| 李妍婷,黄小华,万夕瑶,王渊,罗靖贤,朱之珺,周梦妮.双序列MRI影像组学联合临床预测急性胰腺炎合并代谢综合征患者发生急性胰腺炎后糖尿病[J].中国医学影像技术,2026,42(3):356~360 |
| 双序列MRI影像组学联合临床预测急性胰腺炎合并代谢综合征患者发生急性胰腺炎后糖尿病 |
| Dual-sequence MRI radiomics combined with clinic for predicting post-acute pancreatitis diabetes mellitus in acute pancreatitis patients complicated with metabolic syndrome |
| 投稿时间：2025-06-09 修订日期：2025-12-02 |
| DOI：10.13929/j.issn.1003-3289.2026.03.008 |
| 中文关键词: 胰腺炎 代谢综合征 糖尿病 磁共振成像 影像组学 |
| 英文关键词:pancreatitis metabolic syndrome diabetes mellitus magnetic resonance imaging radiomics |
| 基金项目:南充市科学技术局市校科技战略合作专项(20SXQT0303)、川北医学院临床医学院·附属医院2024年度科研发展计划(2024LC001)。 |
|
| 摘要点击次数: 88 |
| 全文下载次数: 43 |
| 中文摘要: |
| 目的 观察双序列MRI影像组学联合临床预测急性胰腺炎(AP)合并代谢综合征(MS)患者发生急性胰腺炎后糖尿病(PPDM-A)的价值。方法 回顾性纳入121例AP合并MS患者,根据是否发展为PPDM-A分为PPDM-A组(n=50)与非PPDM-A组(n=71),并按7∶3比例随机划分训练集(n=84)与测试集(n=37)。比较组间临床资料,基于差异有统计学意义的参数构建临床模型;提取T2WI及动脉期晚期对比增强T1WI(CE-T1WI)胰腺影像组学特征、构建影像组学模型;以二者构建联合模型。评估各模型预测PPDM-A的效能。结果 组间血糖及高脂血症、脂肪肝、高血压和AP复发占比差异均有统计学意义(P均＜0.05)。基于T2WI及CE-T1WI各筛选出2及6个最优影像组学特征。临床模型、影像组学模型及联合模型预测训练集PPDM-A的曲线下面积(AUC)分别为0.792、0.859及0.927,在测试集分别为0.793、[LM]0.842及0.927,均以联合模型AUC最高(P均＜0.05)。结论 双序列MRI影像组学联合临床可用于预测AP合并MS患者发生PPDM-A。 |
| 英文摘要: |
| Objective To observe the value of dual-sequence MRI radiomics combined with clinic for predicting post-acute pancreatitis diabetes mellitus (PPDM-A) in acute pancreatitis (AP) patients complicated with metabolic syndrome (MS). Methods Totally 121 AP patients complicated with MS were retrospectively enrolled and divided into PPDM-A group (n=50) and non PPDM-A group (n=71) according to developed into PPDM-A or not, also randomly divided into training set (n=84) and test set (n=37) at a ratio of 7∶3. Clinical data were compared between groups, and a clinical model was constructed based on parameters being significantly different between groups. The radiomics features of pancreas in T2WI and late arterial contrast-enhanced-T1WI (CE-T1WI) were extracted, and a radiomics model was constructed. The combined model was constructed based on the above two. The efficacy of each model for predicting PPDM-A was analyzed. Results Significant differences of blood glucose, the proportion of hyperlipidemia, fatty liver, hypertension and AP recurrence were found between groups (all P＜0.05). Finally, 2 and 6 optimal radiomics features were selected from T2WI and CE-T1WI, respectively. The area under the curve (AUC) of clinical model, radiomics model and combined model for predicting PPDM-A in training set was 0.792, 0.859 and 0.927, which was 0.793, 0.842 and 0.927 in test set, respectively, and AUC of combined model was the highest in both sets (both P＜0.05). Conclusion Dual-sequence MRI radiomics combined with clinic could be used to predict PPDM-A in AP patients complicated with MS. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
