Purpose: To develop a framework that can identify optimal imaging plane to monitor tumor motion in MRI guided radiation therapy using combined motion trajectory and plan dose information.
Methods: A total of 10 patients with 6 lung and 4 liver patients underwent IMRT or SBRT were retrospectively analyzed. The target volumes were identified from the maximum inhale and exhale breathing cycles. Voxel-based motion trajectories were obtained by deforming images between these two respiratory phases. The voxelized dose gradient was also calculated between the 50%-100% prescription dose regions. A heatmap was constructed by calculating the products between these two gradient vectors, with higher values representing the motion directions more parallel to the dose gradient directions. The heatmap identified regions that could result more clinical-relevant plan failure thus close motion monitoring for those regions would be needed. Principle component analysis (PCA) was further conducted to identify the primary imaging plane that can capture major clinical-relevant motions. The tumor motion extracted from the optimal imaging angle was then compared to that monitored with the standard image plane used clinically.
Results: From the heat maps generated from all the cases, there was no specific direction or region consistently showing high values. The PCA-determined imaging planes were also patient-specific. Our proposed method was able to capture larger in-plane motion with an average of 2.8 mm(range [1.6-6.8 mm])) compared to 2.1 mm([1.2-5.7mm]) as seen in clinical used image plane. The through-plane motion was also reduced from 2.4mm [0.8-6.6mm] to 1.7mm [0.3-4.2 mm].
Conclusion: The proposed method used the respiratory tumor motion trajectory combined with plan dose to determine the optimal imaging angle before the treatment. Results showed the proposed method had overall better motion tracking performance when compared against the imaging planes used in routine clinic.
Not Applicable / None Entered.