Vienna, Austria

ESTRO 2023

Session Item

Monday
May 15
10:30 - 11:30
Strauss 1
Novel applications of MR imaging
Aswin Hoffmann, Germany;
Casper Beijst, The Netherlands
Proffered Papers
Physics
11:10 - 11:20
Delivered dose reconstruction for MR-guided SBRT of pancreatic tumors with fast 3D cine MRI
Guus Grimbergen, The Netherlands
OC-0780

Abstract

Delivered dose reconstruction for MR-guided SBRT of pancreatic tumors with fast 3D cine MRI
Authors:

Guus Grimbergen1, Giulia Pötgens2, Hidde Eijkelenkamp1, Bas Raaymakers1, Martijn Intven1, Gert Meijer1

1University Medical Center Utrecht, Department of Radiation Oncology, Utrecht, The Netherlands; 2Eindhoven University of Technology, Department of Biomedical Engineering, Eindhoven, The Netherlands

Show Affiliations
Purpose or Objective

Intrafraction motion can be a concern for safe and efficient SBRT of pancreatic tumors. In MR-guided SBRT, the tumor motion during beam-on is typically monitored with (interleaved) 2D cine MRI. However, the tumor surroundings are not captured in these images, and tumor motion might be distorted by through-plane movement. In this study, the feasibility of using highly accelerated 3D cine MRI to reconstruct the delivered dose during MR-guided SBRT was assessed.

Material and Methods

A 3D cine MRI sequence was developed for fast, time-resolved 4D imaging. The sequence features a low spatial resolution (5x5x6 mm³), allowing for rapid volumetric imaging at 430 ms. The 3D cines were acquired during the entire beam-on time of 23 fractions, delivered in five different patients that underwent online adaptive MR-guided SBRT for pancreatic tumors in a 5x8 Gy regimen on a 1.5 T MR-Linac. For every cine dynamic, a 3D deformation vector field (DVF) was extracted using deformable image registration. The linac log files were used to calculate the partial dose plans that had been delivered in the time interval between consecutive cine acquisitions. Each partial dose plan was warped with the corresponding DVF using energy-mass transfer, and afterwards summed to obtain a total delivered dose (see fig. 1). Tumor motion was determined by warping the GTV contour with the same DVF. Key DVH parameters of the GTV and neighboring OARs (duodenum, small bowel and stomach) were calculated in the delivered dose and compared to the planned dose.


Results

The relative DVH parameters of the GTV, duodenum, small bowel and stomach are presented in fig. 2, with respect to the measured GTV respiratory amplitude and drift magnitude. The mean (SD) D99% of the GTV was 0.94 (0.06), and the mean (SD) D0.5cc of the duodenum, small bowel and stomach were respectively 0.98 (0.04), 1.00 (0.07) and 0.98 (0.06). Upon closer inspection of the fractions where the largest degradation in tumor coverage was measured, it was observed that significant lateral drifts (> 4 mm) had occurred during these fractions. The largest relative OAR dose increase of 122% in small bowel D0.5cc turned out to be a negligible absolute increase from 11.8 Gy to 14.3 Gy (scaled to the complete 40 Gy treatment regimen).


Conclusion

We employed a fast, real-time 3D cine MRI sequence for dose reconstruction in the upper abdomen. In this study, we demonstrated that the motion fields, acquired directly from these images, can be used for dose warping. The reconstructed delivered dose showed only a modest degradation of tumor coverage, mostly attainable to baseline drifts. This emphasizes the need for motion monitoring and development of intrafraction treatment adaptation solutions, such as baseline drift corrections.