CT-based classification systems for intra-articular calcaneal fractures: the inter- and intra-observer variations as well as integrality

Background: Several primary fracture classification systems (FCSs) have been widely used for intra-articular calcaneal fractures. The purpose of this study was to measure the inter- and intra-observer variations as well as integrality of the Zwipp, Crosby-Fitzgibbons, Sanders, and Eastwood-Atkins classification systems based on more accurate CT scans. Methods: 549 patients with intra-articular calcaneal fractures taken from a database in our level-I trauma centre (3 affiliated hospitals) were included from January 2018 to December 2019. For each case, normative CT (1 mm slices) scans were available. Four different observers reviewed all CT scans two times according to these 4 most prevalent FCSs within a 2-month interval. For these four FCSs, the kappa [κ] coefficient was used to evaluate interobserver reliability and intraobserver reproducibility, and the percentage that can be classified was used to indicate integrality. Results: The κ values were measured for Zwipp (κ= 0.38 interobserver, κ= 0.61 intraobserver), Crosby-Fitzgibbons (κ= 0.48 interobserver, κ= 0.79 intraobserver), Sanders (κ= 0.40 interobserver, κ= 0.57 intraobserver), and Eastwood-Atkins (κ= 0.44 interobserver, κ= 0.72 intraobserver). Furthermore, the integrality were calculated for Zwipp (100 %), Crosby-Fitzgibbons (100 %), Sanders (92 %) as well as Eastwood-Atkins (89.6 %). Conclusion: Compared with previous literatures, CT scanning with higher accuracy can significantly improve intraobserver reproducibility of Zwipp and Eastwood-Atkins FCSs, but it has no positive effect on variability of Sanders FCS and interobserver reliability of Crosby-Fitzgibbons FCS. While in terms of integrality, Zwipp and Crosby-Fitzgibbons FCSs appear to be superior to the other two FCSs.