Measurement Error in Osteometric, Photographic, and Virtual Methodologies to Quantify the Torsional Profile of the Lower Limb.

OBJECTIVES: The torsional profile of the lower limb consists of femoral torsion, tibial torsion, and talar neck angle. Due to high levels of inter-individual variation and a lack of defining landmarks, these variables are difficult to precisely measure. It is important to ensure torsional profile measurement methodologies are repeatable, so that studies evaluating these variables can be compared. MATERIALS AND METHODS: Two observers collected torsional profile and linear measurements from the femur, tibia, and talus of 20 individuals using osteometric, photographic, and virtual methodologies. Intra- and interobserver error were assessed using the technical error of measurement (TEM), %TEM, and coefficient of reliability. Comparability between methods was evaluated using correlations, reduced major axis regression, and reduced mean squared error. Two methods for measuring the torsional profile were compared: a landmark method and a shape-fitting method. RESULTS: Observer error was low for linear measurements. Torsional profile measurements have higher intra- and interobserver error and lower comparability between methods than linear measurements. Shape-fitting methods for femoral torsion lowered observer error but did not improve methodological comparability. Shape-fitting methods for tibial torsion did not substantially alter observer error but improved method comparability. Shape-fitting methods for talar neck angle greatly improved method comparability, but not observer error. DISCUSSION: Linear measurements have low observer error and are highly comparable between osteometric and virtual methods. There is greater observer error and lower comparability between measurement modalities for angular measurements. Shape-fitting is a promising way to reduce observer error when measuring the torsional profile.