急性胰腺炎（acute pancreatitis，AP）是易导致器官功能障碍和病死率较高的急腹症，及时预判疾病的发生、发展趋势，是及早开展针对性治疗和干预的前提。影像组学能高通量提取医学影像中的定量特征，实现数据深度挖掘，现已经发现其可用于AP的诊断及预测疾病严重程度、疾病进展和复发。CT影像组学可以从功能性腹痛、慢性胰腺炎和复发性AP患者中，鉴别出复发性AP患者，受试者操作特征曲线下面积（aera under curve，AUC）达到0.88，还可以准确预测48个月内的AP复发，AUC达到0.93。MRI影像组学可以预测AP患者未来是否会进展为中重症AP，其AUC达到0.85，效果优于现有临床评分系统；还可以有效预测胰周坏死的发生，其AUC达到0.92。代谢组学已证实，AP发生、发展过程中代谢谱存在动态变化。报道显示，活性代谢物作为预警指标可用于AP的诊断、病因鉴别和严重程度评估。尿液代谢组学被证明能准确诊断AP，其AUC达到0.91。血液代谢组学模型可鉴别胆源性AP、高脂性AP和酒精性AP患者，AUC分别达到了0.89、0.91和0.86，还可以准确预测AP患者未来是否会进展为中重症AP，AUC可达0.99。影像组学和代谢组学2种技术相结合可优势互补，从不同层次共同表征疾病发生、发展的过程，且两者间存在内在联系，能更有效地实现早期预警并及早干预。

Acute pancreatitis (AP) is an acute abdominal disease that is prone to organ dysfunction, with high mortality. Timely prediction of the occurrence and development trend of the disease is the prerequisite for early treatment and intervention. Radiomics can extract quantitative features from medical images with high throughput and realize deep data mining. It can be used for the diagnosis of acute pancreatitis and prediction of severity, progression and recurrence of the disease. For the diagnosis of AP, CT radiomics can distinguish recurrent AP patients from functional abdominal pain, chronic pancreatitis, and recurrent AP patients, with an area under curve (AUC) of 0.88. For predicting AP recurrence, CT radiomics can accurately predict AP recurrence within 48 months, with an AUC of 0.93. For predicting the severity of AP, MRI radiomics can predict whether AP patients will progress to moderate to severe AP in the future, with an AUC of 0.85, which is better than clinical scoring systems. For predicting complications and progression of AP, MRI radiomics can effectively predict the occurrence of peripancreatic necrosis, with an AUC of 0.92. Metabolomics has confirmed that metabolic spectrum changes dynamically during the occurrence and development of AP. It has been reported that active metabolites can be used as early warning indicators for the diagnosis, etiology identification and severity assessment of AP. In addition, urinary metabolomics allows accurate diagnosis of AP, with an AUC of 0.91. For identifying the etiology of AP, the blood metabolomics models can identify patients with biliary AP, hyperlipidemic AP, and alcoholic AP, with AUCs of 0.89, 0.91, and 0.86, respectively. For predicting the severity of AP, the blood metabolomics models can accurately predict whether AP patients will progress to moderate to severe AP in the future, with an AUC of 0.99. The combination of the radiomics and metabolomics can complement each other's advantages and integrate multi-group data, which can jointly characterize the process and internal connections of disease occurrence and development from different levels, for achieving early warning and early intervention more effectively.