;XU Shangcheng;LIU Wenbing(Health Management Center,First Affiliated Hospital of Chongqing Medical and Pharmaceutical College,Chongqing 400060,China;Experimental Medicine Center,First Affiliated Hospital of Chongqing Medical and Pharmaceutical College,Chongqing 400060,China;Department of Public Health,First Affiliated Hospital of Chongqing Medical and Pharmaceutical College,Chongqing 400060,China;Hepatic Biliary and Pancreatic Cancer Center,Chongqing University Cancer Hospital,Chongqing 400030,China)

机构地区：[1]重庆医药高等专科学校附属第一医院健康管理中心,重庆400060 [2]重庆医药高等专科学校附属第一医院实验医学中心,重庆400060 [3]重庆医药高等专科学校附属第一医院公共卫生科,重庆400060 [4]重庆大学附属肿瘤医院肝胆胰肿瘤中心,重庆400030

出　　处：《国际检验医学杂志》2025年第8期965-970,975,共7页International Journal of Laboratory Medicine

基　　金：重庆市科卫联合医学科研面上项目(2022MSXM148);重庆市沙坪坝区科卫联合医学科研面上项目(2023SQKWLH018);重庆医药高等专科学校重点项目(ygzkt2023121)。

摘　　要：目的分析血脂和血常规相关参数在尘肺病患者中的情况,并构建列线图诊断模型探讨其对尘肺病的诊断价值。方法选取2022年1月至2024年1月重庆医药高等专科学校附属第一医院入院的456例尘肺病患者作为尘肺病组,以同期462例接触粉尘相关职业的体检健康者作为对照组。检测并比较两组血脂及血常规相关参数水平,通过单因素和多因素Logistic回归分析与尘肺病相关的血脂及血常规参数。基于机器学习中逻辑回归法构建列线图诊断模型,通过受试者工作特征(ROC)曲线计算C指数、Hosmer-Lemeshow拟合优度绘制模型校准曲线评估列线图诊断模型的诊断效能,并采用决策曲线分析(DCA)评估列线图诊断模型的临床实用性。结果尘肺病组血清高密度脂蛋白胆固醇(HDL-C)、总胆固醇(TC)、红细胞计数(RBC)、血细胞比容(HCT)、血红蛋白(Hb)、淋巴细胞数量(LYM)、淋巴细胞百分比(LYM%)水平低于对照组(P<0.05),中性粒细胞与淋巴细胞比值(NLR)、血小板与淋巴细胞比值(PLR)、全身免疫炎症指数(SII)水平高于对照组(P<0.05)。多因素Logistic回归分析结果显示,血中HDL-C、TC、LYM%、PLR和SII为尘肺病发生的独立影响因素(P<0.05)。将HDL-C、TC、LYM%、PLR和SII作为诊断因子构建尘肺病发生的列线图诊断模型,该诊断模型ROC曲线C指数为0.84(95%CI:0.81~0.86),诊断尘肺病的灵敏度为75.29%,特异度为77.51%,阳性诊断值为83.25%,阴性诊断值为67.88%。对构建的列线图诊断模型进行内部验证,验证集ROC曲线C指数0.84(95%CI:0.80~0.87),灵敏度为80.91%,特异度为72.62%,阳性诊断值为79.46%,阴性诊断值为74.39%。该诊断模型校准曲线斜率接近1,拟合度检验P>0.05。DCA分析发现该诊断模型对尘肺病发生的风险诊断具有临床实用价值。结论HDL-C、TC、LYM%、PLR和SII为尘肺病发生的独立影响因素,基于机器学习原理成功构建了尘肺病发生的列线图诊断模Objective To analyze the situation of blood lipid and blood routine parameters in patients with pneumoconiosis,and construct a column chart diagnostic model to explore their diagnostic value for pneumoconiosis.Methods A total of 456 patients with pneumoconiosis admitted to the First Affiliated Hospital of Chongqing Medical and Pharmaceutical College from January 2022 to January 2024 were selected as the pneumoconiosis group,while 462 healthy subjects exposed to dust during the same period were chosen as the control group.Serum lipids and blood routine parameters related to pneumoconiosis were measured and compared between two groups.Univariate and multivariate Logistic regression analyzes were conducted to examine serum lipids and blood routine parameters associated with pneumoconiosis.A risk prediction model was constructed using logistic regression in machine learning,and the diagnostic efficacy of the column chart diagnostic model was evaluated by calculating the C-index through receiver operating characteristic(ROC)curve and plotting the model calibration curve based on Hosmer Lemeshow goodness of fit.Decision curve analysis(DCA)was used to assess the clinical practicality of the column chart diagnostic model.Results The levels of serum high-density ester protein cholesterol(HDL-C),cholesterol(TC),red blood cell(RBC),hematocrit(HCT),hemoglobin concentration(HGB),lymphocyte number(LYM),and lymphocyte percentage(LYM%)in the pneumoconiosis group were lower than those in the control group(P<0.05).The levels of neutrophil-lymphocyte ratio(NLR),platelet-to-lymphocyte ratio(PLR),and systemic immune inflammation index(SII)were higher than those in the control group(P<0.05).Multivariate Logistic regression analysis showed that HDL-C,LYM%,PLR,and SII were independent influencing factors for pneumoconiosis(P<0.05).A column chart diagnostic model for the occurrence of pneumoconiosis was constructed using HDL-C,TC,LYM%,PLR,and SII as diagnostic factors.The ROC curve C-index of the diagnostic model was 0.84(95%CI:0.81-0.86

关 键 词：尘肺病 血脂 血常规 列线图诊断模型 

分 类 号：R135.2[医药卫生—劳动卫生]