医学
脂肪组织
曲线下面积
急性胰腺炎
接收机工作特性
试验预测值
内科学
计算机断层摄影术
核医学
放射科
作者
Shengqi Wang,Meiping Wang,Li Jiang,Xin Zhao
标识
DOI:10.1016/j.ejrad.2023.111215
摘要
PurposeTo evaluate the association between body composition parameters derived from computed tomography (CT) scans and clinical outcomes in patients with severe acute pancreatitis (AP).MethodsPatients who have been diagnosed AP with a CT scan at ICU admission were included. Body composition parameters were measured on a single slice at L2-3 of the unenhanced CT scans. The intermuscular adipose tissue (IMAT) , visceral adipose tissue (VAT), skeletal muscle area (SMA) and skeletal muscle density (SMD) were assessed using HUs by image analysis software. Univariable and multivariable analyses were performed to analyze the association between body composition parameters and clinical outcomes including all-cause mortality or prolonged ICU stay. The area under the curve (AUC) of a receiver operating characteristic curve was used to explore the predictive value of the body composition on treatment clinical outcomes.ResultsA total of 158 patients were included. The IMAT (8.3cm2 vs 6.0cm2, P=0.001) and VAT (190.3cm2 vs 143.7cm2, P<0.001) were significantly higher in the severe AP group than in the moderately severe group, but were not associated with outcomes. For 1 HU of SMD increased, the risk of poor clinical outcomes decresed 11% (adjusted OR 0.892, 95%CI 0.806-0.987, P=0.026), while an SMD below the median value (32.1 HU for males and 28.5 HU for females) was independently associated with worse outcomes in the multivariable analysis (adjusted OR 8.868, 95% CI 2.146-36.650, P=0.003). The SMD had a good predictive ability for clinical outcomes, AUC was 0.824 (95% CI, 0.715-0.933) for males and 0.803 (95% CI, 0.639-0.967) for females.ConclusionLow SMD was associated with poor outcomes in patients with severe and moderately severe AP and might be used as a novel marker to predict outcomes in patients suffering from severe and moderately severe AP.
科研通智能强力驱动
Strongly Powered by AbleSci AI