Margins in Radiation Therapy
DATE: Tuesday, July 28, 2009
START TIME: 07:30 AM
END TIME: 08:25AM
LOCATION: 211A
Dan Low
Dale Litzenberg
Washington University
University of Michigan
Educational Objectives
• To motivate the need for margins and review ICRU
definitions
• To provide an educational review of the technologies
and methods of measuring target volume positioning
errors.
• To review methods of determining population margins
from measured data.
• To review corrective and intervention strategies for
patients with individualized margins.
Abstract
Conformal radiation therapy has historically used margins to account for geometrical
uncertainties in the position of a target volume. The margins, in their simplest form,
are simply expansions to the shape of a treatment beam, to ensure that dosimetric
planning criteria are met in the presence of inter- and intra-fraction setup variations.
Historically, the size of the margin in a given treatment site was difficult to determine
due to the available technology and the time and effort required to obtaining
accurate data.
Consequently, margins were estimated to accommodate a
population of patients. As new technologies have emerged, target volume position
errors have become easier to measure, their accuracy has increased, and the
measurements can be made much more frequently. As the quantity and quality of
data has increased for patient populations, and even for individual patients, the
conceptual basis of employing margins has evolved. Strategies for ensuring
dosimetric coverage may now be individualized to a specific patient’s geometric
uncertainty characteristics with customized margins, or they may incorporate
geometric correction based on predefined action levels. Other strategies may
incorporate continuous monitoring to gate or modify the beam. And finally, other
strategies seek to eliminate margins by including the estimated geometrical
uncertainty into the development of the dose distribution.
Disclaimer
The content of this presentation is for general
educational purposes only.
The mention of trade names, commercial products
does not imply an endorsement on the part of AAPM
or the presenters.
The views, opinions and comments made in this
refresher course are those of the presenters and not
necessarily those of AAPM.
• To briefly review planning strategies which seek to
eliminate margins.
Department of Radiation Oncology • University of Michigan Health Systems
Conflict of Interest
Outline
I.
DA Low:
Tomotherapy
Varian Medical Systems
DW Litzenberg:
Calypso Medical Technologies
The Need for Margins
II.
Treatment Site Considerations
III.
Simulation Imaging Modalities
IV. Examples of Motion
V.
Treatment Planning
VI. In-room Guidance and Correction Strategies
VII. Off-line Adaptive Strategies
VIII. Planning without margins
Outline
I.
The Need for Margins
II.
Treatment Site Considerations
III.
Simulation Imaging Modalities
IV. Examples of Motion
V.
Treatment Planning
I. The Need for Margins
A. Systematic and random errors
B. Inter- and intra-fraction motion
C. Dosimetric criteria
D. Population and individualized margins
VI. In-room Guidance and Correction Strategies
VII. Off-line Adaptive Stratagies
VIII. Planning without margins
Department of Radiation Oncology • University of Michigan Health Systems
Systematic and Random Errors
Σ small
σ small
Σ large
σ small
Inter- and Intra-Fraction Positioning Errors
Σ small
σ large
Σ large
σ large
Inter-Fraction:
Variations caused by uncertainty in
reproducing target volume position at
isocenter each time patient is placed on
treatment table
Intra-Fraction:
Variations in position while patient is on
treatment table
Each may have Systematic and Random Errors
Systematic Error Example
Systematic Error Example
Systematic shift due to
distention on planning
CT
Department of Radiation Oncology • University of Michigan Health Systems
Importance of Image Guidance
What was Missing?
N=127, patients treated to 78 Gy between 1993 and 1998
No daily image guidance
Outcomes analyzed by rectal distention at the time of simulation
• Method for directly measuring
prostate position
• On-board CT
– CBCT
• Kilovoltage or Megavoltage
– Helical Megavoltage CT
• Markers
–
–
–
–
Imbedded markers
Planar imaging
On-board CT
Radiofrequency
MDACC: de Crevoisier et al., IJROBP, 62, 965-973, 2004
Population-Based Motion
Dosimetric Coverage
1.0
m
Dose
m
Frequency
Prescribed
Dose Level
(eg. 95%)
“Tumor”
0.5
Underdose
x
0.0
-4
-2
0
Position (cm)
2
4
Department of Radiation Oncology • University of Michigan Health Systems
Patient Specific Motion
Patient Specific Margin
Σ=0
1.0
Frequency
Frequency
PTV margin = Setup + Inter + Intra-fraction
0.5
Estimate Σ
And correct
0.0
-4
-2
0
Position (cm)
2
4
Ten Haken
Position (cm)
Ten Haken
Department of Radiation Oncology • University of Michigan Health Systems
Outline
II. Treatment Site Considerations
I.
The Need for Margins
II.
Treatment Site Considerations
III.
Simulation Imaging Modalities
A. Treatment Position
B. Immobilization & Localization
IV. Examples of Motion
V.
Treatment Planning
VI. In-room Guidance and Correction Strategies
VII. Off-line Adaptive Strategies
VIII. Planning without margins
Treatment Position
Accuracy of Localization
Patient Position
Motion
Management
Immobilization
Measurement
Correction
Localization
Simulation &
Planning
1997: No in-room soft tissue guidance available,
Little to no evidence of respiratory motion.
Zelefsky, IJROBP, 1997;37:13-19.
Department of Radiation Oncology • University of Michigan Health Systems
Treatment Position
Intrafraction Motion: Fluoroscopy
Supine Positioning Results
4
2
2
0
-2
-2
-4
LR
AP
IS
Position (cm)
Position (cm)
Supine Positioning Results
8
9
10
LR
AP
IS
0
-2
Ave
Average
Position (cm)
Maximum
Initial Position
7
Patient #
+/- 1σ
2
-4
Initial Position
7
4
9
10
Ave
10
Ave
Patient #
4
2
0
-2
-4
Minimum
8
Final Position
7
8
9
Patient #
Department of Radiation Oncology • University of Michigan Health Systems
Supine vs Prone Treatment Position
• Respiratory motion reduced when supine
Dawson, IJROBP, 2000
Kitamura, IJROBP, 2002
• Significantly less dose to small bowel,
bladder and rectum when supine
Bayley, Radiother & Oncol, 2004
• ~ 95% treated supine
ASTRO IGRT Symposium, 2006
Treatment Position
• Steenbakkers RJHM, Duppen JC, Betgen A, et al. Impact of knee support and shape of tabletop on
rectum and prostate position. International Journal of Radiation Oncology*Biology*Physics
2004;60:1364-1372.
• Adli M, Mayr NA, Kaiser HS, et al. Does prone positioning reduce small bowel dose in pelvic
radiation with intensity-modulated radiotherapy for gynecologic cancer? International Journal of
Radiation Oncology*Biology*Physics 2003;57:230-238.
• Algan Ö, Fowble B, McNeeley S, et al. Use of the Prone Position in Radiation Treatment for Women
With Early Stage Breast Cancer. International Journal of Radiation Oncology*Biology*Physics
1998;40:1137-1140.
• Liu B, Lerma FA, Patel S, et al. Dosimetric effects of the prone and supine positions on image
guided localized prostate cancer radiotherapy. Radiotherapy and Oncology, 2008; 88(1):67-76.
• Stroom JC, Koper PCM, Korevaaar GA, et al. Internal organ motion in prostate cancer patients
treated in prone and supine treatment position. Radiotherapy and Oncology 1999;51:237-248.
• Verhey LJ. Immobilizing and positioning patients for radiotherapy. Seminars in Radiation Oncology
1995;5:100-114.
• Dawson LA, Litzenberg DW, Brock KK, et al. A comparison of ventilatory prostate movement in four
treatment positions. International Journal of Radiation Oncology*Biology*Physics 2000;48:319-323.
• Zelefsky MJ, Happersett L, Leibel SA, et al. The effect of treatment positioning on normal tissue
dose in patients with prostate cancer treated with three-dimensional conformal radiotherapy.
International Journal of Radiation Oncology*Biology*Physics 1997;37:13-19.
Immobilization
Outline
• Fiorino C, Reni M, Bolognesi A, et al. Set-up error in supine-positioned patients immobilized with two
different modalities during conformal radiotherapy of prostate cancer. Radiotherapy and Oncology
1998;49:133-141.
• Halperin R, Roa W, Field M, et al. Setup reproducibility in radiation therapy for lung cancer: a
comparison between T-bar and expanded foam immobilization devices. International Journal of
Radiation Oncology*Biology*Physics 1999;43:211-216.
• Hanley J, Debois MM, Mah D, et al. Deep inspiration breath-hold technique for lung tumors: the
potential value of target immobilization and reduced lung density in dose escalation. International
Journal of Radiation Oncology*Biology*Physics 1999;45:603-611.
• Wong JW, Sharpe MB, Jaffray DA, et al. The use of active breathing control (ABC) to reduce margin
for breathing motion. International Journal of Radiation Oncology*Biology*Physics 1999;44:911-919.
I.
The Need for Margins
II.
Treatment Site Considerations
III.
Simulation Imaging Modalities
• Barnes EA, Murray BR, Robinson DM, et al. Dosimetric evaluation of lung tumor immobilization using
breath hold at deep inspiration. International Journal of Radiation Oncology*Biology*Physics
2001;50:1091-1098.
IV. Examples of Motion
• D'Amico AV, Manola J, Loffredo M, et al. A practical method to achieve prostate gland immobilization
and target verification for daily treatment. International Journal of Radiation Oncology*Biology*Physics
2001;51:1431-1436.
V.
• Gilbeau L, Octave-Prignot M, Loncol T, et al. Comparison of setup accuracy of three different
thermoplastic masks for the treatment of brain and head and neck tumors. Radiotherapy and Oncology
2001;58:155-162.
VI. In-room Guidance and Correction Strategies
• Negoro Y, Nagata Y, Aoki T, et al. The effectiveness of an immobilization device in conformal
radiotherapy for lung tumor: reduction of respiratory tumor movement and evaluation of the daily setup
accuracy. International Journal of Radiation Oncology*Biology*Physics 2001;50:889-898.
VII. Off-line Adaptive Strategies
• Bayley AJ, Giovinazzo J, Rakaric P, et al. Med-Tek type-S immobilization system: calculating the setup margin for radiotherapy of head and neck cancer patients. International Journal of Radiation
Oncology*Biology*Physics 2002;54:289-289.
VIII. Planning without margins
Treatment Planning
• Fuss M, Salter BJ, Cheek D, et al. Repositioning accuracy of a commercially available thermoplastic
mask system. Radiotherapy and Oncology 2004;71:339-345.
Department of Radiation Oncology • University of Michigan Health Systems
Simulation Imaging Modality
A. What you see
1. Computed Tomography
2. Magnetic Resonance
3. Positron-Emission Tomography
B. What you contour
K Langen
PROSTATE DELINEATION:
DIAGNOSTIC KV CT vs ABSOLUTE TRUTH
Radiotherapy and Oncology, 85(2), 239-246, 2007
Visible Human Project - Male
6 radiation oncologists, 120 delineations on KV CTs
Department of Radiation Oncology • University of Michigan Health Systems
PROSTATE DELINEATION:
DIAGNOSTIC KV CT vs ABSOLUTE TRUTH
Simulation Imaging Modality
Radiotherapy and Oncology, 85(2), 239-246, 2007
• 15 brain cases
Visible Human Project - Male
6 radiation oncologists, 120 delineations on KV CTs
• 1 neuro-radiologist
CT volumes on average 30% larger than true
volume
• 1 radiation oncologist
• Contours on CT, MR, registered CT-MR
CT volumes encompassed, on average, 84% of the
true volume.
Extended too anteriorly
Missed posteriorly
Ten Haken, Radiotherapy and Oncology, 25, 121-133, 1992.
Simulation Imaging Modality
GTV
Simulation Imaging Modality
CTV = GTV + 2 cm
12 patients
Inflammation used to
ID lumpectomy cavity
14% CT only
56% Both
Cavity volume ratio
(PET CT)/CT = 1.68
Range = 1.24 – 2.45
30% MR only
PTV volume ratio
(PET CT)/CT = 1.16
Range = 1.08 – 1.64
Adding margin reduces impact
“… the degree of inter-observer variation in volume definition is on the order of the
magnitude of differences in volume definition seen between the modalities, …”
“Given that tumor volumes independently apparent on CT and MRI have equal
validity, composite CT-MRI input should be considered for planning …”
PET CT largely
encompassed CT
Ford, IJROBP, 71, 595-602, 2008.
Ten Haken, Radiotherapy and Oncology, 25, 121-133, 1992.
Department of Radiation Oncology • University of Michigan Health Systems
Outline
III. Examples of Motion
I.
The Need for Margins
A. Head and Neck
II.
Treatment Site Considerations
III.
Simulation Imaging Modalities
B. Lung
IV. Examples of Motion
C. Liver
V.
D. Pancreas
Treatment Planning
VI. In-room Guidance and Correction Strategies
E. Prostate
VII. Off-line Adaptive Strategies
VIII. Planning without margins
Inter-Fraction Motion:
Head and Neck: Implanted Markers
Fraction
Intrafraction Motion and Deformation: 4D CT
Marker
Stability
Markers
Low, Wash U.
Zeidan, ASTRO 2007
Department of Radiation Oncology • University of Michigan Health Systems
Liver – Inter-Fraction Motion
•
•
•
Intrafraction Motion and Deformation: MRI
Implanted Markers (clips okay)
Localized using OBI (orthogonal X-Rays at Exhale)
3 mm threshold for adjustment
Balter, UM
Intrafraction Motion and Deformation: cineMRI
Outline
I.
The Need for Margins
II.
Treatment Site Considerations
III.
Simulation Imaging Modalities
IV. Examples of Motion
V.
Treatment Planning
VI. In-room Guidance and Correction Strategies
VII. Off-line Adaptive Strategies
VIII. Planning without margins
Courtesy Larry Kestin and Michel Ghilezan, WBH
Department of Radiation Oncology • University of Michigan Health Systems
IV. Treatment Planning
Margins
A. Margins
1. ICRU-29, 50, 62 – Definitions
2. Margin Recipes
B. Margin Compromises
ICRU 29 (1978)
2 years after whole-body CT scanners introduced
Target volume:
• contains those tissues that are to be irradiated to a
specified absorbed dose according to a specified timedose pattern.
• Includes expected movements, expected variation in
shape and size, variations in treatment setup
Treated volume:
• The volume enclosed by the isodose surface
representing the minimal target dose.
Irradiated volume:
• The volume that receives a dose considered significant
in relation to normal tissue tolerance.
Hot spot:
• Area that received a dose higher than 100% of the
specified target dose, at least 2cm in a section.
ICRU 50 (1993)
Gross Target Volume
Demonstrable extent of malignant growth
Clinical Target Volume
Microscopic extension, full dose to achieve aim
Planning Target Volume
CTV + uncertainty margins, “dose cloud”
Organs at Risk
Normal tissues that may influence plan
ICRU Reference Point
J Purdy, 2004
J Purdy, 2004
Department of Radiation Oncology • University of Michigan Health Systems
ICRU 62 Supplement (1999)
ICRU Definitions Evolve
ICRU 50
+
Setup Margin
Uncertainties in machine-patient positioning
Internal Margin
Uncertainties in patient-CTV positioning
Internal Target Volume
CTV + internal margin
Planning Organ at Risk Volume
Skip over ICRU definitions!
OAR + uncertainty margins
J Purdy, 2004
ICRU Definitions
J Purdy, 2004
ICRU Definitions
Irradiated volume:
Treated volume:
• Tissue volume that receives a dose
• Tissue volume that is planned to receive at
that is considered significant in relation
least a dose selected and specified by
to normal tissue tolerance.
radiation oncology team as being
appropriate to achieve the purpose of the
treatment.
Department of Radiation Oncology • University of Michigan Health Systems
ICRU Definitions
ICRU Definitions
GTV - Gross Tumor Volume
PTV - Planning Target Volume
• Complete actual or visible/demonstrable extent and
location of the malignant growth.
• Defined by specifying the margins that must be
added around the CTV to compensate for the
variations of organ, tumor and patient movements,
CTV - Clinical Target Volume
• A tissue volume that contains a GTV and/or subclinical
microscopic malignant disease, which has to be
eliminated.
• This volume has to be treated sufficiently in order to
achieve the aim of the therapy: cure or palliation.
inaccuracies in beam and patient setup, and any
other uncertainties.
• A static geometrical concept, can be considered a 3D
envelope in which the tumor and any microscopic
extensions reside and move (Dose cloud)
ICRU Definitions
ICRU Definitions
Organs at risk:
ICRU Reference point:
• Normal tissues (eg, spinal cord) whose
•
radiation sensitivity may significantly
influence treatment planning and/or
prescribed dose.
The system for reporting doses is based on the selection of a
point within the PTV, which is referred to as the ICRU Reference
point
– The dose at the point should be clinically relevant.
– The point should be easy to define in clear and unambiguous way.
– The point should be selected so that the dose can be accurately
determined.
– The point should in a region where there is no steep dose gradient.
Department of Radiation Oncology • University of Michigan Health Systems
ICRU Definitions
ICRU Definitions
Setup margin (SM):
Internal margin (IM):
• Uncertainties in patient-beam positioning in
reference to the treatment machine coordinate
system.
•
Variations in size, shape, and position of the CTV in reference
to the patient’s coordinate system using anatomic reference
points.
• Related to technical factors.
•
Caused by physiologic variations
• Can be reduced by
•
Difficult to control from a practical viewpoint.
•
The volume formed by the CTV and the IM called Internal
– Accurate setup and immobilization of the patient
– Improved mechanical stability of the machine.
ICRU Definitions
Planning organ at risk volume (PRV):
• A margin is added around the organ at risk to
compensate for that organ’s geometric uncertainties.
• Analogous to the PTV margin around the CTV.
target volume (ITV)
Margins Around Organs at Risk
Considerations
– Systematic errors
• Sensitive to shifts in a particular direction
– Random errors
• Impact of dose blurring
– Serial organs at risk
• Sensitive to hot spots
– Parallel organs at risk
• Some tolerance to limited hot spots
Department of Radiation Oncology • University of Michigan Health Systems
Challenges of Implementing ICRU 50 & 62
A. Delineating Volumes
Delineating the GTV
• Based mostly on clinical judgment
• Depends significantly on the imaging modality
B. New Technologies
(CT, MRI, PET, Registration)
C. Adding Errors
• CT is still the principal source of imaging data
used for defining the GTV for 3DCRT and IMRT
• Window and level settings are critical
J Purdy, 2004
Delineating the CTV
GTV to CTV Margins
• More of an art than a science because
current imaging techniques are not
capable of detecting subclinical tumor
• Giraud P, Antoine M, Larrouy A, et al. Evaluation of microscopic tumor extension in
non-small-cell lung cancer for three-dimensional conformal radiotherapy planning.
International Journal of Radiation Oncology*Biology*Physics 2000;48:1015-1024.
• Choo R, Woo T, Assaad D, et al. What is the microscopic tumor extent beyond clinically
delineated gross tumor boundary in nonmelanoma skin cancers? International Journal
of Radiation Oncology*Biology*Physics 2005;62:1096-1099.
• Apisarnthanarax S, Elliott DD, El-Naggar AK, et al. Determining optimal clinical target
volume margins in head-and-neck cancer based on microscopic extracapsular
extension of metastatic neck nodes. International Journal of Radiation
Oncology*Biology*Physics 2006;64:678-683.
• Chao KK, Goldstein NS, Yan D, et al. Clinicopathologic analysis of extracapsular
extension in prostate cancer: Should the clinical target volume be expanded
posterolaterally to account for microscopic extension? International Journal of Radiation
Oncology*Biology*Physics 2006;65:999-1007.
involvement directly.
• Gao X-s, Qiao X, Wu F, et al. Pathological analysis of clinical target volume margin for
radiotherapy in patients with esophageal and gastroesophageal junction carcinoma.
International Journal of Radiation Oncology*Biology*Physics 2007;67:389-396.
• Grills IS, Fitch DL, Goldstein NS, et al. Clinicopathologic Analysis of Microscopic
Extension in Lung Adenocarcinoma: Defining Clinical Target Volume for Radiotherapy.
International Journal of Radiation Oncology*Biology*Physics 2007;69:334-341.
J Purdy, 2004
Department of Radiation Oncology • University of Michigan Health Systems
Delineating the PTV
Compromises Between Volumes
The following should be taken into account:
• PTV and PRV may overlap
– Data from published literature
• Conflict in optimization criteria
– Any uncertainty studies performed in the clinic
• Judgment required in resolving volume conflicts
– the presence of nearby organs at risk.
– The asymmetrical nature of the GTV/CTV
geometric uncertainties.
J Purdy, 2004
J Purdy, 2004
Challenges with New Technologies
Challenges with New Technologies
IMRT
4D CT and Cone Beam CT
– Sharp beam gradients
– Protocols for defining volumes
– More conformal
– Variations in motion after simulation
– More sensitive to geometric uncertainties
– Changes in volumes during therapy
– Time-dependant delivery
– Motion interplay may lead to hot/cold spots
J Purdy, 2004
J Purdy, 2004
Department of Radiation Oncology • University of Michigan Health Systems
Adding systematic and random errors
ICRU 62 – Combining Errors
Systematic errors, Σ
– treatment preparation errors
– influence all fractions
Add systematic and random errors quadratically,
with equal weighting
Random errors, σ
– treatment execution errors
– influence each fraction individually
Combining Errors
SDTot = Σ 2 + σ 2
PTV Margin Recipe
mPTV = α Σ + γ σ’
α ~ 2 – 2.5
γ ~ 0.7
Σ - combined standard deviation of
systematic errors (Σ2 = Σ2m + Σ2s + Σ2d)
Systematic errors are much more important
in determining margins than random errors
σ’ - combined standard deviation of
random errors (σ’2 = σ2m + σ2s)
Department of Radiation Oncology • University of Michigan Health Systems
Margins Around Organs at Risk
Margins Around Organs at Risk
The DVH of the Planning Risk Volume should not
underestimate the contribution of the high-dose
components to the Organ at Risk, in 90% of the cases.
Outline
I.
Dan Low
takes
over!
The Need for Margins
II.
Treatment Site Considerations
III.
Simulation Imaging Modalities
IV. Examples of Motion
V.
Treatment Planning
VI. In-room Guidance and Correction Strategies
VII. Off-line Adaptive Strategies
VIII. Planning without margins
Department of Radiation Oncology • University of Michigan Health Systems
Target Volume Localization
Image Guidance: Positional Variations
Early days
A. Skin marks
B. Anatomical Landmarks and Surrogates
Bones
Implanted markers
Soft tissues
C.
D.
E.
F.
G.
H.
I.
OLD UM STUFF
MV Portal Imaging
kV Imaging
Ultrasound
CBCT
MVCT
Electromagnetic
Surface mapping
N=10
Weekly Images
4-8 films per
patient
Chamfer matching of bony anatomy on digitized port films
Balter, IJROBP, 31, 113, 1995
Balter, IJROBP, 31, 113, 1995
Department of Radiation Oncology • University of Michigan Health Systems
TRANSABDOMINAL ULTRASOUND
Transabdominal Ultrasound: Comparison with Implanted Markers
US (BAT) vs Markers (EPID)
Ant / Post
Sup / Inf
Lateral
Difference: Average ± SD (mm)
-0.7 ± 5.2
2.7 ± 4.5
1.8 ± 3.9
Langen et al., IJROBP, 2003
First widely used targeting technique
Started era of image guidance for routine clinical use
(1998)
Introduced DAILY imaging
US (Sonarray) vs Markers (Exactrac)
Frequency of misalignments
26%
0-5 mm
5-10 mm
48%
17%
10-15 mm
15-20 mm
5%
>20 mm
4%
Scarborough et al., IJROBP, 2006
Transabdominal US: Comparison with CT
Dong et al. (MDACC), ASTRO 2004: BAT vs In-Room CT
KV CT on Rails
PRIMATOM (SIEMENS)
Diagnostic CT
quality images
N=15 patients, 3 CTs per week.
N=342 CT/US pairs Average 23 scans/patient
Physician contours on each CT
Prostate center of volume
Full body field of
view
BAT vs CT differences (Mean + SD):
Lateral
0.5 + 3.6 mm
Ant / Post
0.7 + 4.5 mm
Sup / Inf
0.4 + 3.9 mm
Limitations:
Room size
“…the overall performance … seems unsatisfactory.”
Separate imaging
and treatment
isocenters
Dong et al., IJROBP, 60S, p. S332, 2004
Integration
Wong et al, IJROBP 61, 561–569, 2005
Department of Radiation Oncology • University of Michigan Health Systems
Cone Beam KV CT
Sample CBCT Images: Good for H&N
Elekta Synergy
Same Gantry
Coupled to delivery device
Volume acquisition
Varian Trilogy
Courtesy of Duke Univ Hosp, Durham
Courtesy Varian
Sample CBCT Images: Worse for Pelvis
Sample CBCT Images – Motion Artifacts
TP CT
CBCT - 1
Week 6
Week 2
Week 4
CT scan acquisition & reconstruction time ≈ 2 minutes
Elekta Synergy
Synergy
Images Courtesy of Floris Pos – Netherlands Cancer Institute
Department of Radiation Oncology • University of Michigan Health Systems
Sample CBCT – Breathing Motion Artifacts
Courtesy of Duke Univ Hosp, Durham
Courtesy Varian
Sample CBCT– Respiratory Synchronized CBCT
Courtesy of Duke Univ Hosp, Durham
Courtesy Varian
-152 cases, 7 anatomic sites
-Daily MVCT prior to each treatment; 3788 scans
-Offsets determined for each case
Liu et al, IJROBP, 2007
Li et al, IJROBP, 68; 581-91, 2007
Department of Radiation Oncology • University of Michigan Health Systems
15
LATERAL (mm)
LONGITUDINAL (mm)
25
0
0
-15
-25
Data to help PTV margin determination
6
VERTICAL (mm)
ROLL (mm)
20
0
0
Institutional variations in patient position, immobilization, localization
technique, inter- and intra-fraction motion measurements should also
be taken into account when developing margins.
-20
-6
Li et al, IJROBP, 68; 581-91, 2007
Li et al, IJROBP, 68; 581-91, 2007
Impact of real time motion on dosimetry:
Prostate (Calypso tracks)
LATERAL
Legend
Prostate D95
LONGITUDINAL
Plan
Mean
SD
Max
Min
2.3
2.2
D95 (Gy)
VERTICAL
Li et al: Dosimetric Consequences of Intrafraction Prostate
Motion. IJROBP, 71( 3),801-812, 2008.
Langen et al: Observations on real-time prostate gland motion
using electromagnetic tracking, IJROBP, 71, 1084-1090, 2008
2.1
2.0
Pierburg et al: Dosimetric Consequences of Intrafraction
Prostate Motion on Patients Enrolled on Multi-Institutional
Hypofractionation Study. IJROBP, 69, S23-S24. ASTRO 2007
1.9
1.8
1.7
1
2
3
4
5
6
7
8
9
10
Patient
Li et al, IJROBP, 68; 581-91, 2007
GEOGRAPHIC MISSES
Kupelian, IJROBP, 66, 876-82, 2006
VERSUS
DOSIMETRIC MISSES
Rajendran et al: Daily Isocenter correction with
electromagnetic based localization improves target coverage
and rectal sparing during prostate radiotherapy, IJROBP, In
Press, 2009
Department of Radiation Oncology • University of Michigan Health Systems
WHY NOT IMAGE EVERYDAY?
IMAGING DOSE: MVCT vs KV CBCT
MVCT:
1 - 2 cGy per image*
Relatively uniform dose
40 - 80 cGy per entire course
Can limit length of scan
* Meeks et al., Med Phys. 2005;32(8):2673-81
* Shah et al., IJROBP, 2007; 69(3), S193-194.
In-Room Guidance:
Inter-Fraction
Motion Management Example
ASTRO 07; ABSTRACT 1100
KV CBCT:
1 - 5 cGy per image**, depending on site
Heterogeneous throughout scan volume
40 - 200 cGy per entire course
Cannot limit length of scan
Implanted Markers
** Amer et al., Br J Radiol. 2007;80(954):476-82.
** Wen et al.,Phys Med Biol. 2007;52(8):2267-76.
** Islam et al., Med Phys. 2006;33(6):1573-82.
Advantages of Markers
•
•
•
•
•
•
•
Accuracy
Measured in treatment position
Semi- to fully automated
User independent
No contouring
Real-time position possible
Beam gating
Marker Visualization & Extraction
• 3 – 4 markers
• Centroid of visible markers
σCM < 1 mm
Pouliot, IJROBP, 2003
• 99.6% visualization of 3 markers
18 MV (1x3 mm)
Herman, IJROBP, 2003
• Automated marker extraction
93% per marker, 85% for 3 markers
Aubin, MedPhys, 2003
90% - 100% based on ROI (MVCT)
Chen, AAPM, 2005
Department of Radiation Oncology • University of Michigan Health Systems
Marker Migration & Stability
• 1.3 (0.44 – 3.04) mm (Pouliot, IJROBP, 2003)
• 1.2 +/- 0.2 mm (Kitamura, Radiother Oncol, 2002)
• 0.7 – 1.7 mm (Litzenberg, IJROBP, 2002)
• 1.01 +/- 1.03 mm (Kupelian, IJROBP, 2005)
• 0.8 mm with Beacons (Willoughby, IJROBP, 2006)
Frequency & Time for Adjustment
• 5 mm action threshold with skin mark setup
53% of fractions realigned pre-treatment
1.5 minute mean increase in treatment time
Herman, IJROBP, 2003
• 2 mm action threshold
74% of fractions realigned
< 5 minute increase in treatment time
Area of marker triangle proportional to CT
volume of prostate (Moseley, AAPM, 2005)
Need for implanted fiducials
with CT scans
kV cone beam CT
MV cone beam CT
Helical MV CT
Beaulieu, Radiother & Oncol, 2004
Prostate alignment study
MVCT scans
Radiation Therapist vs Physician
Alignment methods:
Marker
Anatomy
Contours
3 patients, 112 scans
Sagittal
MV CT
Sagittal
CB CT
Langen et al., IJROBP, 62, 1517, 2005
Department of Radiation Oncology • University of Michigan Health Systems
Cone Beam KV CT Alignment techniques
Fiducials vs Soft Tissues
Therapist vs Physician registration
Moseley et al., IJROBP, 67(3), 942–953, 2007
Difference
5 mm
Princess Margaret Hospital: 15 patients, 256 CT sets
Marker
Anatomy
Fiducial vs Soft tissues on cone-beam CT
Contour
PTV margins: 10 mm except 7 mm posteriorly
A/P: 1 %
S/I: 2 %
Lat: 1 %
A/P: 5 %
S/I: 10 %
Lat: 0 %
A/P: 17 %
S/I: 31 %
Lat: 3 %
Langen et al., IJROBP, 62, 1517, 2005
Cone Beam KV CT Alignment techniques
Fiducials vs Soft Tissues
kV Cone Beam CT: Soft Tissues vs Markers
Agreement within 3 mm
Inter-observer Variability (mm)
Lateral (LR)
Vertical (AP)
Long (SI)
Markers
Systematic Error (SD)
Random Error
0.03
0.26
0.15
0.95
0.01
0.47
Soft tissues
on CBCT
Systematic Error (SD)
Random Error
0.61
1.50
1.61
2.86
2.21
2.85
LR
AP
SI
91%
64%
64%
Pearson’s
correlations R2
LR 0.90
AP 0.55
SI
0.41
Less inter-observer variability with markers
Moseley et al., IJROBP, 67(3), 942–953, 2007
Moseley et al., IJROBP, 67(3), 942–953, 2007
Department of Radiation Oncology • University of Michigan Health Systems
Real-Time Marker Tracking
In-Room Guidance:
MOTION
INTRA FRACTION
IRIS
Dual Fluoroscopy
and Flat Panels
Courtesy George Chen, Harvard and Mass General Hospital
Real-Time Marker Tracking
Real-Time Marker Tracking
Department of Radiation Oncology • University of Michigan Health Systems
Marker Implantation Feasibility
AC Wireless Magnetic Tracking
Measurements at 10 Hz
Sub-mm resolution
Real-time tracking
Potential for fast correction
Calypso System
Real Time Motion Data – Patient 1
Real Time Motion Data – Patient 2
IS
AP
LR
Position (mm)
Position (mm)
IS
AP
LR
Department of Radiation Oncology • University of Michigan Health Systems
Percentile Analysis
Percentile: Regular Breather
v85
v85 v90
v90
v95
v95
Regular Breather
v98
Irregular Breather
v98
vx = Volume at
which patient
had v or less
volume x% of
the time
v5
v5
Example V98 (93%
) vs V85 (80%
of time
Amount of Motion We Want to Know
)
of time
In-Room Correction Strategies
Available 3D Image Datasets
A. Action Levels
B. Gating
Ratio 1.38 ± 0.19
Images that cover 80% of breathing cycle show only 72% of the
motion at 93% of the breathing cycle!
Department of Radiation Oncology • University of Michigan Health Systems
Patient Specific Margin
Variation vs Action Level
Σ=0
1.0
0
0
PTV margin = Setup + Inter + Intra-fraction
+ Measurement
+ Correction
0.8
3.0
2.5
0.6
2.0
0.4
1.5
(σmeas2 + σpos2)1/2
1.0
Frequency
3.5
Adjustment Frequency
Corrected Uncertainty, σ (mm)
4.0
0.2
0.5
0.0
Lam, AAPM, 2006
2
4
6 8
0.1
0.0
2
4
6 8
2
4
6 8
1
10
Action Level (mm)
100
Position (cm)
Intra-Fraction Correction
Initial Setup Variations
1
0.1
10
1.0
8
0.8
-1
-0.1
AP
Systematic
σIntra
-2
-0.2
Assume σs = 1 mm
-3
-0.3
Position
(mm)
Position
(cm)
Position
Position(mm)
(cm)
0
0.0
6
0.6
4
0.4
2
0.2
0
0.0
-2
-0.2
-4
-0.4
-4
-0.4
-5
-0.5
0
60
120
180
240
300 360
Time (s)
420
480
540
600
0
60
120
180
240
300 360
Time (s)
420
480
540
600
Department of Radiation Oncology • University of Michigan Health Systems
Margins vs Management Strategy
Radiofrequency transponders: Prostate rotation
• Apply real rotations from tracked data to assess
plan efficacy
• Real-tracking data (sampled at 10Hz) from a
research version of the Calypso System
• Rotational and translational corrections were
applied to the prostate contours for each patient
and the amount of time the prostate was outside
the PTV was determined
C. Noel, Washington University
Plan Evaluation
Plan Evaluation
For 3 patients, over 122 total fractions, the percentage of time
any portion of the prostate is outside of a 5mm PTV: ***
…for a 3mm PTV margin:
% Time out of 5mm PTV
% Time out of 3mm PTV
Patient 1
25%
Patient 1
Patient 2
<1%
Patient 2
5%
Patient 3
51%
Patient 3
93%
***C Noel, ASTRO 2008
80%
***C Noel, ASTRO 2008
Department of Radiation Oncology • University of Michigan Health Systems
Plan Evaluation
Visualize Rotations
Contoured prostate and PTV from plan
…for a 8mm PTV margin:
% Time out of 8mm PTV
Patient 1
<1%
Patient 2
0%
Patient 3
3.3%
***C Noel, ASTRO 2008
Day
Day21Prostate
ProstatePosition
PositionatatSet-up:
Set-up:30˚
9˚
Visualize Rotations
Tracked motion data
Visualize Rotations
Change PTV from 5mm to 3mm
Department of Radiation Oncology • University of Michigan Health Systems
Outline
Offline Adaptive Protocols
• Adaptive Radiation Therapy – ART
I.
The Need for Margins
II.
Treatment Site Considerations
III.
Simulation Imaging Modalities
– N ~ 5 CT scans with contours
– Union of volumes to form patient-specific PTV
Yan, et al, IJROBP, 2000.
• No Action Level protocol – NAL
IV. Examples of Motion
V.
– N ~ 5 portal images to estimate Σ
Treatment Planning
De Boer, et al, IJROBP, 2001
VI. In-room Guidance and Correction Strategies
• Shrinking Action Level protocol – SAL
– Nmax ~ 5 portal images to estimate ΣN
– If ΣN < αN = α / N1/2, apply offset to all fractions
– Else apply and restart at N = 1
VII. Off-line Adaptive Strategies
VIII. Planning without margins
Bel, et al, Radiother Oncol, 1993.
Confidence-Limited PTV (cl-PTV)
transverse
sagittal
Adaptive Radiation Therapy - ART
coronal
Initial CTV + 4 CTVs = ITV
ITV + Random Setup Error of Bones = cl-PTV
Kestin, RTOG, 2006
Department of Radiation Oncology • University of Michigan Health Systems
Adaptive RT Trial
Conventional
PTV
cl-PTV
Margin Reduction (N=600)
• Margins stringently designed to allow <3%
dose reduction within CTV with an 80%
confidence
• PTV volume reduced by mean of 34% (N=600 patients)
– > 30% of conventional PTV unnecessary in 75% of patients
• Equivalent uniform margin: mean = 5.6 mm
• If planning based on single CT scan:
– ~ 14 mm margin (AP) required to meet similar dosimetric
criteria
• Average IMRT dose escalation of 7.5% (86.7 Gy)
Kestin, RTOG, 2006
Outline
I.
The Need for Margins
II.
Treatment Site Considerations
III.
Simulation Imaging Modalities
IV. Examples of Motion
V.
Treatment Planning
VI. In-room Guidance and Correction Strategies
VII. Off-line Adaptive Strategies
VIII. Planning without margins
Meldolesi et al. Int J Radiat Oncol Biol Phys 63:S9063:S90-S91; 2005
Treatment without Margins
MIGA
• Multiple Instance Geometry Approximation (MIGA)
• Approximate positional uncertainties by a modeled
sequence of positions (e.g. independent CT scans)
• Dose calculation is for weighted sum of different patient
geometries
• Assuming that model is correct, dose calculation
accounts for geometric uncertainty
Department of Radiation Oncology • University of Michigan Health Systems
MIGA
The End!
Thanks for attending!
Up Next:
•
•
•
•
Fluence is NOT just a blurring of the PTV-based plan!
No PTV margin
Simultaneous optimization of multiple instances of geometry
Solutions maybe worse than the static geometry solution,
but are much better than the static geometry solution when
compromised by motion
Treatment Planning of Complex Cases
M Hunt and T Nurushev
Department of Radiation Oncology • University of Michigan Health Systems