WU Yuxin, Sachurilatu, LIU Pengfei, WANG Bo, HAI Feng, ZHANG Dongming
Objective: To explore the application value of body composition phenotype based on computed tomography (CT) in the evaluation of clinical and imaging severity of acute pancreatitis (AP). Methods: The clinical and CT imaging data of 206 patients with AP were retrospectively analyzed. The L3 skeletal muscle mass index at the third lumbar vertebra level (L3 SMI), intermuscular adipose tissue (IMAT) and visceral-to-subcutaneous fat ratio (VSR) were measured, and four body composition phenotypes were constructed accordingly. The bedside index for severity in acute pancreatitis (BISAP) and modified CT severity index (MCTSI) were used to evaluate the severity of AP patients. Spearman correlation analysis, Logistic regression analysis and receiver operating characteristic (ROC) curve were used to explore the relationship between body composition parameters and their phenotypes and the severity of AP. Results: ①Clinical severity (BISAP stratification): Univariate analysis showed that VSR, IMAT and sarcopenia (SP) were significantly associated with AP severity (all P＜0.05), while L3 SMI was not statistically significant. Among them, phenotype D had the highest risk of severe AP (OR=10.5, P=0.001). Multivariate regression showed that age, neutrophil-to-lymphocyte ratio (NLR) and C-reactive protein (CRP) were independent risk factors (all P＜0.05), and body composition phenotype had no independent predictive value. ROC analysis showed that the area under the receiver operating characteristic curve (AUC) of the combined prediction model constructed by age, inflammation index and body composition parameters was 0.883 (95%CI: 0.816-0.951), and the prediction efficiency was significantly better than that of single index (VSR was better than CRP in single index).②Imaging severity (MCTSI stratification): Univariate analysis showed that only phenotype D showed a significant correlation (OR=3.333, P=0.046), and the other single items did not show statistical significance (all P＞0.05). Multivariate analysis showed that age (P=0.008) and CRP (P=0.027) were independent risk factors, and there was no significant difference in body composition phenotype and phenotype D (all P＞0.05). ROC analysis showed that the combined model of age and CRP had a certain predictive ability (AUC=0.679), and the age performance was the most stable (AUC=0.643, P=0.001). There was no significant difference between VSR and CRP (P＞0.05). Conclusion: Phenotype D (sarcopenia and visceral fat dominance) based on CT evaluation has important value in the progression of AP patients. Although multivariate analysis shows that its independent predictive value is limited, the body composition phenotype can be combined with independent risk factors such as age, CRP and NLR in clinical evaluation to construct a multi-dimensional evaluation system to achieve early and accurate management of AP patients.
