imrt: treatment planning and dosimetry nesrin dogan, ph.d department of radiation oncology virginia...
Post on 21-Dec-2015
217 views
TRANSCRIPT
IMRT: Treatment Planning and Dosimetry
Nesrin Dogan, Ph.DDepartment of Radiation OncologyVirginia Commonwealth University
Medical College of Virginia HospitalsRichmond, VA, USA
N. Dogan / Nov 2007
Fundamental Issues
• Beam Modeling
• Dose Calculation
• Inverse Planning
• IMRT QA
N. Dogan / Nov 2007
Beam Modeling• For small fields, minor uncertainties due to approximations
in dose calculation models, methods for determining MLC leaf sequences and other factors may form a large fraction of dose delivered, and lead to inaccuracies in delivered dose.
10 cm1 cm1 cm1 cm1 cm1 cm1 cm1 cm1 cm1 cm1 cm
N. Dogan / Nov 2007
Beam Modeling, cont.
• Dosimetric accuracy of the IMRT plan delivery depends on the accurate representation of Accurate Beam Penumbra representation – MLC /
collimator jaws. Adequate characterization and accounting of
transmission and scattering properties of MLC leaves. Output factor for small field size. Accuracy of dose calculation algorithm. Approximations of leaf sequence generation algorithm. Leaf positioning accuracy.
N. Dogan / Nov 2007
Penumbra• Need to be measured with microchamber, film or
diode.• Subtle effects make a difference in IMRT.
Beam model based on penumbra measured with 6 mm diameter chamber
Beam model based on penumbra measured with film
Courtesy of G. Ezzel, Ph.D., Mayo Clinic
N. Dogan / Nov 2007
MLC Leaf Characteristics
• Inter- and intra-leaf transmission
• Tongue-and-groove – can lead to under-dosages ~30% in a 2 mm wide region
• Rounded tip
~12% ~1% ~1.5% ~2.5%
N. Dogan / Nov 2007
Radiation Field Offset for Rounded Leaf Ends
• Offset for between the light and radiation field edge = ~0.6 mm
Measuring the offset
0.6 is best, i.e. subtract 0.6 mm from MLC settings -> TPS should take care of this
Courtesy of G. Ezzel, Ph.D., Mayo Clinic
N. Dogan / Nov 2007
MLC Leaf Transmission and Scattering
• Leaf leakagetransmissionrounded leaf tip
transmissionMLC scatter
• Collimator scatter upstream from the MLC.
Leakage through leaf ~2%
Leakage between neighboring leaf ~5%
• Leakage through closed opposing leaf pair for rounded ends ~20% - leaves should be parked under the jaw
•Minimum gap between opposed leaves = 0.5 -0.6 mm to avoid collisions
Courtesy of G. Ezzel, Ph.D., Mayo Clinic
N. Dogan / Nov 2007
Output Factor Small Fields
D.A. Low et al. “Ionization chamber volume averaging effects in dynamic intensity modulated radiation therapy beams, Med.Phys.30(7): 1706-1711 (2003).
Micro cham: 0.009cc
PTW: 0.125cc
Farmer:0.65cc
N. Dogan / Nov 2007
Courtesy of G. Ezzel, Ph.D., Mayo Clinic
MLC and Small Fields
• Output for small fields very dependent on MLC accuracy.
• 10%/mm for 1 cm segment.
N. Dogan / Nov 2007
Which Detectors to Use?• Need to determine energy dependence
and angular response.• Small field detectors required for small
field characterization. Sensitive to position Detector should be smaller than
homogeneous region of dose to be measured
• Assess electrometer response.
N. Dogan / Nov 2007
DetectorVolume
(cm3)Diameter
(cm) Disadvantages
Micro-chamber
0.009 0.6 Poorer resolution than diodes
Pinpoint chamber
0.015 0.2Over-respond to low energy
photonsMartens et al. 2000p-type Si
diode0.3 0.4
Stereotactic diode
NA 0.45
MOSFET NA NA Non-linear dose response for <30 cGy
Diamond 0.0019 0.73 < resolution than diodes, dose rate dependence, expensive
Courtesy of Jean Moran, Ph.D, UofMichigan
Small 1-D Detectors
N. Dogan / Nov 2007
Pasma Med Phys 26: 2373-2378 (2376) 1999
PredictedEPIDIon Chamber+
Discrepancies in the penumbra region (up to 10%)
Overall: Good agreement
10 MV 25 MV
EPID: DMLC measurements
Courtesy of Jean Moran, UofMichigan
N. Dogan / Nov 2007
Dose Calculation• Current IMRT systems use simplified
dose calculations during plan optimization: e.g., pencil beam -> uses very simple heterogeneity corrections, causing significant dose errors (10% or more non-IMRT cases)
• Final dose calculation is performed using a separate independent dose calculation that incorporate the influence of the MLC: e.g, convolution / superposition; more accurate than Pencil beam; however, inaccuracies persist under certain circumstances
N. Dogan / Nov 2007
Conventional dose algorithms can be inaccurate for • Small fields• Regions of dose gradients (radiation
disequilibrium)• Heterogeneous conditionsIMRT is typically delivered through a sequence of small static fields or with a dynamically moving aperture with a small width. Dose gradients are common place in IMRT fields. For such fields, assumptions used in conventional algorithms regarding scatter equilibrium and output factor variation with field size typically break down.
N. Dogan / Nov 2007
• Significant fraction of the dose within targets and organs at risk is due to scattered or leakage radiation calculated dose distributions have the greatest
uncertainties due to approximations inherent in conventional methods of transforming intensities into MLC leaf sequences
• Experimental checks of IMRT fields routinely shows discrepancies between the planned (desired) and actual.
For IMRT
N. Dogan / Nov 2007
Dose Calculation AlgorithmsC
alc
ula
tio
n S
pe
ed
Calculation Accuracy
Pencil Beam
Monte Carlo
Superposition/Convolution
Courtesy: Jeff Siebers, VCU
N. Dogan / Nov 2007
Comparison of SC and MC
Comparison of a) Superposition-Convolution (SC) and b) MC dose calculations
N. Dogan / Nov 2007
Target
RT Lung
0 5 10 15 20
Dose (Gy)
0
20
40
60
80
100
Vol
um
e (%
)
Cord
Monte Carlo
Pencil Beam
Pawlicki et al., Med Dosim, 26 157 (2001)
Comparison of PB and MC
Pencil BeamMonte Carlo
N. Dogan / Nov 2007
Superposition
Monte Carlo
Slice 45
Monte Carlo
Slice 55
Monte Carlo
Slice 64
Comparison of SC and MCSuperposition Superposition
N. Dogan / Nov 2007
Consequences of Inaccuracy
• Dose Prediction Error (DPE) For a given intensity distribution, dose predicted differs
from that actually delivered to the patient/phantom Can be avoided by performing final calculation with
accurate algorithm
• Optimization Convergence Error (OCE) Consequence of systematic error during optimization Optimization with an inaccurate algorithm results in
different intensities than those predicted by an accurate algorithm
Actual dose is not optimal, a better solution exists Can be avoided by optimization with an accurate
algorithm
N. Dogan / Nov 2007
DPE(same intensities)
PB computed SC computed
Make sure your final dose calculation is with an accurate algorithm
68 Gy 64 Gy 60 Gy 50 Gy 40 Gy 30 Gy
N. Dogan / Nov 2007
OCE(different intensities)
SC optimizationSC calc
PB optimization SC calc
68 Gy 64 Gy 60 Gy 50 Gy 40 Gy 30 Gy
N. Dogan / Nov 2007
Conventional IMRT Optimization Process
Create Leaf Sequence
“Deliverable” DoseCalculation
Create Deliverable Intensities
Optimization
Leaf Sequencer
Leaf positions do not exist
Deliverable Plan
N. Dogan / Nov 2007
Problems with Conventional IMRT process
• Optimized plans are converted to deliverable plans through leaf-sequencing process that takes into account the limitations and effects (leakage/scatter) of the MLC
• The idealized optimal plan is replaced with “deliverable” plan
• Optimized and deliverable IMRT plans differ Different intensity distributions More complex the intensity distribution, the
greater the deviation
N. Dogan / Nov 2007
Comparison of Isodoses
a) An optimized intensity distribution b) A deliverable distribution using DMLC calculated using Convolution/Superposition algorithm
N. Dogan / Nov 2007
InitialIntensity (II(x,y))
EvaluatePlan Objective
Converged?
Ad
just
I(x,
y)
ComputeDose (DO)
Optimized Intensity (IO(x,y)) and Dose DO = DD
NoYes
1
3
4
5
2
Create Leaf Sequence 7
Create Deliverable Intensities(ID(x,y)) 8
6
Final dose is deliverable
Deliverable IMRT Optimization Process
combine optimization and
delivery into one process
Leaf Sequencing
N. Dogan / Nov 2007
Deliverable Optimization
Deliverable optimization can restore
original optimized plan
Original SCopt
Deliverable Plan SC
MC of Deliverable
MCopt
(deliverable)
Head and Neck IMRT plan
N. Dogan / Nov 2007
Heterogeneity Corrections• More important for IMRT than
conventional treatments.• Heterogeneities may effect some
beamlets more than others -> causing different localized dose differences.
• The reliability of clinical experience with DVH prescriptions and results may be significantly compromised if heterogeneity corrections are not used (e.g., Lung).
• Use AAPM Report No:85 Tissue Inhomogeneity Corrections for Megavoltage Photon Beams.
• 4% - 10% error in relative e- density result in ~2% error in dose.
w /Hetero
w/o Hetero
N. Dogan / Nov 2007
• Size of the OARs.• Dose gradients near the OARs.• Finer dose grids are necessary for cases in which high gradients are needed.• Dose grid should be finer than the size of the beamlets or incident fluence map so that the effects of modulation are adequately sampled.
Dose Grid
N. Dogan / Nov 2007
What do we do about differences?
• May need to adjust the beam model.• May need to live with it.
Take known deficiencies into account when evaluating plans
• May be important for OARs.
N. Dogan / Nov 2007
Buildup Region• Important when target regions (PTV)
extend into the buildup region.• Calculated doses are often
inaccurate and lower than delivered doses.
• Likely to cause hot spots in the target and elsewhere as a result of inverse planning engine fighting with the buildup effect – may cause excessive skin reactions and compromise the plan quality.
• Bolus needs to be added if the target is in the buildup region: needs to be included during the scanning of patient.
N. Dogan / Nov 2007
• Definition of Target Volumes GTV, CTV and PTV need to be explicitly definedConsistent with the ICRU definitions (ICRU 50)
• Definition of OARsPlanning OAR Volume (ICRU 62)
• Need to • Use contrast-enhanced CTs• Image fusion (PET, MRI, preopt CTs, etc..)
Target and OARs
N. Dogan / Nov 2007
Margins for Targets• IMRT does not inherently demand for
tight target margins.• CTV to PTV margins depends on each
individual patient and the patient immobilization / location techniques used.
• Tight target margins can be achieved by improved imaging for planning, immobilization and image-guided verification.
N. Dogan / Nov 2007
Realistic for CTV?
Courtesy of G. Ezzel, Ph.D., Mayo Clinic
• Automatic CTV expansions may unrealistically cross tissue boundaries.
Automatic CTV expansions
N. Dogan / Nov 2007
Margins for OARs
• ICRU 62 recommendations suggest the use of margins for OARs.
• Generate expanded OARs if it is possible. e.g.;CordExpand = Cord + 5 mm
BrainstemExpand = Brainstem + 5 mm
• Create “pseudo” structures to achieve sparing at the desired areas.
N. Dogan / Nov 2007
Oral mucosa - avoid
Courtesy of G. Ezzel, Ph.D., Mayo Clinic
Defining Normal Tissues• Tissues to be spared need to be explicitly defined; e.g., oral mucosa when changing from parallel-opposed to IMRT.
N. Dogan / Nov 2007
Target
Nodes
Spinal cord
Avoidance tissue
Avoidance tissue
Gy
60
50
45
30
Gy
60
50
45
30
Courtesy of G. Ezzel, Ph.D., Mayo Clinic
N. Dogan / Nov 2007
Hot Spots Outside of Target Regions
• Occurs in regions that are not contoured.
Work-around• Create “Unspecified Tissue”
Region and include in the optimization.
N. Dogan / Nov 2007
Defining OARs for Optimization
• Create nonPTVOARs for organs overlapping with PTVs:
NonPTVSmallBowell, NonPTVRectum,
NonPTVBladder, etc.
N. Dogan / Nov 2007
Guidelines for Target Expansions Prostate CTV: Expand prostate by 0.5cm in all directions except posteriorly then + seminal vesicles (no expansion for seminal vesicles) Prostate PTV: Expand Prostate CTV by 0.5cm in all directions (3D expansion) Lymph Nodes CTV: Expand lymph nodes by 1.0 cm in anterior, posterior, right and left (2D expansion) with small bowel, bladder, rectum, bones, muscle,
skin1cm and prostate PTV tissues being the limiting organs Lymph Nodes PTV: Expand Lymph Nodes CTV 0.5 cm in all directions (3D expansion) with only skin1cm and Prostate PTV as the limiting structures
N. Dogan / Nov 2007
• Required by the inverse planning process – dose or dose-volume constraints for all structures
• A trial and error process to come up with the proper dose or dose-volume constraints.
• Don’t ask the impossible – set realistic goals – improperly specified constraints will result in inferior plans.
• Create site-specific protocols which can be used for similar cases.
Constraints
N. Dogan / Nov 2007
H&N IMRT Treatment Planning Instruction FormDepartment of Radiation Oncology, VCU Health SystemsStructures Limiting Dose(Gy) Volume (%) / cc Fraction
size
PTV1 7077
97< 20
35
PTV2 56 95 35
PTV3 ----- -----
Brainstem + 0.5 cm 50 0
Cord + 0.5 cm 48 ≤ 0.03 cc
Mandible 60 30
Oral Cavity 45 50
Parotids (L & R) – at least one of them
30 ≤ 50
Larynx – if feasible 45 0
Esophagus 45 30
Brachial Plexus 60 0
Unspecified Tissue 70 ≤ 5
CTV1= GTVt + GTVn + 1cm margin for subclinical disease
PTV1 = GTVt + GTVn + 0.5 setup uncertainty
PTV2 = CTV1 + 0.5 setup uncertainty
PTV3 = CTVnodes + 0.5 setup uncertainty
N. Dogan / Nov 2007
Prostate IMRT Treatment Planning Instruction FormDepartment of Radiation Oncology, VCU Health SystemsStructures Limiting Dose(Gy) Volume (%) Fraction Size
PTV 63 97 28
PTVNodes50.4 95 28
Femurs (L&R) 354045
50 10 2
Periprostatic rectum 556365
5010 2
Remaining rectum 455560
5010 2
Bladder 456065
5010 2
Small Bowel 254550
5010 2
Skin1cm 303545
5010 2
Unspecified Tissue 50 10
N. Dogan / Nov 2007
General Principles for Beam Angle Selection
N. Dogan / Nov 2007
Beam ConfigurationsGeneral Principles
• It is useful to minimize the number of beams for practical reasons
• The minimum number of beams depends upon a complex combination of factors: Shape and size of target volume Locations, tolerances and tissue architecture of normal
tissues Prescription dose (higher doses would normally
require more beams)
• The optimum number may be determined for each class of radiotherapy problems by trial-and-error
N. Dogan / Nov 2007
Beam ConfigurationsGeneral Principles
• If sufficient number of beams are used, the IMRT plan quality is relatively insensitive to beam angles The computer should be able to adjust the weights of
rays to make up for modest imperfections in beam placement
Beams may be placed at equiangular steps
• The larger the number of beams, the better the IMRT plan Rotational IMRT should be better
• Non-coplanar beams should provide additional benefit
N. Dogan / Nov 2007
Beam ConfigurationsGeneral Principles
• In general, if beam angles are optimized the plan optimality should improve the number of beams required for equivalent
dose distribution is smaller than if beams are placed at equi-angular steps
• Computer-aided optimization of the beam angles is a difficult and as yet inadequately solved problem Extremely large number of plans need to be
compared
N. Dogan / Nov 2007
• Choose shortest path to irradiate target(s)• Avoid OARs• Keep large beam separation if it is
possible• Beam angle may become important for
tumors that are not centrally located.• It depends on the optimizer
Beam ConfigurationsGeneral Principles
N. Dogan / Nov 2007
H&N:5 Beam
9 Beam
7 Beam
15 Beam
N. Dogan / Nov 2007
Isocenter Placement
• Better plans can be achieved by selective isocenter placement.
• Desirable to shift isocenter to provide best separation between target and tissues.
• Desirable to have isocenter in region of reliable bony anatomy.
Center of All Targets
Center of PTV
N. Dogan / Nov 2007
7 rows to cover target
One row hits target and structure
6 rows to cover target
Split between target and structure
Spatial quantization effects• Shift isocenter to provide best
separation between target and tissues.
Courtesy of G. Ezzel, Ph.D., Mayo Clinic
N. Dogan / Nov 2007
Dose-Volume Based vs. EUD-Based Optimization – H&N
ExampleEUD + tumor as
“virtual normal tissue”
3045
50(c)
70
60
EUD unconstrained
(b)3045
50
80
60
70
45
Dose-Volume
(a)
30
5060
70
45
N. Dogan / Nov 2007
N. Dogan / Nov 2007
60
50
70
40
(a)
80
40
80
6070
50
(c)
9080
4090
6050
70
(b)
Bladder
Target
Bladder
Target
BladderRectum
Dose-Volume EUD unconstrainedEUD + tumor as
“virtual normal tissue”
Dose-Volume Based vs. EUD-Based Optimization – Prostate
Example
N. Dogan / Nov 2007
N. Dogan / Nov 2007
SIB IMRTTwo-Phase IMRT
70
5045
60
3540
45
50
35
40
45
45
50
60
70
50Mean
dose to nodes 59 Gy
Mean dose to nodes 51 Gy
Sequential vs. SIB
N. Dogan / Nov 2007
Non-Target Tissue Volumes Receiving Specified Dose
Dose Level (cGy)
Two-phase IMRT
Simultaneous integrated boost
IMRT
1000 2183 2169 0.6
2000 1975 1941 1.8
3000 1557 1459 6.7
4000 1096 1016 7.9
5000 732 604 21.2
6000 388 238 63.0
7000 83 62 33.9
% Difference between SIB IMRT and 2-phase IMRT
Volume (cc) outside the target (tumor and nodes) at specified
dose level or higher
N. Dogan / Nov 2007
Minimize Number of Segments
N. Dogan / Nov 2007
Minimize Number of Segments
50 Segments
75 Segments
100 Segments
150 Segments
200 Segments
Segments MU
50 550
75 582
100 604
150 619
200 631
N. Dogan / Nov 2007
Impact of Degree of Fluctuations (“Complexity”) in Intensity Patterns
on MUs for IMRT
100
Total MUs= 100
100
Total MUs = 300
100 100
10 cm
10 cm
100 100
Total MUs = 200
10 cm
Total MUs = 150
50
50 50
10 cm
N. Dogan / Nov 2007
Attention to Objective Function
N. Dogan / Nov 2007
Target Volumes
Critical Structure
TLDs in Target VolumesRadiochromic film through multiple plansDelivery is required by RTOG for participation in IMRT trials
Removable DryInsert
Water
Water
Anthropomorphic: RPC Head Phantom
Courtesy of Jean Moran, UofM
N. Dogan / Nov 2007
Anthropomorphic: RPC Head Phantom
N. Dogan / Nov 2007
Examples
N. Dogan / Nov 2007
HN IMRT w / Supraclavicular NodesTreat Nodes with
AP ScV field• Requires matching of
IMRT fields w/AP field
• May cause hot or cold spots
• ScV field needs to be included in the IMRT optimization
• Feathering
• Watch out for overlaps if the IMRT plan wants to open the jaws into the ScV area -> may need to adjust the jaw
ScV field
IMRT field
N. Dogan / Nov 2007
This row might be used by the IMRT plan if the target is drawn too close to the isocenter plane
Courtesy of G. Ezzel, Ph.D., Mayo Clinic
Human planners sometimes have to trim IMRT beams…..
N. Dogan / Nov 2007
HN IMRT w / Supraclavicular Nodes
Treat Nodes with
IMRT• No matching –
eliminates junction issues.
• Needs extra care treating shoulders – avoid hot spots.
• Immobilization.
N. Dogan / Nov 2007
Nasopharynx
Nasopharynx field
arrangement
7 posterior fields
N. Dogan / Nov 2007
SIB IMRT
PTV1: 54 Gy/30 fx’s
(1.8Gy/fx)
PTV2: 60 Gy/30 fx’s
(2.0Gy/fx)
PTV3: 67.5 Gy/30 fx’s
(2.25Gy/fx)
• The field length on the 2 lateral fields is stopped at the top of the shoulder.
• The s/clav portion of the target volume is treated with the remaining 5 fields.
Nasopharynx
N. Dogan / Nov 2007
Sinus
• Single Tx Volume• 60 Gy/30fx’s• Challenging due to
the proximity and sometimes overlap of tumor volume and critical optical structures and large air cavity.
N. Dogan / Nov 2007
7 beam field arrangement
N. Dogan / Nov 2007
• A tunnel was carved out in the dose cloud around the optic structures.
• Achieved with the use of expanded structures and the ability to manipulate both the imp. Weighting and overlap priorities.
N. Dogan / Nov 2007
Sinus DVH
N. Dogan / Nov 2007
Nasal Cavity and
lymph nodes
• 2 Tx volumes• PTV1 46 Gy/ 23 fx’s
nasal cavity & nodes• PTV2 24 Gy/12 fx’s
nasal cavity boost• Sequential IMRT
plans• 0.5 cm Bolus over
nose
N. Dogan / Nov 2007
PTV1• 7 beams• Gantry angles
205, 285, 315, 0, 20, 75, and 153
N. Dogan / Nov 2007
PTV2
• Sequential IMRT plan
• 4 beams• Gantry angles
285, 315, 20, and
75
N. Dogan / Nov 2007
Nasal cavity isodose lines
N. Dogan / Nov 2007
Prostate and Lymph Nodes
1. Prostate PTV+ Lymph Nodes PTV: 50.4 Gy / 28 fx
Prostate PTV : 63 Gy (BED = 70 Gy) / 28 fx
2. Sequential 9 Gy IMRT Boost to Prostate PTV : 72 Gy (BED = 80 Gy)
or
Upfront 6 Gy HDR boost
SIB-IMRT
N. Dogan / Nov 2007
Heart
IM nodes
Heart
IM nodes
Breast IMRT w/IM nodes
N. Dogan / Nov 2007
Breast IMRT W/regional Nodes
N. Dogan / Nov 2007
Breast IMRT W/regional Nodes
N. Dogan / Nov 2007
Lung IMRT
N. Dogan / Nov 2007
Boost1
Primary Tumor
Bladder
RectumCTV
Brachytherapy + IMRT of Cervix
• Stage IIIB cervix cancer
• Patient receives both brachytherapy (BRT) and external radiotherapy (XRT)
Boost2(Involved Node)
N. Dogan / Nov 2007
GYN Brachytherapy + IMRTIMRT+BRTConventional 3DCRT+BRT
Primary Tumor
N. Dogan / Nov 2007
0
20
40
60
80
100
0 20 40 60 80
Rectum
XRT + BRTIMRT + BRT
Dose (Gy)
0
20
40
60
80
100
0 20 40 60 80
CTV
XRT + BRTIMRT + BRT
Dose (Gy)
0
20
40
60
80
100
0 20 40 60 80
Bladder
XRT + BRTIMRT +BRT
Dose (Gy)
0
20
40
60
80
100
0 20 40 60 80
Boosts
Boost 1 XRT + BRTBoost 1 IMRT + BRTBoost 2 XRT + BRTBoost 2 IMRT + BRT
Dose (Gy)
Conventional 3DCRT+BRT vs. IMRT+BRT
N. Dogan / Nov 2007
QA tasks for IMRT
• Machine QA- Acceptance and routine QA of the MLC for IMRT delivery - dosimetric and geometric characteristics
• Algorithm QA for IMRT - QA of planning system and data consistency with machine
• Patient Specific QA – prove plan works 1D and 2D dosimetry of treatment components such
IM beams and segments 3D dosimetry of entire treatment delivery
• Post Treatment QA• Log-file analysis
N. Dogan / Nov 2007
Phantom Dose Verification
Beams on Patient Beams on Phantom
N. Dogan / Nov 2007
SampleFilm Dosimetry Results
Other Analysis Distance to Agreement Gamma …
Isodose Comparison
Profiles
Dose Differences
N. Dogan / Nov 2007
Compare isodoses (film) and absolute dose (chamber)
Current Practice
Courtesy of G. Ezzel, Ph.D., Mayo Clinic
N. Dogan / Nov 2007
Gamma Analysis
Measured Film
Adaptive Convolution
Monte Carlo
N. Dogan / Nov 2007
Calculation to Measurement Comparison
(b)
Measured Calculated
54% of points have a dose difference <2% or a DTA <2 mm
N. Dogan / Nov 2007
MC to Measurement Comparison
(b) (c)
Measured Calculated
Measurement and MC w transport through MLC 97% within 2%,2 mm
Measurement and MC using Tx planning systems Intensity Matrix
N. Dogan / Nov 2007
=10%Superposition Monte Carlo
MC Verification
N. Dogan / Nov 2007
The percentage of points, averaged over all of the plan’s treatment fields for each patient with ≥1 with 2% tolerance and 2 mm DTA.
MC (8.1% ± 3.7% points failed; range = 4.9% – 18.4%) SC (16.7% ± 5.6% points failed; range = 11.3% – 30.7%)
MC Verification of Prostate IMRT Plans
N. Dogan / Nov 2007 Courtesy of G. Ezzel, Ph.D., Mayo Clinic
Check standard patterns for constancy
N. Dogan / Nov 2007
DMLC field 14x14 cm2
at SSD =100 cm, 2 cm separated strips
• Using radiographic films Intensity-modulated
pattern field Check leaf position,
acceleration, motion stability
Check for hot and cold density
Visual check
Routine DMLC QA
Courtesy of Jean Moran, Ph.D, UofMichigan
N. Dogan / Nov 2007
Summary
• Inverse IMRT Planning is not intuitive – however, easy to establish protocols and class solutions for each specific site.
• Necessary to define realistic goals and constraints.
• Simplify IMRT plan as much as possible once you have acceptable solution.
• Know the limitations of your inverse treatment planning system.
N. Dogan / Nov 2007
Summary• Need to characterize the MLC system for
IMRT with special emphasis on penumbra, leaf leakage and transmission.
• Need to know the limits of the mechanical systems and interactions with controller and accelerator software for delivery.
• Continued need for improvements to software for delivery system, measurement devices, phantoms, and dose analysis tools.
N. Dogan / Nov 2007
Acknowledgements
Jeffrey Siebers – VCUGary Ezzel – Mayo ClinicMark Oldham – Duke UniversityJean Moran – U of MichiganIvaylo Mihalov - UofArkansas