MR only simulation for radiation therapy – DRR generation
Y Yang*1, F Han2, M Cao1, K Sheng1, M Selch1, T Kaprealian1, N Agazaryan1, D Low1, P Hu2
1. Department of Radiation Oncology, UCLA School of Medicine, Los Angeles, CA
2. Department of Radiology, UCLA School of Medicine, Los Angeles, CA
- MRI only radiation therapy has been increasingly gaining research interests.
- Conventionally, MRIs are fused with CT to take advantage of its superior soft tissue contrast.
- Bone imaging is an important challenge that needs to be overcome. The ultra-short T2 and T1 of cortical bone result in low signal intensity, which presents difficulty for both dose calculation and patient setup in terms of digitally reconstructed radiograph (DRR) generation.
- Current solutions will either require manual bone contouring or multiple MR scans.
- To develop a technique which can potentially enable high quality MR bone imaging for radiation therapy based on MRI simulation.
- Four meningioma patients were scanned at 1.5 Tesla using a 3D radial UTE (Ultrashort Echo Time) gradient echo sequence in addition to their standard clinical MRI and CT simulation.
- The UTE sequence acquires two sets of images at two different echo times (0.07ms and 3.61ms). The two echoes are acquired asymmetrically within the same TR.
- The two images were processed using software built in house, which performs pixel-by-pixel magnitude subtraction of the two images. Further post-processing includes mask based noise removal and scaling for better visualization.
- The resultant bone images were subsequently loaded into commercial treatment planning system to generate DRRs.
- For each patient, orthogonal kV/kV pair was simulated from patient CT which was well registered with UTE bone MRI.
- Known shifts in three translational directions were applied to simulated kV/kV pair.
- Auto matching between MRI-DRR and simulated kV/kV pair was done with commercial imaging registration system.
| || |
|Figure1: Three planes of MR bone image and a DRR generated from a right lateral beam from a patient. ||Figure2: Three planes of CT image and a DRR generated from a right lateral beam from the same patient. |
| || |
a)-c) Three planes of the CT image;
d)-f) three planes of MR bone image;
g)-i) three planes of the fusion between these two images.
Figure 1 shows the bone image of one patient and the generated DRR comparing to his CT.
- The fused image of the CT image from the same patient with the MR bone image (Fig. 2) clearly demonstrates the accuracy of automatic bone identification using our technique.
- The majority of cranial, facial and vertebral bones are well visualized in all patients.
- The generated DRR is of acceptable quality but there is still remaining artifacts from tissue/air interface and brain tissue interface.
- Auto matching was used to align the simulated kV/kV pair to MR DRR for each patient. The accuracy is within 1mm with no subjective interference for all four patients.
| || |
Figure3: Auto matching between MRI bone DRR and simulated kV pair for one patient.
- This study shows the potential of cranial patient setup with on-board kV pair with single MR sequence.
- Further work will be needed on MR sequence development and post-processing procedure to achieve MR bone image with comparable quality to CT.
 Kapanen M, Tenhunen M. T1/T2*-weighted MRI provides clinically relevant pseudo-CT density data for the pelvic bones in MRI-only based radiotherapy treatment planning. Acta Oncol, 2012. doi: 10.3109/0284186X.2012.692883.
 Chen L, Nguyen TB, Jones E, Chen Z, Luo W, Wang L, Price RA, Pollack A, Ma C. Magnetic resonance-based treatment planning for prostate intensity-modulated radiotherapy: creation of digitally reconstructed radiographs. Int J Radiat Oncol Biol Phys 2007, 68: 903–911.
 Johansson A, Karlsson M, and Nyholm T, “CT substitute derived from MRI sequences with ultrashort echo time,” Med. Phys. 2011, 38: 2708–2714.