Computational method for cryoprobe-layout optimization via finite sphere packing

Cryosurgery is the destruction of undesired tissues by freezing, for example, in prostate cryosurgery. Minimally-invasive cryosurgery is currently performed by means of an array of cryoprobes, each in the shape of a long hypodermic needle. The optimal arrangement of the cryoprobes, which is known to have a dramatic effect on the quality of the cryoprocedure, is based on the cryosurgeon's experience. This thesis focuses on an automated computerized technique for cryosurgery planning, in an effort to improve the quality of cryosurgery. A two-phase optimization method is applied for this purpose, based on two previous and independent developments. Phase I is based on bubble-packing method, a sphere packing method in finite domain previously used as an efficient method for finite elements meshing. Phase II is based on a force-field analogy method which is proven to be robust at the expense of a typically long runtime. Both optimization phases are used to seek an optimal layout and an optimal insertion depth of cryoprobes. Two-dimensional and three-dimensional models of the prostate and urethral warmer have been reconstructed from patient's ultrasound images. The quality of planning for each method was evaluated based on bioheat transfer simulations. Both optimization phases contributed an efficient cryosurgery planning while Phase I may be sufficient for the clinical application for most cases due to time constraints in a clinical setup. Bubble packing has been extended for planning cryosurgery with pullback procedure and the intra-operational re-planning.