Click here to

Session: Multi-Disciplinary General ePoster Viewing [Return to Session]

End-To-End Deviation Validation of Image Registration in MR-Linac Online Adaptive Radiotherapy

p yin*, Z Li, Y Yin,

Presentations

PO-GePV-M-236 (Sunday, 7/10/2022)   [Eastern Time (GMT-4)]

ePoster Forums

Purpose: Accurate image registration is required for image-guided radiotherapy (IGRT). This study evaluated the registration accuracy of positional deviations during MR-guided adaptive radiotherapy (MRgRT) and the impact of the registration results on dose delivery.

Methods: The MRgRT motion management QA phantom was used for the MR daily adaptive workflow . To investigate directional registration errors, the phantom was moved in 4 directions (1 cm in each direction) relative to the CT-simulated position. The adaptive plans were re-optimized, and the spatial deviations and DVH parameters were recorded to evaluate the reliability of the registration algorithm and the relationship between deviations and the delivered dose.

Results: The average registration deviation (ΔX, ΔY, ΔZ) was 0.02±0.06cm, 0.07±0.10cm, -0.16±0.01 cm and 0.04±0.04 cm, 0.10±0.08 cm, 0.18±0.01 cm respectively when the phantom remains stationary and motion. Deviation A was X=0.13 cm, Y=-0.12 cm, and Z=-0.14 cm when the spinal cord was used as the ROI. When the tumor target was added as ROI, there was no significant change in the LR and AP directions; however, the deviation in the IS direction was significantly better than that in spinal-cord-only registration. The registration deviations of ΔX, ΔY, and ΔZ from the original positions were -0.675 ± 0.238 mm, 1.500 ± 0.177 mm, and 0.975 ± 1.163 mm while the phantom remains stationary. There was no significant correlation between the spatial deviations (ΔX, ΔY, ΔZ) and DVH parameters.

Conclusion: The phantom registration in static mode was better than that in motion mode. There were no significant differences between the motion modes. Selecting the bone structure or combining the bone structure and soft tissue as the registration ROI was better than selecting only soft tissue. There was no strong correlation between the DVH parameters of the daily adaptive plan and registration deviations.

Keywords

Not Applicable / None Entered.

Taxonomy

Not Applicable / None Entered.

Contact Email

Share: