Tumor tracking by integrating multiple sensing results for radiation therapy

摘要

In radiation therapy, X-ray image-based tumor tracking is a major technique for managing a moving tumor induced by patient's respiration. However, the tumor in X-ray images is often unclearly captured due to mainly the overlapping of other internal structures/tissues, and it leads to the degradation of tracking accuracy. In this study, we present a tumor tracking method based on multiple sensing results including not only the tumor directly but also both the other respiration-induced moving structures/tissues taken by X-ray fluoroscopy and body surface monitored by an IR camera. These sensing results are then combined by a regression technique for improved accuracy. The experimental results using clinical datasets showed that the tracking errors of the proposed method and a conventional tracking method with a single sensing result were 0.67 mm and 1.30 mm, respectively. These results suggest that the effective integration of different sensing results can improve the tracking accuracy.