A deformable finite element model of the breast for predicting mechanical deformations under external perturbations

Rationale and Objectives. Live guidance during needle breast procedures is not currently possible with high-field-strength (1.5-T), superconducting magnetic resonance (MR) imaging. The physician can calculate only the approximate location and extent of a tumor in the compressed patient breast before inserting the needle, and the tissue specimen removed at biopsy may not actually belong to the lesion of interest. The authors developed a virtual reality system for guiding breast biopsy with MR imaging, which uses a deformable finite element model of the breast. Materials and Methods. The geometry of the model is constructed from MR data, and its mechanical properties are modeled by using a nonlinear material model. This method allows the breast to be imaged with or without mild compression before the procedure. The breast is then compressed, and the finite element model is used to predict the position of the tumor during the procedure. Three breasts of patients with cancer were imaged with and without compression. Deformable models of these breasts were built, virtually compressed, and used to predict tumor positions in the real compressed breasts. The models were also used to register MR data sets of the same patient breast imaged with different amounts of compression. Results. The model is shown to predict reasonably well the displacement by plate compression of breast lesions 5 mm or larger. Conclusion. A deformable model of the breast based on finite elements with nonlinear material properties can help in modeling and predicting breast deformation. The entire procedure lasts less than half an hour, making it clinically practical.