| 何瑛,吕静,胡茜,邵九杰,朱彦芳,朱永琪,王艺霖,王霈,刘云.增强皮质期CT影像组学评分联合CT特征列线图预测肾细胞癌同时性远处转移[J].中国医学影像技术,2024,40(12):1894~1899 |
| 增强皮质期CT影像组学评分联合CT特征列线图预测肾细胞癌同时性远处转移 |
| Nomogram based on enhanced cortical phase CT Radscores combined with CT features for predicting synchronous distant metastasis of renal cell carcinoma |
| 投稿时间：2024-06-06 修订日期：2024-09-05 |
| DOI：10.13929/j.issn.1003-3289.2024.12.019 |
| 中文关键词: 肾肿瘤 肿瘤转移 体层摄影术,X线计算机 影像组学 列线图 |
| 英文关键词:kidney neoplasms neoplasm metastasis tomography, X-ray computed radiomics nomograms |
| 基金项目: |
|
| 摘要点击次数: 363 |
| 全文下载次数: 91 |
| 中文摘要: |
| 目的 观察增强皮质期CT影像组学评分(Radscore)联合CT特征列线图预测肾细胞癌(RCC)同时性远处转移(SDM)的价值。方法 回顾性分析A中心139例RCC,将其分为SDM组(n=46)与无SDM组(n=93),并按照7∶3比例分为训练集(n=97)与测试集(n=42),以B中心20例RCC为验证集(SDM 8例、无SDM 12例)。于增强皮质期CT图像中提取并筛选影像组学特征,计算Radscore;以多因素logistic回归分析筛选临床及CT特征中对于RCC SDM的独立影响因素,联合Radscore构建逻辑回归模型,并以列线图可视化;以受试者工作特征曲线及曲线下面积(AUC)评估列线图预测RCC SDM的效能。结果 肿瘤最大径、CT-T分期及肾周脂肪条索影均为RCC SDM的独立影响因素(P均＜0.01)。基于5个最优特征计算Radscore;基于肾周脂肪条索影、CT-T分期及Radscore构建的模型预测训练集、测试集及验证集RCC SDM的AUC分别为0.964、0.921及0.885。结论 利用增强皮质期CT Radscore联合肾周脂肪条索影及CT-T分期可有效预测RCC SDM。 |
| 英文摘要: |
| Objective To observe the value of nomogram based on enhanced cortical phase CT Radscore combined with CT features for predicting synchronous distant metastasis (SDM) of renal cell carcinoma (RCC). Methods Totally 139 RCC patients from center A were retrospectively enrolled and divided into SDM group (n=46) and non-SDM group (n=93), also classified as training set (n=97) and test set (n=42) at a ratio of 7∶3. Additionally, 20 RCC patients from center B were included as validation set (8 cases with SDM and 12 cases without SDM). Radiomics features were extracted and screened based on enhanced cortical phase CT images to calculate Radscore. Multivariate logistic regression analysis was performed to identify independent predictors of RCC SDM among clinical and CT features. Then a logistic regression model was constructed combining Radscore and independent predictors of RCC SDM and visualized as a nomogram. The receiver operating characteristic curve and the area under the curve (AUC) was used to assess the efficacy of the nomogram for predicting RCC SDM. Results The maximum tumor diameter, CT-T stage and perirenal adipose stranding were all independent predictors of RCC SDM (all P＜0.01). Radscore was calculated based on 5 optimal features. The nomogram was constructed based on perirenal adipose stranding, CT-T stage and Radscore. AUC of the model for predicting RCC SDM in training set, test set and validation set was 0.964, 0.921 and 0.885, respectively. Conclusion Enhanced cortical phase CT Radscore combined with perirenal adipose stranding and CT-T stage could effectively predict RCC SDM. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
