Fragilidad relativa de fémures osteoporóticos evaluados con DXA y simulación de caídas con elementos finitos guiados por radiografías de urgencias

Ruiz Wills, Carlos; Tassani, Simone; Di Gregorio, Silvana; Martínez, Silvia; González Ballester, Miguel Ángel, 1973-; Humbert, Ludovic; Noailly, Jérôme; del Río Barquero, Luis Miguel

doi:http://dx.doi.org/10.4321/S1889-836X2020000200005

Fragilidad relativa de fémures osteoporóticos evaluados con DXA y simulación de caídas con elementos finitos guiados por radiografías de urgencias

Llicència

Citació

Ruiz Wills C, Tassani S, Di Gregorio S, Martínez S, González Ballester MA, Humbert L, Noailly J, Del Río LM. Fragilidad relativa de fémures osteoporóticos evaluados con DXA y simulación de caídas con elementos finitos guiados por radiografías de urgencias. Rev Osteoporos Metab Miner. 2020 Feb 10;12(2):62-70. DOI: 10.4321/S1889-836X2020000200005

Resum

Objetive: The diagnosis of osteoporosis has been based on the measurement of bone mineral density, although this variable has a limited capacity in discriminating patients with or without fractures. The application of finite element analysis (FE) on computed tomography volumetric images has improved the classification of subjects by up to 90%, although the radiation dose, complexity, and cost do not favor their regular practice. Our objective is to apply FE analysis to threedimensional models with dualenergy xray absorptiometry (3DDXA), to classify patients who present osteoporotic fracture of the proximal femur and those without fracture. Material and methods: A cohort of 111 patients with densitometric osteoporosis was selected: 62 with fracture and 49 without it. Subjectspecific FE models for impact were used, such as static simulation of lateral fall. Impact simulations allow identifying the critical region in 95% of cases, and the mechanical response to maximum lateral force. An analysis was performed using a discriminative classifier (Support Vector Machine) by fracture type, tissue and gender, using DXA measurements and biomechanical parameters. Results: The results showed a classification sensitivity of 100%, and a false negative rate of 0% for cases of neck fracture for trabecular bone in women. The variable major main stress (MPS) is identified as the best parameter for the classification. Conclusion: The results suggest that using 3DDXA models help in order to better discriminate patients with raised fracture risk.