Assessment of Setup Accuracy for Prostate Radiotherapy using Onboard imaging and Conebeam Computed T

1
Assessment of Setup Accuracy for Prostate
Radiotherapy using On-board imaging and Cone-beam
Computed Tomograph

• Duke University Medical Center
• Department of Radiation Oncology
• Sua Yoo, Fang-Fang Yin, Devon Godfrey, Mark
  Oldham, and Shiva Das

2
Introduction

• Accurate patient setup is essential for
  delivering radiation therapy.
• Conventional setup methods rely on external
  markers or bony anatomy as seen on 2D orthogonal
  radiographic views.
• Cone-beam Computed Tomography (CBCT) provides 3D
  soft tissue information utilizing On-board imager
  (OBI).
• This study investigates the extent of mismatch of
  the soft tissue in CBCT compared to that in the
  planning CT for the prostate treatment, after
  shift corrections based on 2D bony anatomy in
  orthogonal radiographs.

3
Protocol for 2D setup verification
CT simulation
Recorded In Vision
Eclipse - RTP
Treatment
Vision RV
Lat MV using PV Gantry _at_270o
AP MV Gantry _at_0o
4DTC
No Shift
AP MV using PV Lat KV using OBI Gantry _at_0o
2D/2D Match
Couch shift
Shift
4
2D/2D Match
Matching tool - shift KV/MV images to match with
DRR
DRR seen through a blend window
DRR seen through a blend window
MV image on top of DRR
KV image on top of DRR
Lat KV
AP MV
Lat DRR
AP DRR
Couch shift info based on matching
Pass couch shift info to Clinac
5
Protocol for 3D setup verification
Recorded In Vision
CT simulation
Q Shift based on - Soft-tissue? - Bony anatomy?
Treatment
Eclipse - RTP
3D/3D Match
Vision RV
4DTC
CBCT
Lat MV using PV Gantry _at_270o
AP MV Gantry _at_0o
AP MV using PV Lat KV using OBI Gantry _at_0o
No Shift
2D/2D Match
Couch shift
Shift
6
3D/3D Match
Matching tool - shift CBCT to match with planning
CT
Contour of prostate drawn on CT in Eclipse
CBCT on top of planning CT
Planning CT
Transversal
Frontal
Sagittal
Couch shift info based on matching
Pass couch shift info to Clinac
7
Methods and materials

• Setup the patient on the treatment couch based on
  tattoos, laser, cross hair, SSD etc
• Take AP MV/LAT KV
• Analyze 2D/2D Match based on bony anatomy match
  (2DBone) ? Record shift
• Shift the couch based on 2D/2D Match
• Take CBCT
• Analyze 3D/3D Match based on bony anatomy match
  (3DBone) ? Record shift
• Do not apply the 3DBone shift, but reset the
  shift
• Analyze 3D/3D Match based on soft tissue match
  (3DST) ? Record shift
• 13 patients were set on the couch on the first
  day of treatment without simulation, then 2DBone,
  3DBone and 3DST were acquired.
• 1 patient (patient M) was followed up 10
  sessions, but 2D/2D Match was not acquired.
• 1 patient (patient L) was followed up 5 sessions

8
Methods and materials

• 2DBone
• Shift information based on bony anatomy match
  using 2D orthogonal images
• This demonstrates how accurate the initial setup
  could be without 2D imaging
• 3DBone
• Shift information based on bony anatomy match
  using 3D CBCT
• This is acquired after 2DBone is applied for
  correction.
• This demonstrates residual corrects to be made
  after 2DBone. Thus, this demonstrates how
  accurate the 2D images perform setup
  verification.
• 3DST
• Shift information based on soft tissue match
  using 3D CBCT
• This demonstrates variations between matching
  based on bony anatomy and matching based on
  soft-tissue.

9
Results 13 patients
Patient in head first supine position Ver
Anterior/ Posterior, Lng Inferior/
Superior, Lat Left/ Right
10
Results 13 patients
11
Results Patient M and Patient L
Patient M
Patient L
12
Results patient M and patient L

• Patient M
• 10 sessions
• 2DBone available
• only for 1st session

• Patient L
• 5 sessions

13
ResultsDisagreement between 3DBone and 3DST
14
ResultsDisagreement between 3DBone and 3DST
13 Patients
Patient M
Patient L
Patient
Session
Session
15
Discussion

• 2DBone data show large shifts (-0.8cm to 1.0cm)
  to all directions, which means that setting up a
  patient without imaging is not reliable.
• 3DBone data show small shifts (-0.5cm to 0.8cm)
  because the patients had already been
  repositioned based on 2DBone. However, 3DBone
  data are not zero, which means that setting up a
  patient based on the bony anatomy in 2D is not
  the same as in 3D. Only few cases have shift
  less than 0.2cm to all direction.
• 3DST data show a wide range of shifts (-0.4cm to
  1.3cm) even though the patients had already been
  repositioned based on 2DBone.
• The difference between 3DBone and 3DST represents
  disagreement between bony anatomy and the soft
  tissue, which is the prostate, the target of the
  treatment.
• Note that 3DBone or 3DST can not match anything
  completely by shifting the couch to only three
  directions because of rotation, organ
  motion/deformation, patient condition (weight
  loss/gain), and so on.

16
Discussion

• Matching CT and CBCT based on bony anatomy
  cannot be accomplished due to rotation

Planning CT
Planning CT
They match well.
CBCT
They dont match.
CBCT

• Matching CT and CBCT based on soft tissue cannot
  be accomplished due to organ motion and
  deformation.

Planning CT
CBCT

• Rectum and bladder were full during the CT
  scanning, but almost empty during the CBCT
  scanning and treatment. The prostate and seminal
  vesicle change in their positions and shapes
  significantly.

Planning CT
CBCT