Algorithms For Determining Leaf Trajectories With Consideration Of Marker Visibility

Background and Objectives:Human tumors are mostly in the thorax and abdomen (such as lung cancer and prostate cancer), and so the tumor positions would be affected by the respiratory motion and the gastrointestinal peristalsis during radiotherapy (RT) treatment delivery. Recently, a numbe of investigations have been carried out to solve the target motion problem, and show that four-dimensional radiotherapy (4D RT) and real-time tracking radiotherapy are both promising methods to handle this problem. The purpose of this study was to combine these two methods, and to develop the algorithms for determining leaf trajectories with consideration of marker visibility. This study mainly focused on:(A) considering marker visibility during leaf sequencing for segmental intensity-modulated radiation therapy (SMLC-IMRT). For real-time tracking radiotherapy, fiducial markers need to be implanted into or nearby the patient’s target, and then the markers will be monitored by using the imaging system mounted on LINAC (such as EPID). Due to the different apertures shaped by MLC during IMRT treatment delivery, the detection probability of fiducial marker needs to be as high as possible, that is, the visible time of markers on EPID needs to be as long as possible. In order to quantify the detection probability, this study proposed the concept of marker visibility, that is, the percentage ratio of the visible time of markers to the beam-on time;(B) considering marker visibility for dynamic IMRT (DMLC-IMRT). This research was related to the previous one, and the objective was also to maximize the marker visibility.Materials and Methods:We firstly developed the optimization models and the algorithms’flowcharts of these studeis, and then programmed in MATLAB language. The proposed optimization algorithm was evaluated by6randomly generated test fields containing1or3markers (large size field:20×20; middle size field:10×10;small size field:5×5) and15clinical test fields containing3markers (3clinical cases of prostate cancer with5-fields treatment planning). Results:The evaluation results of the optimal algorithm for SMLC showed that:a) the total delivered intensities were kept constant (i.e., beam-on time was not increased);b) the marker visibility was maximized after the leaf sequence optimization. The results for DMLC also showed that the marker visibility was maximized after the leaf trajectory optimization without increasing the total delivered intensities. Moreover, the computation efficiency was very high for both optimization algorithms (less than1s for every test field).Conclusions:This work proposed the concept of marker visibility which to quantify the detection probability of fiducial markers during the real-time tracking radiotherapy treatment delivery. The marker visibility could be maximized by using the optimization algorithms for both SMLC and DMLC mode. Moreover, these algorithms were feasible and high efficient, and could be the theoretical basis of developing real-time tracking radiotherapy.