Proposal ID: 253924
Role: Cordinator
Acronym: MOSAIC
Topic: FP7-PEOPLE-2009-IEF: Marie Curie Action: "Intra-European Fellowships for Career Development
Call identifier: FP7-PEOPLE-2009-IEF

MOSAIC: Method of Osteo-fracture Study through Automatic Identification and Classification: biomechanical analysis of bone trabecular structure

Duration in months: 24
Free keywords: Bone Fracture, Trabecular Bone, Micro-CT, BMD, Meso-level, Image Processing, Registration, Classification.

Bone fractures represent a major health problem in the increasingly elder population. The correct prediction of bone fracture risk in bone disease, e.g. osteoporosis, is a mandatory issue for the prevention of traumatic events and related medical efforts. The assessment of the fracture risk is based on the clinical evaluation of bone mineral density, but with an estimation error ranging from 20% to 40%. Trabecular structure is known to play an important role in the mechanical behaviour of bone, but clinical imaging does not provide sufficient information. In order to perform a systematic and in depth analysis of the trabecular framework, and its relation with the mechanical properties, micro-tomographic devices are required. Using these devices, structural analysis of bone specimens could be performed by investigating the structure of the broken region, in order to examine the differences between broken and unbroken regions on specimens that were mechanically tested. The aim of the proposed project is the development of a system that will permit the prediction of in-vitro trabecular mechanical behaviour with possible applications on clinical meso-level x-ray images. The presented system will be comprised mainly of a classifier that will identify the fracture probability of bones’ regions using estimated mechanical parameters. In order to define the classifier, a fracture identification module, based on image processing, will be also developed with the aim to locate firstly broken and unbroken regions of mechanically tested specimens, and then estimate specific mechanical parameters related to the different regions. This initial knowledge will be used to train the classifier and to predict the broken region of not tested bone specimens.