S2216
Physics - Intra-fraction motion management and real-time adaptive radiotherapy
ESTRO 2026
navigator slices. Imaging parameters provided complete thorax coverage with high spatial resolution (1.04-1.28 mm in-plane resolution) and high temporal resolution (approx. 1.5Hz). 10-minute continuous free- breathing acquisitions were acquired per volunteer. For each of 19 cases, a patient-specific network was trained using their individual 10-minute acquisitions (split 11/12 coronal slices training, 1/12 validation) using photometric loss and Jacobian regularization. Results: The framework achieved sub-second end-to-end volumetric reconstruction (<1s per 3D volume) after a 15-minute one-time, patient-specific training, versus ~5 hours for conventional sorting [3]: a >300 × speedup enabling real-time adaptive radiotherapy [1]. Quantitative validation across 19 volunteers demonstrated significant reductions in both 2D motion tracking, and 3D image reconstruction, errors (Fig. 1b-c, 2b-c). The method can capture both regular and irregular breathing, overcoming the motion underestimation artifacts inherent to binning-based methods. Maximum intensity projections (Fig. 2d) confirmed continuous vessel and bronchial tree geometry. Learned deformation fields exhibited plausible motions with Jacobian determinants near unity, indicating volume preservation.
Conclusion: This first template-free, binning-free 4D-MRI framework addresses the four primary barriers limiting clinical implementation: spatial-temporal resolution trade-offs, processing time, workflow complexity, and inadequate handling of irregular free- breathing. The combination of superior motion accuracy, anatomical fidelity, and sub-second latency positions this approach for future integration into real- time adaptive radiotherapy workflows requiring continuous volumetric monitoring for adaptive dose delivery. References: [1] Lombardo E, Dhont J, Page D, et al. Real-time motion management in MRI-guided radiotherapy: Current status and AI-enabled prospects. Radiother Oncol. 2024;190:109970.[2] Karaev N, et al. Cotracker: It is better to track together. European conference on computer vision. Cham: Springer Nature Switzerland; 2024.[3] von Siebenthal M, et al. 4D MR imaging of respiratory organ motion and its variability. Phys Med Biol. 2007;52(6):1547-1560. Keywords: 4D-MRI, Motion Management, Adaptive Radiotherapy
Made with FlippingBook - Share PDF online