Integrating active shape models into ultrasound elastography to diagnose musculoskeletal injuries: A 2D simulation study

We build on ultrasound elastography (UE) by offering a new method for diagnosing musculoskeletal injuries from estimated tissue displacements. Our strategy is to isolate the portion of tissue displacements that arise due to injury. Active shape models are constructed capturing displacement variation among normal tissue. New tissue is then evaluated by estimating displacements with (1) the active shape models and (2) a traditional UE tracking algorithm. The difference between the two estimates defined virtual axial displacement and used to identify injured tissue. Our method was tested by simulating planar tissue examined with ultrasound elastography. Images are presented of axial displacement and virtual axial displacement as well as axial strain and virtual axial strain, i.e. partial derivative of the respective displacements with respect to the axial coordinate. Injured tissue and uninjured tissue were not statistically different when comparing mean absolute value of axial strain covarying with the loading conditions. In contrast, uninjured tissue and injured tissue were statistically different when comparing absolute value of virtual axial strain covarying with loading conditions (p <; 0.0001). Statistical significance was considered p <; 0.05.