| 李庆龙,詹鹏超,刘星,李莉明,邢静静,梁盼,高剑波.CT影像组学联合临床指标列线图预测免疫联合化学治疗用于胃癌效果[J].中国医学影像技术,2025,41(12):2026~2031 |
| CT影像组学联合临床指标列线图预测免疫联合化学治疗用于胃癌效果 |
| Nomogram of CT radiomics and clinical indicators for predicting efficacy of immunotherapy combined with chemotherapy of gastric cancer |
| 投稿时间：2025-03-19 修订日期：2025-09-17 |
| DOI：10.13929/j.issn.1003-3289.2025.12.019 |
| 中文关键词: 胃肿瘤 免疫疗法 药物治疗 体层摄影术,X线计算机 影像组学 |
| 英文关键词:stomach neoplasms immunotherapy drug therapy tomography, X-ray computed radiomics |
| 基金项目: |
|
| 摘要点击次数: 171 |
| 全文下载次数: 26 |
| 中文摘要: |
| 目的 观察CT影像组学联合临床指标列线图预测免疫联合化学治疗(化疗)用于胃癌效果的价值。方法 回顾性收集136例接受免疫联合化疗的胃癌患者,按7∶3比例随机将其分为训练集(n=96)与验证集(n=40),并根据疗效划分应答组与无应答组。以logistic回归分析于训练集临床资料中筛选独立预测因子并构建临床模型。于治疗前静脉期CT中提取及筛选组影像学特征并构建组学模型,联合前述独立预测因子构建联合模型并绘制列线图。评估各模型预测效能,以及联合模型的拟合度和临床净收益。结果 应答组77例、无应答组59例。临床模型、组学模型及联合模型在训练集的曲线下面积(AUC)分别为0.711、0.808及0.844,在验证集分别为0.670、0.714及0.732。联合模型在训练集的AUC高于临床模型(P=0.008),其他各模型在训练集和验证集的AUC差异均无统计学意义(P均＞0.05)。联合模型预测结果与实际结果的一致性良好、拟合度优,且临床净收益较高。结论 CT影像组学联合临床指标列线图能有效预测免疫联合化疗用于胃癌效果。 |
| 英文摘要: |
| Objective To observe the value of nomogram of CT radiomics and clinical indicators for predicting efficacy of immunotherapy combined with chemotherapy of gastric cancer. Methods Totally 136 gastric cancer patients who underwent immunotherapy combined with chemotherapy were retrospectively collected and randomly divided into training set (n=96) and validation set (n=40) at the ratio of 7∶3. The therapeutic efficacy was evaluated, and the patients were categorized into response group and non-response group. Based on clinical data in training set, logistic regression was used to screen independent predictors of efficacy of combined therapy of gastric cancer and to develop a clinical model. Radiomics features of lesions in pre-treatment venous phase CT were extracted and screened to construct a radiomics model, then a combined model was built by integrating independent predictive factors, and its nomogram was drawn. The predictive performance of each model was assessed, and the calibration and clinical net benefit of combined model were evaluated. Results There were 77 cases in response group and 59 cases in non-response group. The area under the curve (AUC) of clinical, radiomics and combined models in training set was 0.711, 0.808 and 0.844, respectively, while in validation set was 0.670, 0.714 and 0.732, respectively. Except for combined model which had higher AUC in training set than clinical model (P=0.008), no significant difference of AUC of other models was found in training set and validation set (all P＞0.05). The prediction results of combined model exhibited highly consistent to actual results, with excellent fitting performance and high clinical net benefit. Conclusion Nomogram of CT radiomics and clinical indicators could effectively predict efficacy of immunotherapy combined with chemotherapy of gastric cancer. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
