A study on the accuracy of motion tracking of thoracic tumors at radiotherapy with external surrogates

Abstract

<strong>Introduction</strong>: In radiotherapy with external surrogates, exact information of tumor position is one of the key factors that improves treatment delivery. Many dynamic tumors in thorax region of patient move mainly due to respiration and are known as intra-fractional motion error that must be compensated, as well. One of clinical strategy is using Stereotactic Body Radiation Therapy with external surrogates has been proposed to compensate motion error. In this strategy, prediction model is one of main component as responsible to track tumor motion in real time. In this study, a prediction model based on the fuzzy logic concept is proposed for real time tumor tracking during tumor treatment. Several approaches have been proposed for this aim using linear and non-linear prediction models, but a fuzzy environment may be optimal due to its robustness and benefits.

<strong>Materials and Methods:</strong> The fuzzy model is configured using training dataset provided by monitoring systems with detecting tumor motion and external rib cage motion of ten patients treated with Cyberknife Synchrony System at Georgetown University Hospital. After configuring, the model is ready to trace tumor motion during therapeutic beam irradiation. In our study, we investigated the effect of the number of data clusters, fuzzy inference system type on the performance accuracy of our model. In order to assess the performance of our model, the predicted tumor motion was compared with respect to the state of the art model.

<strong>Results:</strong> In this work we utilized Root Mean Square Error (RMSE) as common available mathematical metric for illustrating the performance uncertainty error of fuzzy prediction model. As resulted, the mean value of RMSE over ten patients is 5.45 mm Moreover, by implementing sugeno method the RMSE is remarkably less than the same calculation by means of Mamdani method. Moreover, the computational time will be significantly decreased using Sugeno type. As example, the run time improvement is almost 33% for one typical patient with left lob lung cancer.

<strong>Conclusion:</strong>A fuzzy logic based prediction model was assessed in this work to predict tumor motion as real time during external radiotherapy. The benefits of implemented fuzzy inference system in model learning and its simplicity during running has made it feasible, robust and very promising for real clinical application.