Intra-fraction motion until, and during, beam on for MR guided adaptive prostate treatments
PO-1898
Abstract
Intra-fraction motion until, and during, beam on for MR guided adaptive prostate treatments
Authors: David Tilly1, Samuel Fransson2,3, Martin Lundmark3, Ulf Isacsson1,3, Adam Johansson1,4,3, Nina Tilly1, Petra Witt Nyström5
1Uppsala University, Immunology, Genetics and Pathology, Uppsala, Sweden; 2Uppsala University, Surgery and Radiology, Uppsala, Sweden; 3Uppsala University Hospital, Medical Physics, Uppsala, Sweden; 4Uppsala University, Radiology and Surgery, Uppsala , Sweden; 5Uppsala University Hospital, Oncology, Uppsala, Sweden
Show Affiliations
Hide Affiliations
Purpose or Objective
The MR-Linac has the potential to reduce the safety margins for prostate radiotherapy, and thus reduce the dose to OARs, through online adaptive treatments where the anatomy of the day is segmented on images taken with the patient in treatment position. However, due to continuous organ motion, the integrity of the segmentation and thus the treatment plan may be compromised if the segmentation and replanning are too time consuming. This is of particular of interest in adaptive hypo-fractionated treatments with long time on the table and strict requirement for accuracy.
The current work is part of assessing if the current safety margin can be reduced. Therefore, the aim was to retrospectively quantify, for a patient cohort, the systematic intra-fraction prostate displacement as a function of time on the treatment table.
Material and Methods
The clinical workflow at the MR-Linac involves acquisition of T2 weighted 3D imaging for 1) PRE - online segmentation and re-planning, 2) VER - verification before treatment start, and finally 3) END close to treatment end. The images from 21 prostate radiotherapy patients, in total 112 fractions, treated with 6.1Gy x 7fx, was used to quantify the prostate displacement as a function of time on the table. A total of 203 image registrations were performed for 111 verification image and 92 post-fraction images with their respective planning images.
The prostate displacement was semi-automatically determined with multiresolution rigid registration with SimpleITK using a mask (PTV + 2cm) and cross correlation as similarity measure. Each registration was manually inspected and adjusted if necessary. The acquisition time was extracted from the Dicom meta data of the images. The displacements were then binned into time after acquisition of the PRE image.
Results
The trend is that the prostate moves 0.5 mm (0.9 mm std) to beam on, and 1.0 mm (1.1 mm std) in the S-I direction during the entire fraction, see figure 1. The displacement in L-R and A-P directions are smaller (1.2 mm std). The average time from the acquisition of PRE to beam-on was 26 minutes and to beam-off 33 minutes, and during that timeframe the average motion is very small.
Conclusion
The mean prostate displacement from the time of planning image acquisition to beam is largest in the S-I direction, approximately 0.5 mm to beam on and 1.0 mm to beam off.