Thompson Atalie C, Olcott Eric W, Poullos Peter D, Jeffrey R Brooke, Thompson Matthew O, Rosenberg Jarrett, Shin Lewis K
Stanford University School of Medicine, 291 Campus Drive, Li Ka Shing Building, Stanford, CA, 94305, USA,
Emerg Radiol. 2015 Aug;22(4):385-94. doi: 10.1007/s10140-015-1297-6. Epub 2015 Feb 17.
Confident diagnosis of appendicitis when the appendix is borderline (6 to 7 mm) in size can be challenging. This retrospective study assessed computed tomography (CT) findings that are most predictive of appendicitis when the appendix is borderline in diameter. Three radiologists conducted separate, blind retrospective reviews of 105 contrast-enhanced CTs with borderline appendices. Presence or absence of appendicitis was confirmed by chart review of clinical or surgical outcomes. Logistic regression was used to determine the odds ratio (OR) and the receiver operating characteristic for CT features predictive of appendicitis. Absence of intraluminal air (OR = 5.11, p < 0.001), wall hyperemia (OR = 3.92, p = 0.002), wall thickening (OR = 29.7, p < 0.001), and fat stranding (OR = 3.85, p = 0.003) were significant findings in univariate logistic regression. Using a multivariate model, we found that the absence of intraluminal air (OR = 6.04, p = 0.002) and wall thickening (OR = 24.6, p < 0.001) remained statistically significant and were unaffected by adjustment for gender and pediatric age. The area under the curve was significantly greater for the multivariate model than the initial, clinical CT impressions (p = 0.024). The combination of wall thickening and absence of intraluminal air was 92.6 % (95 % CI 75.7-99.1) sensitive and 82.4 % (95 % CI 65.5-93.2) specific for appendicitis. Wall thickening and the absence of intraluminal air are prominent predictors of appendicitis and, if present together, these features may aid in identifying appendicitis on CT when the appendix is borderline in size.
当阑尾大小处于临界值（6至7毫米）时，对阑尾炎进行确切诊断可能具有挑战性。这项回顾性研究评估了在阑尾直径处于临界值时，对阑尾炎最具预测性的计算机断层扫描（CT）表现。三位放射科医生分别对105例阑尾处于临界值的增强CT进行了独立的、盲法回顾性分析。通过查阅临床或手术结果的图表来确认是否存在阑尾炎。采用逻辑回归来确定预测阑尾炎的CT特征的优势比（OR）和受试者工作特征曲线。在单变量逻辑回归中，腔内无气体（OR = 5.11，p < 0.001）、壁充血（OR = 3.92，p = 0.002）、壁增厚（OR = 29.7，p < 0.001）和脂肪条索（OR = 3.85，p = 0.003）是显著发现。使用多变量模型，我们发现腔内无气体（OR = 6.04，p = 0.002）和壁增厚（OR = 24.6，p < 0.001）在统计学上仍然显著，并且不受性别和儿童年龄调整的影响。多变量模型的曲线下面积显著大于最初的临床CT印象（p = 0.024）。壁增厚和腔内无气体的组合对阑尾炎的敏感性为92.6%（95%可信区间75.7 - 99.1），特异性为82.4%（95%可信区间65.5 - 93.2）。壁增厚和腔内无气体是阑尾炎的突出预测指标，如果同时出现，这些特征可能有助于在阑尾大小处于临界值时通过CT识别阑尾炎。
