摘要
目的联合临床数据、常规超声、超声造影及超声影像组学特征,构建肾透明细胞癌患者术后无进展生存期预后模型。资料与方法回顾性分析2015年1月—2021年12月解放军总医院第一医学中心536例肾透明细胞癌患者,根据患者术后是否发生进展分为术后进展组83例及术后未进展组453例,按照8∶2分为训练集和验证集,收集患者临床、常规超声及超声造影数据,提取超声影像组学特征,并计算组学评分。建立临床-多模态超声组学预测模型,通过受试者工作特征曲线、校准曲线、决策曲线评估模型性能。结果中位肿瘤直径为3.8(2.6,5.5)cm,83例(15.49%)术后出现进展。基于训练集数据构建临床-多模态超声组学模型,纳入年龄、肿瘤直径、肿瘤是否凸向集合系统、肿瘤回声特征、肿瘤边界及超声影像组学评分6个预测因子,其中年龄(HR=1.05,95%CI 1.02~1.07)、肿瘤凸向集合系统(HR=2.26,95%CI 1.25~4.08)、肿瘤回声不均质(HR=2.75,95%CI 1.42~5.33)、肿瘤边界不清(HR=4.94,95%CI 2.19~11.16)及超声影像组学评分(HR=1.86,95%CI 1.46~2.38)为术后进展的独立危险因素。模型预测1年、2年及3年无进展生存期的训练集曲线下面积分别为0.89、0.92、0.92,验证集曲线下面积分别为0.88、0.87、0.92。结论基于临床数据及多模态超声的预测模型能够在术前对肾透明细胞癌患者术后发生进展风险进行较为精准的评估,为术前制订个性化治疗策略提供重要依据。
Purpose To develop a prognostic model for postoperative progression-free survival in patients with clear cell renal cell carcinoma by integrating clinical data,conventional ultrasound,contrast-enhanced ultrasound,and ultrasound radiomic features.Materials and Methods A retrospective analysis was performed on 536 patients with clear cell renal cell carcinoma who underwent surgery at the First Medical Center of Chinese PLA General Hospital from January 2015 to December 2021.Patients were classified into a postoperative progression group(n=83)and a non-progression group(n=453)according to whether disease progression occurred after surgery.The cohort was divided into a training set and a validation set at a ratio of 8:2.Clinical,conventional ultrasound,and contrast-enhanced ultrasound data were collected.Ultrasound radiomic features were extracted and a radiomics score was calculated.A clinical-multimodal ultrasound radiomics prediction model was developed,and model performance was evaluated using receiver operating characteristic curves,calibration curves,and decision curve analysis.Results The median tumor diameter was 3.8(2.6,5.5)cm.Postoperative disease progression occurred in 83 patients(15.49%).Based on the training set data,a clinical-multimodal ultrasound radiomics model was constructed,incorporating six predictors:age,tumor diameter,protrusion into the collecting system,tumor echogenicity pattern,tumor boundary,and the ultrasound radiomics score.Among these,age(HR=1.05,95%CI 1.02-1.07),tumor protrusion into the collecting system(HR=2.26,95%CI 1.25-4.08),heterogeneous echogenicity pattern(HR=2.75,95%CI 1.42-5.33),unclear tumor boundary(HR=4.94,95%CI 2.19-11.16),and the ultrasound radiomics score(HR=1.86,95%CI 1.46-2.38)were identified as independent risk factors for postoperative progression.The model achieved AUCs of 0.89,0.92,and 0.92 for predicting 1-,2-,and 3-year progression-free survival in the training set,and 0.88,0.87,and 0.92 in the validation set,respectively.Conclusion A prediction model integrating clinical data and multimodal ultrasound enables relatively accurate preoperative assessment of postoperative progression risk in patients with clear cell renal cell carcinoma and may inform individualized preoperative treatment planning.
作者
宋禄达
梁蕾
陈子为
张奥
赵萍
姜波
蒋文莉
李秋洋
SONG Luda;LIANG Lei;CHEN Ziwei;ZHANG Ao;ZHAO Ping;JIANG Bo;JIANG Wenli;LI Qiuyang(Department of Ultrasound,the First Medical Centre of Chinese PLA General Hospital,Beijing 100853,China)
出处
《中国医学影像学杂志》
北大核心
2026年第2期230-236,242,共8页
Chinese Journal of Medical Imaging
关键词
癌
肾细胞
超声检查
超声造影
影像组学
无进展生存期
预后
Carcinoma,renal cell
Ultrasonography
Contrast-enhanced ultrasound
Radiomics
Progression-free survival
Prognosis
