Distinguishing benign and life-threatening pneumatosis intestinalis in patients with cancer by CT imaging features.
Academic Article
Overview
abstract
OBJECTIVE: The purpose of this study is to determine the overall proportion of clinically worrisome and benign pneumatosis intestinalis (PI) occurring in patients with cancer and to evaluate associated risk factors and CT features. MATERIALS AND METHODS: We retrospectively studied the CT examinations of 84 patients treated at our tertiary cancer center. Reviewers who were blinded to clinical data and classification analyzed PI in terms of location, pattern (linear, cystic, or both), and associated CT features, including pneumoperitoneum, portomesenteric venous air, bowel wall thickening, bowel dilatation, and ascites. On the basis of the review of clinical information and criteria derived from prior literature, the cases were classified as clinically worrisome PI (underlying bowel disease) or benign PI (diagnosis of exclusion that resolved on follow-up imaging without targeted therapy). Clinical factors reviewed included age, sex, cancer type, steroid use, and chemotherapy administration. RESULTS: Forty-seven patients were classified as having benign PI (56%) and the remainder as having clinically worrisome PI (44%). The following imaging features correlated significantly with clinically worrisome PI: bowel wall thickening (p < 0.001), mesenteric stranding (p < 0.001), ascites (p < 0.001), bowel dilatation (p = 0.004), location confined to small bowel (p = 0.012), and portomesenteric venous gas (p = 0.02). Benign PI was significantly associated with PI confined to the colon (p = 0.004). CONCLUSION: Benign PI was slightly more prevalent than clinically worrisome PI in our cohort of patients with cancer. The presence of certain CT features (mesenteric stranding, bowel wall thickening, and ascites) and the location of PI may be indicators of more significant bowel disease and, therefore, of clinically worrisome cases. There was no statistical significance achieved for nonimaging clinical factors.
