Validation of a software‐based torsional measurement method of the lower limb: A retrospective study

TL;DR

This study validates a software tool for measuring lower limb torsion using CT scans, showing it is reliable and accurate compared to manual methods.

Contribution

The study confirms the reliability and validity of a software-based torsion measurement method for lower limb CT imaging.

Findings

Software-based measurements showed excellent intra-rater reliability (ICC 0.870–0.993) for both tibia and femur.

Inter-rater reliability was also high for both manual and software-based methods (ICC 0.851–0.993).

Software-based measurements demonstrated strong validity with correlation coefficients above 0.9.

Abstract

The aim of this study was to evaluate a software‐based measurement tool of computed tomography imaging to assess torsional alignment of the lower limb—usually used for patients with posttraumatic deformities and patellofemoral issues—and compare these measurements with manual measurements of two experienced raters.

This study was a retrospective analysis of 58 lower limbs (47 patients, 20 men and 27 women) aged between 19 and 97 years. Inclusion criteria were the clinical indication for torsion measurement of the lower limb. Legs with incomplete imaging and age less than 18 years were excluded. Femoral and tibial torsion were measured separately. Torsional alignment was assessed software‐based at two time points by two raters. Rater one additionally assessed manual‐based measurement at two time points. The software used was mediCAD 3D Knee Version 2.5.33 (Hectec). Subsequently, intra‐ and inter‐rater reliability was calculated using the intraclass correlation coefficient (ICC). Validity testing was performed by means of precision, concordance correlation coefficient according to Lin (ccc) and Pearson correlation coefficient.

High intra‐rater reliability for software‐based as well as manual‐based evaluation of torsional alignment was found for the tibia as well as the femur (ICC ranging between 0.870 and 0.993). Inter‐rater reliability also showed highly significant results of both manual and software‐based measurements (ICC ranging between 0.851 and 0.993). For almost all comparisons (except the tibia left), software‐based measurements showed higher ICC scores above 0.9, and, therefore, classified as ‘excellent reliability’. For validity testing, correlation coefficients and precision showed very good correspondence of the measurements (all values > 0.9), without systematic deviations.

Software‐based measurement of torsional alignment according to the measurement method developed by H.‐A. Waidelich proved to be a reliable and valid technique. Especially for inexperienced surgeons, software‐based measurement, therefore, might improve confidence in reliable medical decisions in diagnostics and treatment.

Level III.

Figures