Evaluation of intrafraction duodenum positional stability for pancreatic SABR on the MR-Linac
PD-0748
Abstract
Evaluation of intrafraction duodenum positional stability for pancreatic SABR on the MR-Linac
Authors: Mairead Daly1, Ananya Choudhury1,2, Robert Chuter1,3, Beth Erickson4, William Hall4, Alan McWilliam1,3, Eliana Vasquez Osorio1, Ganesh Radhakrishna2, Helena Verkooijen5, Cynthia Eccles1,6
1The University of Manchester, Division of Cancer Sciences, Manchester, United Kingdom; 2The Christie NHS Foundation Trust, Clinical Oncology, Manchester, United Kingdom; 3The Christie NHS Foundation Trust, Medical Physics & Engineering, Manchester, United Kingdom; 4Medical College of Wisconsin, Department of Radiation Oncology, Milwaukee, USA; 5University Medical Center Utrecht, Imaging Division, Utrecht, The Netherlands; 6The Christie NHS Foundation Trust, Radiotherapy, Manchester, United Kingdom
Show Affiliations
Hide Affiliations
Purpose or Objective
The duodenum is a dose-limiting organ at risk (OAR) in pancreatic stereotactic ablative radiotherapy (SABR). MR-Linac SABR delivery can be associated with longer treatment times than conventional Linacs, increasing the possible impact of intrafraction motion on planned dosimetry. The purpose of this study is to evaluate the intrafraction stability of the duodenum, the impact of treatment duration on stability, and whether a 5 mm planning at risk volume (PRV) is adequate.
Material and Methods
Pancreatic SABR patients with intact duodenum treated with abdominal compression on the MR-Linac under an international prospective registry study (NCT04075305) were included. 3D T2-weighted or motion compensated MRIs were acquired before treatment (MRIpre) and either before treatment delivery (verification) or post-treatment (MRIpost). Intrapatient sequences were consistent. The duodenum was contoured on all images by one observer and divided into 4 anatomical segments (Fig. 1a). We assessed intrafraction duodenum changes by 1) difference in volume, 2) displacement of centre of mass (COM) of duodenum and individual segments and 3) mean distance to agreement (mDTA) of whole duodenum on MRIpre (Dpre) and on MRIpost (Dpost). Time between images was calculated for each patient on all fractions.
A 5 mm PRV was generated using Dpre, and then rigidly copied to MRIpost. The volume of underlap of Dpost with the Dpre PRV was calculated. The underlap between this volume and the planning target volume (PTV) expanded by 2cm was evaluated to determine whether the duodenum was displacing towards the high-dose region. The difference in pooled mean absolute 3D vector of segment COM of Dpre and Dpost was tested for significance with single factor ANOVA, and association with time between scans assessed using Pearson’s correlation.
Results
15 patients received 33-40 Gy in 5 fractions (#). MRIpost treatment/verification data were not available for 3 patients for 1 # each. These were excluded from analysis, leaving 72 evaluable fractions.
There was evidence of some intra-fraction duodenal drift; the number of fractions where Dpost was outside of PRV, and number of # where this underlap was within 2cm of PTV are shown in Table 1. No strong association was found between time between scans and duodenum COM displacement in X, Y or Z planes (r=-0.17, 0.02, -0.04 respectively), nor for absolute 3D vector of displacement (r=0.15) (Fig. 1B). The difference in pooled mean absolute 3D displacement vector between segments was not significant (p=0.6).
Conclusion
Whole and sub-organ duodenum position and volume is stable for the majority of patients across a treatment fraction of approximately 40mins, with a 5mm PRV being adequate within 2 cm of PTV. Longer treatment durations may increase the risk of motion so position should be monitored throughout. Future work is required to evaluate individual patient factors associated with increased duodenum motion.