Abstract Objective To evaluate the classification proposed by David Dejour to describe trochlear dysplasia of the knee through inter- and intraobserver reproducibility measurements. Methods Ten patients with trochlear dysplasia were studied. Three physicians, members of the Sociedade Brasileira de Cirurgia do Joelho (Brazilian Society of Knee Surgery), were invited to evaluate the images. Intra- and interobserver analyses were performed at one-week intervals.

Reproducibility was evaluated in four scenarios:

using only radiography; using radiography and tomography; using radiography and consulting the classification; and using radiography and tomography, consulting the classification. Results The intraobserver evaluation presented discordant results. In the interobserver analysis, the degree of agreement was low for the analyses that used only radiography and excellent for those in which both radiography and tomography were used. Conclusion The Dejour classification presented a low intra- and interobserver reproducibilitywhenonly the profile radiography was used. It was demonstrated that the use of the radiography alone for classificationmay generate lack of uniformity even among experienced observers. However, when radiography and tomography were combined, reproducibility improved.