Agreement between semiautomatic and manual measurement of selected parameters on weight-bearing computed tomography images in total ankle replacement: a retrospective study

Abstract

Objective: To assess the agreement between semiautomatic 3D measurements and manual measurements derived from WBCT images in patients with end-stage ankle osteoarthritis (AO) who underwent total ankle replacement (TAR). Methods: In this retrospective, IRB-approved study (ID #201904825), we evaluated patients who underwent TAR via the lateral trans-fibular approach for end-stage ankle OA. The study included 14 ankles from 14 patients. Raw multiplanar data were analyzed using CubeVue® software. Lateral talar station (LTS) was obtained in the sagittal plane, while hindfoot moment arm (HMA) and talar tilt angle (TTA) were calculated in the coronal view. Semiautomatic 3D measurements were performed using Disior® software. Intra-rater reliabilities were analyzed using the intraclass correlation coefficient (ICC). Agreement between methods was tested with Bland-Altman plots. Each measurement was assessed using the Wilcoxon signed-rank test. Alpha risk was set to 5% (α=0.05). P-values of ≤0.05 were considered significant. Results: ICC-measured reliabilities ranged from moderate to almost perfect for manual and semiautomatic WBCT measurements in the preoperative and postoperative groups for HMA and LTS. There was high correlation between parameters calculated from manual and semiautomatic measurements, and strong agreement between the readers and software in both groups. Conclusions: Manual (M) and semiautomatic (SA) 3D measurements expressed excellent agreement for pre- and postoperative groups, indicating a high correlation between the parameters calculated and strong agreement between the readers and the software in both groups. Level of Evidence III; Therapeutic Studies; Comparative Retrospective Study.