CT scans do not reliably identify appendicoliths in acute appendicitis

Background Both the CODA trial and the 2020 WSES Jerusalem guidelines concluded that nonoperative management of appendicitis (NOMA) was non-inferior for uncomplicated appendicitis but appendicoliths increase risk for complications. Thus “Appendicoliths on CT” are a relative contraindication to NOMA but accuracy of CT in identifying appendicoliths is just assumed. Materials and methods We conducted an EMR review of 1552 appendectomy patients, who had pre-operative CT scans for suspected acute appendicitis from 2001-2019. Two radiologists reinterpreted images from 2011 and 2019 for presence of appendicoliths. Appendicoliths identified on CT were compared to appendicoliths in corresponding pathology specimens. Cohorts A (2001-10) & B (2011-19) were created on account of changes in CT scanners. PPV, NPV, sensitivity and specificity rates were calculated for contemporaneous readings and for 2011 and 2019 reinterpretations. Results 397 CT scans were read as positive for appendicoliths; 167 for A, 230 for B. 246 corresponding specimens had appendicoliths. PPV was 62% overall; 69% for A, 57% for B. 1155 scans were negative for appendicoliths. 191 corresponding specimens had appendicoliths; NPV was 83% overall; 80% for A, 87% for B. 447 specimens had appendicoliths; only 246 were identified by CT. Sensitivity was 55% overall; 48% for A, 66% for B. Specificity was 86% overall (964/1115); 91% for A, 82% for B. Radiologist accuracies varied. Conclusions Complicated appendicitis is often associated with appendicoliths. The assumption that CT identification of appendicoliths is reliable is unsupported by this study. CT finding of an appendicolith should not be used to exclude patients from antibiotics treatment but rather used in the shared decision-making conversation about management with patients.