Use of Clinical Data to Predict Appendicitis in Patients with Equivocal US Findings

Purpose To determine the incremental value of clinical data in patients with ultrasonographic (US) examinations that were interpreted as being equivocal for acute appendicitis. Materials and Methods Institutional review board approval, with a waiver of informed consent, was obtained for this analysis of clinical and imaging data in patients younger than 18 years old who were evaluated for acute appendicitis. Findings from US examinations were reported in a structured fashion, including two possible equivocal impressions. Clinical data were captured as Pediatric Appendicitis (PAS) or Alvarado scores and considered as categoric (high, intermediate, or low likelihood) and continuous variables to simulate stratification of equivocal US examinations to predict appendicitis. Receiver operating characteristic curves were used to define score cutoffs, and logistic regression was used to assess individual clinical variables as predictors of appendicitis. Results The study population was made up of 776 patients (mean age, 11.7 years ± 3.7), with 429 (55.2%) girls. A total of 203 (26%) patients had appendicitis. US had a negative predictive value of 96.2% and a positive predictive value of 93.3% for depicting appendicitis, with 89 of 782 (11.4%) equivocal examinations. Categoric PAS and Alvarado scores were equivocal for 59.5% (53 of 89) and 50.6% (45 of 89) of equivocal US examinations, respectively. Categoric low- and high-likelihood PAS and Alvarado scores correctly predicted the presence of appendicitis in 61.1% (22 of 36) and 77.3% (34 of 44) of equivocal US examinations, respectively. As continuous variables, a PAS or Alvarado score of 5 or lower could be used to exclude appendicitis, with a 80.8% (21 of 26) and 90% (18 of 20) negative predictive value, respectively. Conclusion The study confirms the excellent performance of US for depicting pediatric appendicitis. In the subset of equivocal US examinations, a low clinical score (≤5) may be used to identify patients with a low likelihood of having appendicitis. (©) RSNA, 2016.