Differences between computed tomoghaphy and surgical findings in acute complicated diverticulitis.

Background/objective: A preoperative reliable classification system between clinical and computed tomography (CT) findings to better plan surgery in acute complicated diverticulitis (ACD) is lacking. We studied the inter-observer agreement of CT scan data and their concordance with the preoperative clinical findings and the adherence with the intraoperative status using a new classification of diverticular disease (CDD).

Methods: 152 patients operated on for acute complicated diverticulitis (ACD) were retrospectively enrolled. All patients were studied with CT scan within 24 h before surgery and CT images were blinded reanalyzed by 2 couples of radiologists (A/B). Kappa value evaluated the inter-observer agreement between radiologists and the concordance between CDD, preoperative clinical findings and findings at operation. Univariate and multivariate analysis were used to evaluate the predicting values of CT classification and CDD stage at surgery on postoperative outcomes.

Results: Overall inter-observer agreement for the CDD was high, with a kappa value of 0.905 (95% CI = 0.850-0.960) for observers A and B, while the concordance between radiological and surgical findings was weak (kappa values = 0.213 and 0,248, respectively and 95% CI = 0.106 to 0.319 and 95% CI = 0.142 to 0.355, respectively). When overall morbidity, mortality and the need of a terminal colostomy were considered as main endpoints no concordance was observed between surgical and radiological findings and the CDD (P=NS).

Conclusions: The need for a more accurate classification of ACD, able to better stage this emergency, and to provide surgeons with reliable information for the best treatment is advocated.