无功能垂体大腺瘤神经内镜下经蝶窦切除术后垂体前叶功能减退发生风险预测模型构建与验证  

Construction and validation of a nomogram model for predicting hypopituitarism after endoscopic transsphenoidal resection of non-functional pituitary macroadenoma

作　　者：王斐[1] 符秀舒 陈开宁[1] 全会标[1] WANG Fei;FU Xiushu;CHEN Kaining;QUAN Huibiao(Department of Endocrinology,Hainan Provincial People's Hospital,Haikou China)

机构地区：[1]海南省人民医院海南医科大学附属海南医院内分泌科,海南海口570311 [2]海口市第三人民医院内分泌科,海南海口571100

出　　处：《山东医药》2025年第1期19-23,共5页Shandong Medical Journal

基　　金：海南省卫生健康行业科研项目(22A200002)。

摘　　要：目的构建预测无功能垂体大腺瘤(NFPMA)神经内镜下经蝶窦切除术后垂体前叶功能减退(HAP)发生风险的列线图模型,并验证该模型的效能。方法选择接受神经内镜下经蝶窦切除术的NFPMA患者103例,术后1年复查垂体前叶功能,HAP 38例(HAP组)、垂体前叶功能正常65例(NHAP组)。收集所有研究对象术前一般资料、实验室检查资料、临床症状、影像学资料、手术并发症以及术后病理资料,采用单因素和多因素Logistic回归模型分析NFPMA神经内镜下经蝶窦切除术后发生HAP的独立危险因素。根据其独立危险因素构建预测NFPMA神经内镜下经蝶窦切除术后HAP发生风险的列线图模型,并通过受试者工作特征(ROC)曲线验证该列线图模型的预测效能。结果基于单因素和多因素Logistic回归模型筛选出4个NFPMA神经内镜下经蝶窦切除术后发生HAP的独立危险因素:糖化血红蛋白(HbA1c)、垂体最大径、侵袭性评分、中枢性尿崩症,构建NFPMA神经内镜下经蝶窦切除术后HAP发生风险的Logictic回归方程:logit(P)=-6.309+0.333×HbA1c+0.817×垂体最大径+0.532×侵袭性评分+3.402×中枢性尿崩症(有=1,无=0),并通过R4.2.0软件构建其列线图模型;ROC曲线分析发现,该列线图模型预测NFPMA神经内镜下经蝶窦切除术后发生HAP的曲线下面积为0.842(95%CI:0.763~0.920),其预测敏感度为81.6%、特异度为76.9%。结论HbA1c、垂体最大径、侵袭性评分、中枢性尿崩症是NFPMA神经内镜下经蝶窦切除术后HAP发生的独立危险因素,基于上述独立危险因素构建的列线图模型可以较好地预测NFPMA神经内镜下经蝶窦切除术后HAP的发生风险。Objective To establish the nomogram model for predicting the risk of hypopituitarism(HAP)after endoscopic transsphenoidal resection of non-functional pituitary macroadenoma(NFPMA),and to verify the efficacy of this model.Methods Totally 103 patients with NFPMA who underwent endoscopic transsphenoidal resection were selected.According to the pituitary function one year after surgery,they were divided into the HAP group(n=38)and normal anterior pituitary function group(n=65).Basic information,medical history,blood lipids,blood glucose results,compression symptom scores,endocrine symptom scores,pituitary-target gland hormone levels,pituitary imaging characteristics including tumor size and invasiveness scores,postoperative complications,and pathological features were collected for analysis.Univariate and multivariate Logistic regression models were used to analyze the risk factors for HAP after endoscopic transsphenoidal resection of NFPMA.We constructed a nomogram model based on the risk factors of HAP after endoscopic transsphenoidal resection of NFPMA.The predictive performance of the nomogram model was using the receiver operating characteristic(ROC)curve.Results Four risk factors of HAP after endoscopic transsphenoidal resection of NFPMA were selected based on the univariate and multivariate Logistic regression models:HbA1c,diameter of the pituitary,invasive score,and central diabetes insipidus.The Logictic regression equation for predicting the risk of HAP after endoscopic transphenoidal resection of NFPMA was:logit(P)=-6.309+0.333×HbA1c+0.817×maximum diameter of the pituitary+0.532×invasive score+3.42×central diabetes insipidus(yes=1,no=0).We established its nomogram model through the R4.2.0 software.The area under the ROC curve was 0.842(95 CI:0.763–0.920),with the sensitivity of 81.6%and the specificity of 76.9%.Conclusions HbA1c,maximum diameter of the pituitary,invasive score,and central diabetes insipidus are risk factors of HAP after endoscopic transsphenoidal resection of NFPMA.The nomogram model

关 键 词：无功能垂体大腺瘤 垂体前叶功能减退 经蝶窦切除术 预测模型 

分 类 号：R736.4[医药卫生—肿瘤]