DVH-based inverse planning for LDR pancreatic brachytherapy.

Abstract

Pancreatic cancer is one of the common types of malignant cancer. Low-dose-rate (LDR) brachytherapy has shown great efficacy in curing pancreatic cancer. However, a long preoperative planning time and an insufficient tumor dose are common issues. In this paper, we present and validate a method for inverse planning using simulated annealing (SA) for the treatment of pancreatic cancer. The SA method was used for the inverse planning process. With the help of parallel computation and a quick dose field estimation algorithm. This method allowed inverse planning to be performed quickly. A novel length-control method was used to consider and limit the dose of the organ at risk. The effect of this system was validated by calculating the dose-volume histogram metric and time consumption. Regarding the percentage of the volume receiving 100% of the prescribed dose (V100), this approach yielded an average difference in V100 of 5.01% for the tumor and of 1.32% for the organ at risk in the small tumor group; in the large tumor group, the average difference in V100 was 2.3% for the tumor and - 4.49% for the organ at risk. The average time required for inverse planning was 1.63 ± 0.26s for small tumors and 3.81 ± 0.51s for large tumors. Compared with other inverse planning methods, the optimal quality of the plans yielded by this method was further improved. This paper presents a new type of inverse planning method for the treatment of pancreatic cancer based on SA. Compared with other LDR inverse planning methods, the method presented here can provide the prescribed dose to the tumor while considering the dose of the organ at risk. Also, the required time is significantly lower than other methods. All the experimental results indicate that this method is ready to be applied in further clinical studies.