Radiographic Classification of Coronal Plane Femoral Deformities in Polyostotic Fibrous Dysplasia

Background: Fibrous dysplasia of bone is a skeletal dysplasia with a propensity to affect the femur in its polyostotic form, leading to deformity, fracture, and pain. The proximal femur is most commonly involved with a tendency to distal progression, thereby producing the typical shepherd's crook deformity. However, there are few data on the spectrum and progression of femoral deformities in polyostotic fibrous dysplasia.

Questions/purposes: The purposes of this study were (1) to develop a radiographic classification for polyostotic fibrous dysplasia; (2) to test this classification's intra- and interobserver reliability; and (3) to characterize the radiographic progression of polyostotic fibrous dysplasia in a population of patients with the condition who were treated with a variety of approaches at several centers.

Methods: We retrospectively reviewed radiographs of 127 femurs from 84 adult patients affected by polyostotic fibrous dysplasia. Fifty-nine femurs had undergone one or more operations. The radiographs were evaluated in the coronal plane for neck-shaft angle and angular deformities along the whole femoral shaft down to the distal epiphysis. Four observers evaluated each film two times at intervals; intra- and interobserver reliability testing was performed using the kappa statistic. Eighty-nine femurs (70%) were available for followup to evaluate for progression at a mean of 10 years (range, 6-20 years).

Results: Six reproducible patterns of deformity were identified in both untreated and operated femurs: type 1 (24%), normal neck-shaft angle with altered shape of the proximal femur; type 2 (6%), isolated coxa valga with neck-shaft angle > 140°; type 3 (7%), isolated coxa vara with neck-shaft angle < 120°; type 4 (20%), lateral bowing of the proximal half of the femur associated with normal neck-shaft angle; type 5 (14%), like in type 4 but associated with coxa valga; and type 6 (29%), like in type 4 but associated with coxa vara. Interobserver and intraoberver kappa values were excellent, ranging from 0.83 to 0.87. In 46 of the 89 femurs (52%) for which longitudinal radiographic documentation was available, there was progressive worsening of the original deformity, although the pattern remained the same; types 1 and 2 tended not to progress, whereas types 3 to 6 did.

Conclusions: A reproducible radiographic classification of polyostotic fibrous dysplasia-associated femoral deformities is proposed, which can serve as a tool for assessing and treating these deformities. After reviewing the radiographs of 127 femurs, we identified six reproducible patterns of femoral deformities.

Level Of Evidence: Level III, diagnostic study. See Guidelines for Authors for a complete description of levels of evidence.