基于术前增强CT特征的列线图预测腹膜后肉瘤组织病理分级价值  

Value of a nomogram model based on preoperative contrast-enhanced CT features in predicting the histopathological grade of retroperitoneal sarcomas

作　　者：孙湛[1] 胡尊英 梁皓昱 赵涵[3] 郝大鹏[2] 王鹤翔[2] SUN Zhan;HU Zunying;LIANG Haoyu;ZHAO Han;HAO Dapeng;WANG Hexiang(Department of Health Materials Management,The Affiliated Hospital of Qingdao University,Qingdao 266003,China)

机构地区：[1]青岛大学附属医院放疗技术中心,山东青岛266003 [2]青岛大学附属医院放射科,山东青岛266003 [3]青岛大学附属医院病理科,山东青岛266003

出　　处：《青岛大学学报(医学版)》2024年第6期850-854,共5页Journal of Qingdao University(Medical Sciences)

基　　金：山东省自然科学基金项目(ZR2021MH159);青岛大学附属医院临床医学+X项目(QDFY+X2021015);山东省医药卫生科技发展计划项目(2019WS373)。

摘　　要：目的探讨基于术前增强CT特征的列线图在术前预测腹膜后肉瘤(RPS)病理分级的价值。方法回顾性收集156例术后病理证实为RPS病人的术前增强CT资料。根据法国癌症中心联合会组织学分级系统,高级别RPS病人(Ⅱ~Ⅲ级)113例,低级别RPS病人(Ⅰ级)43例。基于单因素及多因素Logistic回归分析筛选独立预测因子,建立列线图预测模型。通过10折交叉验证评估模型性能,采用中位受试者工作特征曲线下面积(AUC)及中位准确度对模型预测效能进行评估。列线图模型的校准度和临床适用度应用校准曲线、决策曲线评估。结果根据单因素Logistic回归结果,筛选出病灶数量、形状、强化程度、脂肪成分、液体成分、淋巴结转移纳入多因素Logistic回归分析,最终增强CT影像学特征的病灶数量、强化程度、液体成分被确定为RPS病理分级预测因子。预测模型的中位AUC为0.858(95%CI=0.776~0.944),中位准确度为84.0%。结论基于增强CT影像学特征的病灶数量、强化程度和液体成分的列线图模型可有效预测RPS病理分级,为病人制定临床治疗方案提供依据。Objective To investigate the value of a nomogram model based on preoperative contrast-enhanced CT(CECT)features in predicting the histopathological grade of retroperitoneal sarcomas(RPS).Methods A retrospective analysis was performed for the preoperative CECT data of 156 patients with pathologically confirmed RPS.According to the French Federation of Cancer Centers histological grading system,there were 113 patients with high-grade(grade Ⅱ-Ⅲ)RPS and 43 patients with low-grade(grade Ⅰ)RPS.Univariate and multivariate Logistic regression analyses were used to identify independent predictive factors,and a nomogram model was established.The method of 10-fold cross validation was used to assess the performance of the model,and the median area under the receiver operating characteristic curve(AUC)and median accuracy were used to evaluate the predictive performance of the nomogram model.The calibration curve and the decision curve were used to assess the degree of calibration and clinical applicability of this model.Results According to the results of the univariate Logistic regression analysis,lesion number,shape,degree of enhancement,fat composition,fluid composition,and lymph node metastasis were included in the multivariate Logistic regression analysis,and finally the number of lesions,the degree of enhancement,and fluid composition were identified as predictive factors for the pathological grade of RPS.The predictive nomogram model showed a median AUC value of 0.858(95%CI=0.776-0.944)and a median accuracy of 84.0%.Conclusion The nomogram model based on lesion number,degree of enhancement,and fluid composition can effectively predict the pathological grade of RPS,which provides a basis for developing clinical treatment regimens for patients.

关 键 词：体层摄影术 X线计算机 肉瘤 腹膜后间隙 列线图 病理学 临床 

分 类 号：R445.3[医药卫生—影像医学与核医学] R739.9[医药卫生—诊断学]