Reliability of lower limb frontal plane alignment measurements using plain radiographs and digitized images

Abstract

This study evaluated the reliability of lower limb frontal plane alignment measures obtained from plain radiographs measured manually and digitized images measured using a custom computer software package (TheHTO Pro; Fowler Kennedy Sport Medicine Clinic, London, Ontario, Canada). Radiographic measurements used in the planning of high tibial osteotomy, including the mechanical axis angle and mechanical axis deviation, were measured on 42 hip-to-ankle radiographs on two separate occasions by two different raters (A.V.S., J.J.D.). Intraclass correlation coefficients (0.96-0.99) indicated excellent agreement between the manual and computer measurements, suggesting both methods can be used interchangeably. Although test-retest and inter-rater reliability tended to be slightly better when using TheHTO Pro, intraclass correlation coefficients were excellent for both methods (0.97-0.99). The standard errors of measurement were <1 degree for mechanical axis angle and <2 mm for mechanical axis deviation, regardless of method or rater. Based on the observed standard errors of measurement, conservative estimates for the error associated with an individual's mechanical axis angle at one point is approximately 1.5 degrees, and the minimal detectable change on reassessment is approximately 2 degrees. The error associated with an individual's mechanical axis deviation at one point is approximately 4 mm, and the minimal detectable change on reassessment is approximately 6 mm. These results suggest that manual and computer measurements of lower limb frontal plane alignment can be calculated with minimal measurement error. However, the small errors associated with both methods should be considered when making clinical decisions.