MRI-guided Treatment Planning for Skin Brachytherapy with PETRA
OC-0022
Abstract
MRI-guided Treatment Planning for Skin Brachytherapy with PETRA
Authors: Casey Y. Lee1, Evangelia Kaza1, Phillip M. Devlin2, Robert A. Cormack2, Ivan Buzurovic2
1Dana-Farber/ Brigham Women’s Cancer Center, Harvard Medical School, , Department of Radiation Oncology, Boston, USA; 2Dana-Farber/ Brigham Women’s Cancer Center, Harvard Medical School, Department of Radiation Oncology, Boston, USA
Show Affiliations
Hide Affiliations
Purpose or Objective
Freiburg Flap (FF) is the most commonly used
applicator in HDR surface brachytherapy. Treatment planning is performed on CT, as
opposed to MRI that provides superior soft tissue contrast. An optimized MRI
sequence - Pointwise Encoding Time Reduction with Radial Acquisition (PETRA) –
presented in this study demonstrated a potential to make an MR-only treatment
planning feasible. PETRA was used
to digitize FF and create an MR-only treatment plan and compared to CT.
Material and Methods
A FF (Elekta, Stockholm, Sweden) consisting of 12 catheters (24 spherical beads per catheter, 10
mm diameter) was sandwiched between 2 flat plexiglass plates (30 x 20 x 1.7 cm3). An optimized coronal 3D
PETRA sequence (TR\TE\TI 3.3\0.07\100 ms, FOV 306x306 mm2, isotropic
0.8 mm-resolution, flip angle 4°, BW 407 Hz/px) was used to acquire MR images on a 3T SIEMENS Vida stimulator with a UltraFlex
Large 18 andSpine 32 coil. The phantom was subsequently scanned on a helical CT
(115 mA, 120 kV, 1.25 mm slice thickness). Applicators were localized and plans were created to the target defined as volume at 3 mm depth in the Oncentra
Brachy (Elekta Brachytherapy, Netherlands) Treatment Planning System (TPS) for
both CT and MRI. The TPS dwell points were calculated with step size of 10 mm
from projected tip position in the TPS. MR images were registered onto CT
images via rigid registration. Catheter difference (distance between the corresponding
dwell positions between CT and MR-based digitization) were compared for each
spatial dimension and for 3D. Relative
percent dose from CT or MR-only treatment planning at a point 8, 10, 15, 20,
and 25 mm beneath each dwell point (1380 points in total) were compared.
Results
The projected catheter trajectory between
CT and MR showed good agreement (Hausdorff distance of 0.99 ± 0.36 mm for dwell
points, Fig 1a). The catheter differences
between the MR and CT-generated trajectories along each
spatial dimension for all dwell points were 0.19 ± 0.42, 0.28 ± 0.11, 0.27 ±
0.66 mm for x-, y-, and z-dimension, respectively (Fig 1b-d). The catheter difference in 3D was
0.82 ± 0.37 mm (Fig 1e). The difference
in the relative dose between CT and MR-only plans (Fig 2a) were not
statistically significant (paired t-test,
α = 0.05). Differences in the mean relative dose at each depth were all <
0.04% (Fig 2c).
Fig.
1 a) Projected catheter trajectory in CT and MR,
b-e) mean catheter differences in x, y, z, and 3D for individual catheters.
Fig.
2 a) Treatment plans generated with CT and MR, b) phantom
setup, c) mean (± SD) relative dose at each depth.
Conclusion
The MR-only digitization of FF using PETRA
sequence was shown to achieve <
1 mm accuracy compared to the CT. The optimized PETRA sequence created an MR-only
plan that provides comparable dose profile to the plan generated via CT-based
approach.