Automatic and Optimal Trajectory Planning for Laser Interstitial Thermal Therapy in Malignant Glioma.

Abstract

Laser interstitial thermal therapy (LITT) is a novel and effective treatment for malignant glioma patients who could not undergo surgical resection. Preoperative planning for better tumor coverage and less probe use is a critical step of LITT's success, which can be challenging for surgeons as different constraints need to be considered. Here we proposed an automatic algorithm to plan trajectory combination for LITT, extending previous studies on LITT that used only one probe for treatment. Furthermore, we combined iterative optimization with Lagrangian relaxation and thus achieved both mathematic optimality and solving efficiency, which existing ablation planning research may fail to balance. The planning algorithm was evaluated in 16 cases of tumors and was compared with surgeon's planning result. In all cases our planning result achieved an average ablation rate of 99.83% and satisfied multiple constraints, outperforming the manual planning in probe usage and extent of ablation. By solving the relaxation model, our proposed method proved to be capable of reducing the number of probes to minimum with an average time cost of 39.73 seconds.