Accuracy of computed tomography in predicting appendiceal perforation

Abstract

Background: Some surgeons use nonoperative management with or without interval appendectomy for patients who present with perforated appendicitis. These strategies depend on accurately delineating perforation by computed tomography (CT). Since 2005, our institution has used an evidence-based definition for perforation as a hole in the appendix or fecalith in the abdomen. This has been shown to clearly separate those with a high risk of abscess from those without. To quantify the ability of CT to identify which patients would meet these criteria for perforation, we tested 6 surgeons and 2 radiologists who evaluated blinded CT scans.

Methods: A junior and senior surgical residents, 2 staff interventional radiologists, and 4 attending pediatric surgeons with 3 to 30 years of experience reviewed 200 CT scans of pediatric patients who had undergone a laparoscopic appendectomy. All CT scans were reviewed electronically, and the reviewers were blinded to the results, outcome, and intraoperative findings. None of the patients had a well-formed abscess on CT. The reviewers were asked to decide only on perforated or nonperforated appendicitis according to our intraoperative definition. Clinical admission data were reviewed and compared between groups.

Results: In total, the reviewers were correct 72% of the time with an overall sensitivity of 62% and a specificity of 81%. The overall positive predictive value was 67%, and the negative predictive value was 77%.

Conclusions: This study shows that in the absence of a well-formed abscess, the triage of patient care based on a preoperative diagnosis of perforation from CT may be imprudent and subject a portion of the population to an unnecessarily prolonged course of care.