Retrospective case-control study to predict a potential underlying appendiceal tumor in an acute appendicitis context based on a CT-scoring system.

ABSTRACT

OBJECTIVES: To assess CT signs to discriminate an appendiceal tumor versus a non-tumoral appendix in an acute appendicitis context. METHODS: A 10-year bicentric retrospective case-control study was performed in adults. Patients with a histopathological appendiceal tumor and appendicitis were paired for age and sex with patients with non-tumorous appendicitis (1/3 ratio, respectively). Two senior radiologists blindly analyzed numerous CT findings with final consensus to perform univariate and multivariate statistical analyses. A diagnostic CT scan score was calculated with a bootstrap internal validation. Reproducibility was assessed based on the kappa statistic. RESULTS: A total of 208 patients (51 +/- 21 years; 114 males) were included (52 patients in the tumor group and 156 in the non-tumor group). In the multivariate analysis, an appendicolith and fat stranding were protective factors with OR = 0.2 (p = 0.01) and OR = 0.3 (p = 0.02), respectively, while mural calcifications (OR = 47, p = 0.0001), an appendix mass (OR = 7.1, p = 0.008), a focal asymmetric wall abnormality (OR = 4.9, p = 0, 001), or a ≥ 15 mm diameter (OR = 3.5, p = 0.009) were positive predictive factors of an underlying tumor. Using a ≥1 cut-off, our diagnostic score had an AUC = 0.87 (95 % CI, 0.82-0.93) and a positive likelihood ratio = 13.5 (95 % CI, 6.7-27.1). CONCLUSION: We developed a reliable scoring system based on CT findings, which is highly predictive of an underlying appendiceal neoplasm in an appendicitis context using a ≥1 cut-off.