Abstract:
Traditional path planning methods for lung biopsy are facing the unique clinical problems, such as the high risk of surgical complications, the large impact of physician proficiency, and the inability to quantitatively analyze the risk of surgical path. In order to solve the problems, a set of optimal path planning programs for lung biopsy are designed. Firstly, image segmentation for vital organs of chest is accomplished based on the chest CT images. Then, the surgical risk is quantified into 3 constraint conditions and 6 target conditions, to measure the pros and cons of different paths, and to evaluate the surgery risk of different paths. 3 constraint conditions are the puncture depth, the obstacle avoidance of vital organs and the puncture angle, and 6 target conditions are the thickness of the chest wall, the puncture length in the lung, the distance between vital organs, the distance between the extension line of the path and the vital organs, the skin penetration angle and the pleural penetration angle. Finally, the planning system selects the optimal path by using an adaptive agglomerative nesting (AGNES) algorithm to cluster the paths into
clusters and a multi-objective optimization formulation to quantitatively analyze the cluster center points. The experiment results show that the optimal path obtained by automatic planning is confirmed by the doctors to meet the surgical requirements, and all calculated optimal paths rank among the top three in the doctors' ranking. This proves the rationality and effectiveness of the path planning method, which can meet the clinical needs of path planning for lung biopsy, and can provide doctors with effective 3-dimensional visual puncture path guidance.