Differentiation of cancerous and inflammatory colorectal perforations using multi-detector computed tomography.
Abdom Radiol (NY). 2017 Apr 11. doi: 10.1007/s00261-017-1134-3. [Epub ahead of print] Gong XH1, Zhuang ZG1, Zhu J1, Feng Q1, Xu JR1, Qian LJ2.
PURPOSE: To determine reliable CT features to distinguish cancerous from inflammatory colorectal perforations.
MATERIALS AND METHODS: A total of 43 patients with surgically and pathologically confirmed colorectal perforation caused by either colorectal cancer (n =27) or an inflammatory conditions (n = 16) were identified. Two radiologists independently assessed the contrast-enhanced CT features for locations of perforation, mural configurations, soft-tissue alterations, lymphadenopathy, and metastases. Intergroup comparisons for univariate analysis were performed using Fisher's exact test or chi-square test for categorical data and Mann-Whitney test for numeric data. Stepwise logistic regression analysis was conducted with features that were found significant under the univariate analysis. Interobserver agreement was assessed using intraclass correlation coefficient (ICC) and kappa test.
RESULTS: Maximal mural thickness >1.39 cm (sensitivity, 100%; specificity, 68.75%), luminal mass or shoulder formation (sensitivity, 88.89%; specificity, 68.75%), absence of diverticula (sensitivity, 96.30%; specificity, 50.00%), irregular mural thickening (sensitivity, 92.59%; specificity, 81.25%), lymphadenopathy (sensitivity, 40.74%; specificity, 93.75%), and metastases (sensitivity, 25.93%; specificity, 100%) were significantly frequent in cancerous perforations. The maximal mural thickness (P = 0.0493, odds ratio = 439.83) and irregular mural thickening (P = 0.0343, odds ratio = 4.69) were identified as the highly distinguished identifiers.
CONCLUSIONS: The CT manifestations of cancerous and inflammatory colorectal perforations overlap. Definitive diagnosis is not always possible with imaging alone. The maximal mural thickness >1.39 cm and irregular configuration of the thickened bowel wall were the two highly statistically significant CT features that may help order the difference between the two entities.