Room: AAPM ePoster Library
Purpose: The TomoTherapy treatment planning system utilizes daily megavoltage-computed tomography (MVCT) data to monitor the effects of anatomical changes over a patient’s treatment course using deformable image registration (DIR). The purpose of this study was to evaluate the accuracy of DIR in the thoracic and male pelvic region using the TomoTherapy treatment planning system.
Methods: Our dataset consisted of four-dimensional computed tomography (CT) images of ten patients with thoracic cancer. The dataset for these patients was provided by DIR-Lab and included a coordinate list of anatomical landmarks (e.g., 300 bronchial bifurcations). The DIR accuracy was quantified in each patient by using target registration errors (TREs) for the 300 anatomical landmarks. We also analyzed MVCT scans of ten patients with prostate cancer treated with intensity-modulated radiation therapy with TomoTherapy. DIR accuracy was quantified by Dice similarity coef?cient (DSC) of the bladder and rectum.
Results: The mean ± SD TREs for the ten thoracic cancer patients was 2.17 ± 2.61 mm. The largest displacement occurred between peak-inhale and peak-exhale (3D displacement, 14.99 mm), and the registration errors were the largest with the highest standard deviations (6.74 ± 9.07 mm). The mean ± SD DSCs for the ten prostate cancer patients of the bladder and rectum were 0.84 ± 0.05 and 0.79 ± 0.04, respectively. Bladder filling and rectal gas caused small DSCs because daily MVCT images differed from the planning CT images (rectum: 0.64, bladder: 0.62).
Conclusion: This study confirmed that the DIR accuracy incorporated in the TomoTherapy treatment planning system is reasonable in the thoracic and male pelvic region. However, the accuracy of DIR may be relatively low in the case of large internal variability innate to the bladder and rectum.