Computed tomography findings and surgical outcomes in acute mesenteric ischemia: a retrospective single-center cohort study

Background: Acute mesenteric ischemia (AMI) is a time-dependent condition associated with high mortality, frequently related to diagnostic delay and progression to irreversible bowel necrosis. Computed tomography (CT) plays a pivotal role not only in diagnosis but also in guiding therapeutic decision-making. However, the prognostic role of individual CT findings and their relationship with irreversible ischemic injury remain incompletely defined. This study evaluated predefined CT findings, AMI subtype, treatment pathways, intraoperative infarction or necrosis, and mortality in a real-world cohort of patients with CT-confirmed AMI. Methods: A retrospective single-center cohort study was conducted including 102 adult patients with CT-confirmed AMI admitted between January 2018 and February 2026. Patients were classified by AMI subtype as arterial occlusive AMI, venous AMI, non-occlusive mesenteric ischemia (NOMI), or secondary/mechanical ischemia. Eight predefined CT findings were analyzed: pneumatosis intestinalis, porto-mesenteric venous gas, bowel wall thickening, colitis/ileitis, occlusion/volvulus/intussusception, free intraperitoneal air, free peritoneal fluid, and vascular suffering or bowel wall hypoenhancement. Radiological diagnostic certainty was analyzed separately as a synthetic imaging variable. The primary endpoint was intraoperative evidence of bowel infarction, necrosis, or irreversible ischemia among surgically explored patients. Secondary endpoints included clinically significant ischemia, bowel resection, invasive treatment, in-hospital mortality, and 30-day mortality. Univariate associations, diagnostic performance measures, ROC/AUC analysis, and parsimonious multivariable logistic regression models for mortality were performed. Results: Arterial occlusive AMI was the most frequent AMI subtype (59/102, 57.8%), followed by NOMI (23/102, 22.5%), secondary/mechanical ischemia (15/102, 14.7%), and venous AMI (5/102, 4.9%). In-hospital mortality was 41.2% (42/102), while 30-day mortality was 46.5% (47/101), including five deaths after discharge. The primary endpoint was assessable in 65 patients; intraoperative infarction or necrosis was present in 53 (81.5%). No individual CT finding showed strong standalone discriminatory performance for intraoperative infarction or necrosis. Pneumatosis showed the strongest descriptive performance for the primary endpoint (OR: 2.72, 95% CI: 0.76-9.80; p = 0.188; AUC 0.62). Pneumatosis showed an exploratory association with 30-day mortality (OR: 2.42, 95% CI: 1.08-5.43; p = 0.045). In the multivariable model for 30-day mortality, acute kidney injury was independently associated with death (adjusted OR: 9.26, 95% CI: 2.40-35.74; p = 0.001), while non-surgical admission showed a borderline association after adjustment (adjusted OR: 2.70, 95% CI: 0.94-7.71; p = 0.065). The model AUC was 0.82. Conclusions: In this heterogeneous AMI cohort, individual CT findings were clinically meaningful but showed limited standalone discriminatory performance for surgically confirmed infarction or necrosis. Pneumatosis showed an exploratory association with 30-day mortality, while porto-mesenteric venous gas and free air were specific but infrequent. CT remains essential for diagnosis, subtype classification, and treatment planning, but CT findings should be interpreted within an integrated clinical and multidisciplinary framework. The exploratory CT ischemic burden score should be considered descriptive and not a validated triage or prognostic tool.