Toward a generic real-time compression correction framework for tracked ultrasound.

PURPOSE: Tissue compression during ultrasound imaging leads to error in the location and geometry of subsurface targets during soft tissue interventions. We present a novel compression correction method, which models a generic block of tissue and its subsurface tissue displacements resulting from application of a probe to the tissue surface. The advantages of the new method are that it can be realized independent of preoperative imaging data and is capable of near-video framerate compression compensation for real-time guidance.
METHODS: The block model is calibrated to the tip of any tracked ultrasound probe. Intraoperative digitization of the tissue surface is used to measure the depth of compression and provide boundary conditions to the biomechanical model of the tissue. The tissue displacement field solution of the model is inverted to nonrigidly transform the ultrasound images to an estimation of the tissue geometry prior to compression. This method was compared to a previously developed method using a patient-specific model and within the context of simulation, phantom, and clinical data.
RESULTS: Experimental results with gel phantoms demonstrated that the proposed generic method reduced the mock tumor margin modified Hausdorff distance (MHD) from 5.0 ± 1.6 to 2.1 ± 0.7 mm and reduced mock tumor centroid alignment error from 7.6 ± 2.6 to 2.6 ± 1.1mm. The method was applied to a clinical case and reduced the in vivo tumor margin MHD error from 5.4 ± 0.1 to 2.9 ± 0.1mm, and the centroid alignment error from 7.2 ± 0.2 to 3.8 ± 0.4 mm.
CONCLUSIONS: The correction method was found to effectively improve alignment of ultrasound and tomographic images and was more efficient compared to a previously proposed correction.