LAUSR

OBJECTIVES We hypothesize that three-dimensional (3D) geometric analyses in weight bearing CT-images of the foot and ankle are more reproducible compared to two-dimensional (2D) analyses. Therefore, we compared 2D and 3D analyses on bones of weight-bearing and non weight-bearing cone-beam CT images of healthy volunteers. METHODS Twenty healthy volunteers (10 male, 10 female, mean age 37.5 years) underwent weight-bearing and non weight-bearing cone-beam CT imaging of both feet. Clinically relevant height and angle measurements were performed in 2D and 3D (for example: cuboid height, calcaneal pitch, talo-calcaneal angle, Meary's angle, intermetatarsal angle). Three-dimensional measurements were obtained using automated software. Intra-observer and inter-observer agreement were evaluated for all 2D measurements. RESULTS Overall intraclass correlation coefficients (ICC's) were higher than 0.750 for most 2D measurements, ranging from 0.352 to 0.995. Calcaneal pitch, angle between the first metatarsal (MT1) and proximal phalange 1, between the fifth metatarsal (MT5) and the calcaneus and heights of the sesamoid bones, navicular, cuboid and talus decreased during weight-bearing in both 2D and 3D results (p < 0.01). Meary's angle was not statistically different in 2D (p = 0.627) and 3D (p = 0.765). Higher coefficients of variation in 2D geometric analysis parameters (0.27 versus 0.16) indicate that 3D analyses are more precise compared to 2D (p < 0.01). Results of left and right feet are comparable for 2D and 3D analyses. CONCLUSION Although 2D and 3D geometrical analyses are fundamentally different, automated 3D analyses are more reproducible and precise compared to 2D analyses. In addition, 3D evaluation better demonstrates differences in bone configurations between weight-bearing and non weight-bearing conditions, which may be of value to demonstrate pathology.