Assessment and predictive value of acute pancreatitis severity based on CT radiomics and psoas muscle index
DOU Rui-xin,GONG Yan,ZHAO Guang
Abstract:
Objective To investigate the value of CT radiomics combined with the psoas muscle index (PMI) in assessing and predicting the severity of acute pancreatitis (AP). Methods A retrospective analysis was conducted on 160 AP patients admitted to our hospital from January 2023 to December 2023. Patients were divided into a severe group (41 cases) and a non-severe group (119 cases). Pancreatic radiomics features and PMI were extracted from CT images by two radiologists. Correlation analysis of PMI and patient nutritional risk with AP severity. Patients were randomly divided into a training set (n=112, with 24 cases in the severe group and 88 cases in the non-severe group) and a testing set (n=48, with 17 cases in the severe group and 31 cases in the non-severe group). Logistic regression (LR) was used to construct three models: a clinical model, a CT radiomics model, and a PMI-CT radiomics model incorporating PMI features. The predictive performance of the three models was evaluated using the receiver operating characteristic (ROC) curve. Results PMI was significantly negatively correlated with nutritional risk (r=-0.72, P <0.05) and significantly positively correlated with AP severity (r=0.93, P <0.05). The areas under the ROC curve (AUC) for the clinical model, CT radiomics model, and PMI-CT radiomics model in the training set were 0.78, 0.89, and 0.91, and in the testing set were 0.77, 0.88, and 0.90. Conclusion PMI is strongly associated with the severity and prognosis of acute pancreatitis. The PMI-CT radiomics model, incorporating PMI features, can serve as a reliable tool for more accurately assessing the severity and prognostic risk in AP patients.
