927 Can Active Shape Modelling Be Used to Distinguish Those at Risk of Femoral Fracture in Osteoporosis?

The prevention of fractures is the primary treatment goal for osteoporosis patients (Klibanski et al., 2001), therefore the early identification of individuals who are at high risk of hip fracture is crucial and could delay incidence of fracture and the often-crippling consequences that accompany it. The analysis of the shape of the proximal femur has been previously identified as a useful tool in the prediction of hip fracture. The aim of this systematic review was to evaluate published literature which investigated the accuracy of the application of active shape modelling (ASM) on radiographs of the femur and hip joint in patients with osteoporosis in the prediction of hip fracture compared to controls. This review compiles data from studies applying ASM on hip radiographs taken from patients with osteoporosis and control groups. Initial literature search produced 120 papers but due to exclusion criteria and where full text was not available, only two papers were included in the final analysis. Both studies used a 29-point model for ASM of the femoral head. Statistical significance was found between fracture and control groups for femoral neck and trochanter in both studies when applying ASM. It was shown that combining BMD and ASM improved the accuracy of the correct identification of fracture group to 90% compared to Pshape (74%) and BMD (66%) alone. No statistical significance was found between fracture and control groups for physiological factors on bone structure. It was concluded that incorporating ASM into analysis of hip radiographs in osteoporotic patients would aid in fracture prediction.