Acute diverticulitis: beyond the diagnosis: predictive role of CT in assessing risk of recurrence and clinical implications in non-operative management of acute diverticulitis.

Abstract

The aim of the study is to identify CT findings that are predictive of recurrence of acute uncomplicated colonic diverticulitis, to better risk-stratify these patients for whom guidelines recommend a conservative outpatient treatment and to determine the appropriate management with an improvement of health costs. Over the past year, 33 patients enrolled in an outpatient integrated care pathway (PDTA) for uncomplicated acute diverticulitis with 1-year follow-up period, without recurrence, and 33 patients referred to Emergency Department for a recurrent acute diverticulitis were included. Images of admission CT were reviewed by two radiologists and the imaging features were analyzed and compared with Chi-square and Student t tests. Univariate and multivariate Cox regression models were employed to identify parameters that significantly predicted recurrence in 1-year follow-up period and establish cutoff and recurrence-free rates. The maximally selected rank statistics (MSRS) were used to identify the optimal wall thickening cutoff for the prediction of recurrence. Patients with recurrence showed a greater mean parietal thickness compared to the group without recurrence (16mm vs. 11.5mm; HR 1.25, p < 0.001) and more evidence of grade 4 of peridiverticular inflammation (40% vs. 12%, p = 0.009, HR 3.44). 12-month recurrence-free rates progressively decrease with increasing thickness and inflammation. In multivariate analysis, only parietal thickness maintained its predictive power with an optimal cutpoint > 15mm that causes a sixfold increased risk of recurrence (HR 6.22; 95% CI, 3.05-12.67; p < 0.001). Beyond thickness and peridiverticular inflammation, predictive value of early recurrence within 90days from the 1st episode resulted also an Hinchey Ib on admission CT. The maximum wall thickening and the grade of peridiverticular inflammation can be considered as predictive factors of recurrence and may be helpful in selecting patients for a tailored treatment to prevent the risk of recurrence.