Room: AAPM ePoster Library
Purpose: The uncertainty of the daily respiratory motion that exceeds its internal target volume (ITV) margin can negatively impact to target dose distribution, especially lung SBRT. This study aims to develop the daily ITV margin verification system based on markerless 4D-CBCT tumor motion detection. The system accuracy and performance were examined through a phantom study.
Methods: The tumor templates and locations at various angles were constructed from planned 4DCT and GTV contour data of the exhalation phase. The tumor motion was automatic detected from daily CBCT projection images using the defined templates matching with the normalized cross-correlation algorithm. The daily ITV margin in superior-inferior (SI) direction was then constructed to verify the daily margin to the plan. The system was developed using MATLAB. The system accuracy was tested with the phantom (CIRs Thorax Dynamic phantom) experiments by introducing various tumor motion amplitude (±8 mm, ±12 mm, and ±16 mm) and tumor sizes (1 and 3 cm). Two test scenarios including consistent and different amplitudes were verified the system accuracy as well as the accuracy of determined ITV margin converted from the tumor motion.
Results: The results of system accuracy showed the algorithm can detect the tumor motion from markerless 4D-CBCT with root mean square error (RMSE) less than 1 mm in all amplitudes and tumor sizes. For the consistent and different amplitude, the RMSE was less than 1 mm in all tumor sizes. Moreover, this proposed system can determine the daily ITV margin from markerless 4D-CBCT with the RMSE less than 1 mm.
Conclusion: The markerless 4D-CBCT based daily ITV margin verification can benefit to the patient position setup process in prior to treatment delivery. Beside a patient position verification, the patient respiratory motion pattern, as daily ITV margin, can be verified through CBCT without additional equipment setup.
Cone-beam CT, Templates, Setup Verification