Impact of two- and three-dimensional computed tomography use on intraobserver and interobserver reliabilities of pilon fracture classification and treatment recommendation - 19/09/19

Doi : 10.1016/j.otsr.2019.07.011

Seong-Eun Byun ^a, Wonchul Choi ^a, Youngrak Choi ^a,⁎ , Tae-Keun Ahn ^a, Hyung Kyung Kim ^b, Sangchul Yoon ^c, Jongwook Lee ^d, Dae-Sung Choi ^a
^a Department of Orthopedic Surgery, CHA Bundang Medical Center, CHA University, 59, Yatap-ro, Bundang-gu, Seongnam-si, Gyeonggi-do, Republic of Korea
^b Department of Pathology, Kyung Hee University, Hospital of Gangdong, Seoul, Republic of Korea
^c Center for Global Health and Innovation, National Medical Center, Seoul, Republic of Korea
^d Department of Applied Economics, University of Minnesota, Minneapolis, USA

^⁎Corresponding author.

Sous presse. Épreuves corrigées par l'auteur. Disponible en ligne depuis le Thursday 19 September 2019
Cet article a été publié dans un numéro de la revue, cliquez ici pour y accéder

Abstract

Background

Two-dimensional (2D) and three-dimensional (3D) computed tomography (CT) have been increasingly used in various intra-articular fractures including pilon fracture. However, no study has investigated intraobserver and interobserver reliabilities of pilon fracture classification using 3D CT images.

Hypothesis

(1) Intraobserver and interobserver agreements of fracture classifications and treatment recommendations will improve by using 2D CT images compared to using plain radiographs only; (2) agreements will improve by adding 3D CT images compared to adding 2D CT images; and (3) agreements of orthopedic residents rather than specialists will be influenced more by imaging modality.

Materials and methods

Ten orthopedic specialists and 10 residents completed a survey to classify the fractures according to the Rüedi-Allgöwer and AO/OTA classifications and to select treatment options using 25 pilon fracture images. The survey was conducted using plain radiographs, with 2D and 3D CT images introduced 3 and 6weeks later, respectively. Kappa coefficients were calculated to determine reliabilities.

Results

Intraobserver reliabilities for fracture classifications in specialists significantly improved by using 2D images compared to using plain radiographs only. Addition of 3D CT did not significantly improve intraobserver reliabilities compared to those with 2D CT. Use of 2D CT images significantly improved overall interobserver agreement of both classifications, with the improvement being greater for residents. Use of 3D CT images did not improve the interobserver reliability of both classifications. Overall interobserver reliabilities for treatment recommendations did not significantly differ according to the imaging modality. However, interobserver agreement among residents significantly improved from slight agreement using radiographs only to fair agreement using 2D CT images.

Discussion

Intraobserver and interobserver reliabilities of pilon fracture classification and treatment recommendations did not improve between using 3D and 2D CT. Using 2D CT images improved the intraobserver and interobserver reliabilities of the fracture classifications in specialists and the interobserver reliabilities of the fracture classifications and the treatment recommendations in residents.