Observer variability of tibial plateau slope measurement in 40 dogs with cranial cruciate ligament-deficient stifle joints.

Abstract

To determine (1) the inter- and intraobserver variability in measurement of tibial plateau angle (TPA), (2) whether this inter- and intraobserver variability is related to the characteristics of the dog (age, size, and amount of degenerative joint disease [DJD]) and the experience level of the observer, and (3) the extent of any relationship between interobserver variability of TPA and the variability of the observers' selection of the specific cranial and caudal points along the tibial plateau. Examination of tibial radiographs of 40 dogs clinically affected with a cranial cruciate ligament (CrCL)-deficient stifle joint. Eleven different observers, divided into 3 groups based on their level of experience with the tibial plateau leveling osteotomy (TPLO) technique, measured the TPA on all 40 radiographs on 5 different occasions. The degree of DJD present in the stifle joint was independently graded as an overall measure and then again as it specifically related to the cranial and caudal points along the tibial plateau. The total observed variabilities of the TPA were assessed with reference to interobserver differences, intraobserver differences, and among the groups of observers with respect to the different dog characteristics. Finally, the specific points selected on the radiographs were reexamined to determine whether any variability was present in cranial and caudal point selection. The interobserver standard deviation of the TPA measurements for each dog was 0.8 degrees, and the intraobserver standard deviation was 1.5 degrees. The TPA measurements obtained by the 11 observers differed significantly from each other (P <.001); however, there was no significant difference of TPA among the different groups of observers (P =.67). There was no significant correlation observed between either the inter- or intraobserver variability and the dog characteristics. Specific point data and their relationship to the various variables of dog characteristics and inter- and intraobserver TPA variability revealed significant correlations only to the amount of DJD present at the caudal point (P =.001). Interobserver variation, but no significant group variation, was present. Overall DJD did not appear to be related to the variability in TPA angle measurement. Most of the interobserver variability was attributable to variability in horizontal point selection at both the cranial and caudal points and vertical point selection at the caudal point. It appears that degenerative changes that specifically obscure the points on the tibial plateau, especially at the caudal point, are responsible for most of the interobserver variation. The desired postoperative TPA of 5 degrees is dependent on a precise initial measure of TPA preoperatively. This study indicates that there is statistically significant interobserver variability with measurement of TPA, which, therefore, can result in a similar amount of variability with the final tibial plateau slope obtained postoperatively.