Abstract
Purpose: A new radiographic classification of the femoral trochlea was proposed by David Dejour in 1998 to quantify the severity of bony dysplasia. The purpose of this work was to evaluate the reproducibility of this classification system and to determine its contribution to the identification of trochlea with a high-risk of femoropatellar instability.
Material and methods: Nine independent observers (one resident, four junior surgeons, four senior surgeons) with no knowledge of the patient’s history read 68 strict lateral views of knees with femoropatellar instability (53 objective instabilities (OI) and 15 potential instabilities (PI)). The classification system includes four types determined with three signs: crossing (defining the dysplasia and present in all four types), supratrochlear spike, double contour. The four types are: type A crossing alone, type B crossing and spike, type C crossing and double contour, type D crossing, spike and double contour. The kappa test was used to assess reproducibility and chi square test to analyse data by category.
Results: Twenty-one radiographs were excluded by one or several observers due to insufficient quality or the impossibility to identify the signs of the new classification. Interob-server reproducibility assess on 47 radiographs was fair (kappa = 0.48). The crossing sign was identified by the nine observers on the 47 radiographs. Reproducibility of identification of the spike was good (κ= 0.62), but was fair for the double contour (κ = 0.51). there was no difference in reproducibility by level of experience of the observers. The new classification system was not correlated with severity of femoropatellar instability: presence of spike 80% OI, absence of spike 67% OI; presence of double contour 74% OI, absence of double contour 75% OI.
Discussion, conclusion: This new classification system is more reproducibly than the former 3-type system proposed by Henri Dejour. The crossing sign and the spike are the most reproducible signs. There presence is however insufficient to quantify the dysplasia and predict the severity of the femoropatellar instability. A quantitative measure of the depth of the trochlea, which shows excellent reproducibility (interclass coefficient 0.65) could be added to better quantify the morphological anomaly and determine the most adapted treatment.
The abstracts were prepared by Pr. Jean-Pierre Courpied (General Secretary). Correspondence should be addressed to him at SOFCOT, 56 rue Boissonade, 75014 Paris, France
