基于术前CT及临床特征构建卵巢上皮性癌肿瘤间质比预测模型的研究  

Research on prediction model of epithelial ovarian cancer stroma ratio based on preoperative CT and clinical characteristics

作　　者：孙艳[1] 罗祥桐 范雨涵 向往 SUN Yan;LUO Xiang-tong;FAN Yu-han;XIANG Wang(Hunan Cancer Hospital,Changsha 410013,China,;School of Medicine,Nankai University,Tianjin 300350,)

机构地区：[1]湖南省肿瘤医院,湖南长沙410013 [2]南开大学医学院,天津300350

出　　处：《中国临床医学影像杂志》2024年第4期268-273,共6页Journal of China Clinic Medical Imaging

基　　金：湖南省科技创新计划项目(编号:2021SK51119);湖南省卫生健康委科研计划项目(编号:202209015365);湖南省肿瘤医院2021年启航青年基金项目(编号:QH2021005)。

摘　　要：目的:探讨不同肿瘤间质比(Tumor-stroma ratio,TSR)的卵巢上皮性癌(Epithelial ovarian carcer,EOC)的特征性影像学表现,并基于术前临床及影像特征建立列线图预测模型。方法:收集我院2013年7月-2016年6月经手术病理证实为EOC的患者96例,根据术后组织病理结果将其分为高间质比组(TSR≥50%)和低间质比组(TSR<50%)两组。回顾性分析两组病例术前全腹部增强CT的形态学特征,包括原发灶的单侧或双侧、囊变程度、囊变形态、边界特点、长短径比、体积、平扫CT值、增强幅度、是否存在钙化、腹水、腹膜后淋巴结肿大以及腹膜转移情况。采用单因素及多因素Logistic回归分析筛选出评估EOC TSR的独立影响因子,并以此构建列线图预测模型。通过受试者工作特征(ROC)及曲线下面积(AUC)评估列线图模型的诊断效能,通过决策曲线评估列线图模型的临床适用性。结果:新辅助化疗、肿瘤长短径比、增强幅度、边界特点、囊变形态、是否存在腹膜后淋巴结肿大均是EOC TSR的独立影响因子。据此结果构建列线图,其ROC曲线的AUC值为0.966(95%CI 0.936~0.997),灵敏度为96.1%,特异度为88.9%。结论:高间质比组的EOC在CT上多表现出更高的长短径比、更明显的强化、边界不清、微囊型的囊变形态及腹膜后淋巴结肿大。基于EOC患者的术前CT和临床特征构建的列线图可以无创评估TSR,有助于辅助临床风险分层和治疗决策。Objective:To investigate the distinctive imaging features of epithelial ovarian cancer(EOC)with varying tumor-stroma ratio(TSR)and to develop a nomogram prediction model based on preoperative clinical and imaging characteristics.Methods:Retrospectively collected data from 96 patients diagnosed with EOC through surgery and pathology confirmation,spanning from July 2013 to June 2016 at our hospital.Patients were categorized into high TSR group(TSR≥50%)and low TSR group(TSR<50%)based on postoperative histopathology.The morphological features of preoperative contrast-enhanced CT of the two groups were retrospectively analyzed,including the location of the primary lesion,degree and morphology of cystic changes,border characteristics,longest and shortest diameters,length-to-shortness ratio,volume,CT values on plain scans,enhancement degree,presence of calcification,ascites,retroperitoneal lymph node enlargement,and peritoneal metastasis.Uni-variate and multivariate Logistic regression analyses were performed to identify independent factors influencing the evaluation of the TSR of EOC.Subsequently,a nomogram prediction model was developed based on these factors.The diagnostic perfor-mance of the nomogram model was evaluated by ROC curve and area under the curve(AUC),and the clinical applicability of the nomogram model was evaluated by decision curve analysis.Results:Preoperative neoadjuvant chemotherapy,length-to-shortness ratio of the tumor,enhancement degree,boundary,morphology of cystic changes,and enlarged lymph nodes were in-dependent influencing factors for TSR of EOC.Based on these results,a nomogram was constructed with an AUC value of 0.966(95%CI 0.936~0.997),sensitivity of 96.1%,and specificity of 88.9%.Conclusion:The high TSR group of E0C has a higher length-to-shortness ratio,higher enhancement degree,obscure boundary,microcystic form and enlarged retroperitoneal lymph nodes.The nomogram,built on the preoperative CT features of EOC patients,offers a non-invasive method for evaluat-ing the TSR and aiding in

关 键 词：卵巢肿瘤 肿瘤 腺和上皮 体层摄影术 X线计算机 

分 类 号：R737.31[医药卫生—肿瘤] R814.42[医药卫生—临床医学]