acknowledgement clinical implementation of imrt for …€“ zhongxing liao , md – joe chang, md...
TRANSCRIPT
1
Therapy Continuing Education Course
Clinical Implementation of IMRT for Lung Cancers
Therapy Continuing Education Course
Clinical Implementation of IMRT for Lung Cancers
H. Helen Liu, PhDDepartment of Radiation Physics,
U.T. MD Anderson Cancer Center, Houston, TXAAPM, Seattle, 2005
H. Helen Liu, PhDH. Helen Liu, PhDDepartment of Radiation Physics, Department of Radiation Physics,
U.T. MD Anderson Cancer Center, Houston, TXU.T. MD Anderson Cancer Center, Houston, TXAAPM, Seattle, 2005AAPM, Seattle, 2005
AcknowledgementAcknowledgement
Physics colleagues– Xiaochun Wang, PhD– Xiaodong Zhang, PhD– Maria Jauregui– Lei Dong, PhD– Siyoung Jang, PhD– Radhe Mohan, PhD
Physics colleaguesPhysics colleagues–– XiaochunXiaochun Wang, PhDWang, PhD–– XiaodongXiaodong Zhang, PhDZhang, PhD–– Maria Maria JaureguiJauregui–– Lei Dong, PhDLei Dong, PhD–– SiyoungSiyoung Jang, PhDJang, PhD–– RadheRadhe Mohan, PhDMohan, PhD
Oncology colleagues– Hussan Murshed, MD– Craig Stevens, MD– Thomas Guerrero, MD– ZhongXing Liao, MD– Joe Chang, MD– Melinda Jeter, MD– Ritsuko Komaki, MD– Jim Cox, MD
Oncology colleaguesOncology colleagues–– HussanHussan MurshedMurshed, MD, MD–– Craig Stevens, MDCraig Stevens, MD–– Thomas Guerrero, MDThomas Guerrero, MD–– ZhongXingZhongXing LiaoLiao, MD, MD–– Joe Chang, MDJoe Chang, MD–– Melinda Jeter, MDMelinda Jeter, MD–– RitsukoRitsuko Komaki, MDKomaki, MD–– Jim Cox, MD Jim Cox, MD
OverviewOverview• Introduction
– Clinical Rationales– Concerns and Myths– Clinical Applications
• Methodology– Patient Selection– Treatment Simulation– Target Delineation– Treatment Planning– Plan Evaluation
• Treatment Verification and QA
•• IntroductionIntroduction–– Clinical RationalesClinical Rationales–– Concerns and MythsConcerns and Myths–– Clinical ApplicationsClinical Applications
•• MethodologyMethodology–– Patient SelectionPatient Selection–– Treatment SimulationTreatment Simulation–– Target DelineationTarget Delineation–– Treatment PlanningTreatment Planning–– Plan EvaluationPlan Evaluation
•• Treatment Verification and QATreatment Verification and QA
Prerequisites: Basic Concepts of IMRT
Prerequisites: Basic Concepts of IMRT
• Intensity modulation• Fluence modulation (Open density matrix)• Pencil beam or beamlets• Beam delivery systems
– DMLC• Delivery: step-shoot, sliding window• Control of Dose: Control points, segments• MUs in IMRT
– Compensators
• Inverse planning or treatment planning optimization
•• Intensity modulationIntensity modulation•• FluenceFluence modulation (Open density matrix)modulation (Open density matrix)•• Pencil beam or Pencil beam or beamletsbeamlets•• Beam delivery systemsBeam delivery systems
–– DMLCDMLC•• Delivery: stepDelivery: step--shoot, sliding windowshoot, sliding window•• Control of Dose: Control points, segmentsControl of Dose: Control points, segments•• MUsMUs in IMRTin IMRT
–– CompensatorsCompensators
•• Inverse planning or treatment planning Inverse planning or treatment planning optimizationoptimization
2
[Introduction]
Clinical Rationales[Introduction]
Clinical Rationales• The benefits of IMRT for Lung Cancers
– Dose conformity to target volumes– Dose avoidance to normal structures
• Sharper dose gradient for adjacent critical organs: spinal cord, esophagus, etc
• Dose sparing for parallel organs: lung, heart, liver
•• The benefits of IMRT for Lung CancersThe benefits of IMRT for Lung Cancers–– Dose conformity to target volumesDose conformity to target volumes–– Dose avoidance to normal structuresDose avoidance to normal structures
•• Sharper dose gradient for adjacent critical organs: Sharper dose gradient for adjacent critical organs: spinal cord, esophagus, etcspinal cord, esophagus, etc
•• Dose sparing for parallel organs: lung, heart, liver Dose sparing for parallel organs: lung, heart, liver
Feasibility of sparing lung and other thoracic structures with intensity-modulated radiotherapy for non-small-cell lung cancer. IJROBP, 2004 Mar 15;58(4):1268-79. •Murshed, Liu, Liao, et al, Dose and volume reduction for normal lung using intensity-modulated radiotherapy for advanced-stage non-small-cell lung cancer. IJROB, 2004 Mar 15;58(4):1258-67.
[Clinical Rationales]Comparison of IMRT vs 3D for NSCLC
[Clinical Rationales]Comparison of IMRT vs 3D for NSCLC
IMRT 3D
[Clinical Rationales]
Comparison of IMRT vs 3D for NSCLC[Clinical Rationales]
Comparison of IMRT vs 3D for NSCLC
10
20
30
40
50
60
70
80
Per
cen
t lu
ng
vo
lum
e
V5-3D V5-IMRT V10-3D V10-IMRT V20-3D V20-IMRT
*Comparison based sliding-window technique, IMRT results will be improved further with step-shoot technique.
[Introduction]
Concerns and Myths[Introduction]
Concerns and Myths• IMRT spreads low dose volume for lung and
normal tissues– Lung parenchyma may be sensitive to low doses (5-
20 Gy)– Use of multiple IMRT beams ( > 6) may result in
increase of low-dose volume to normal tissues– The effect depends on
• Beam angle selection• Inverse planning process• Type of leaf sequences, leaf leakage, MU efficiency
•• IMRT spreads low dose volume for lung and IMRT spreads low dose volume for lung and normal tissuesnormal tissues–– Lung parenchyma may be sensitive to low doses (5Lung parenchyma may be sensitive to low doses (5--
20 20 GyGy))–– Use of multiple IMRT beams ( > 6) may result in Use of multiple IMRT beams ( > 6) may result in
increase of lowincrease of low--dose volume to normal tissuesdose volume to normal tissues–– The effect depends onThe effect depends on
•• Beam angle selectionBeam angle selection•• Inverse planning processInverse planning process•• Type of leaf sequences, leaf leakage, MU efficiencyType of leaf sequences, leaf leakage, MU efficiency
3
[Introduction]
Concerns and Myths[Introduction]
Concerns and Myths• Inter- and intra-fractional organ motion
– Respiratory motion can be a significant source of uncertainty for target delineation
– Interplay effect between tumor motion and leaf motion may increase dosimetry uncertainty
• The effect maybe minor for treatment courses with large numbers of fractions
– Respiratory motion may affect dose to normal structures (lung, heart, esophagus, cord, etc)
– Patient anatomies may change during treatment courses
•• InterInter-- and intraand intra--fractional organ motionfractional organ motion–– Respiratory motion can be a significant source of Respiratory motion can be a significant source of
uncertainty for target delineationuncertainty for target delineation–– Interplay effect between tumor motion and leaf Interplay effect between tumor motion and leaf
motion may increase motion may increase dosimetrydosimetry uncertaintyuncertainty•• The effect maybe minor for treatment courses with large The effect maybe minor for treatment courses with large
numbers of fractionsnumbers of fractions
–– Respiratory motion may affect dose to normal Respiratory motion may affect dose to normal structures (lung, heart, esophagus, cord, etc)structures (lung, heart, esophagus, cord, etc)
–– Patient anatomies may change during treatment Patient anatomies may change during treatment coursescourses
• Complexity of treatment planning and delivery– Simulation and planning process requires
experience and more effort/time – Increase of treatment and delivery time may
reduce patient compliance and comfort– More challenges in QA and dosimetry
verification compared to 3DCRT
•• Complexity of treatment planning and Complexity of treatment planning and deliverydelivery–– Simulation and planning process requires Simulation and planning process requires
experience and more effort/time experience and more effort/time –– Increase of treatment and delivery time may Increase of treatment and delivery time may
reduce patient compliance and comfortreduce patient compliance and comfort–– More challenges in QA and More challenges in QA and dosimetrydosimetry
verification compared to 3DCRTverification compared to 3DCRT
[Introduction]
Concerns and Myths[Introduction]
Concerns and Myths
[Introduction]
Clinical Applications[Introduction]
Clinical Applications• Lung cancers
– Non-small cell lung cancer• Superior sulcus tumors: improvement of target conformity
and sparing of spinal cord• Advanced stage (stage III, IV): improvement of target
coverage and sparing of lung and other OARs
– Small cell lung cancer• Limited and advanced stage: same as above
• Mesothelioma– Improvement of target coverage and sparing of
contra-lateral lung, liver, kidneys, cord, heart
•• Lung cancersLung cancers–– NonNon--small cell lung cancersmall cell lung cancer
•• Superior Superior sulcussulcus tumors: improvement of target conformity tumors: improvement of target conformity and sparing of spinal cordand sparing of spinal cord
•• Advanced stage (stage III, IV): improvement of target Advanced stage (stage III, IV): improvement of target coverage and sparing of lung and other OARscoverage and sparing of lung and other OARs
–– Small cell lung cancerSmall cell lung cancer•• Limited and advanced stage: same as aboveLimited and advanced stage: same as above
•• MesotheliomaMesothelioma–– Improvement of target coverage and sparing of Improvement of target coverage and sparing of
contracontra--lateral lung, liver, kidneys, cord, heartlateral lung, liver, kidneys, cord, heart
[Clinical Applications]
Non-small cell lung cancers[Clinical Applications]
Non-small cell lung cancers
Single Lesion Multiple Hilar Lesions
4
[Clinical Applications]
Lung cancers[Clinical Applications]
Lung cancers
Superior Sulcus Tumors Mesotheliomas
[Methodology]
Patient Selection[Methodology]
Patient Selection
• Patient selection based on disease characteristics– Nearby critical structures– Complex target volumes– Suitable target size
• IMRT may not offer significant advantage over 3DCRT for small lesions (earlier stage) or extremely large lesions (late stage)
– Primary NSCLC stage III lesions are ideal candidates for IMRT
•• Patient selection based on disease characteristicsPatient selection based on disease characteristics–– Nearby critical structuresNearby critical structures–– Complex target volumesComplex target volumes–– Suitable target sizeSuitable target size
•• IMRT may not offer significant advantage over 3DCRT for small IMRT may not offer significant advantage over 3DCRT for small lesions (earlier stage) or extremely large lesions (late stage)lesions (earlier stage) or extremely large lesions (late stage)
–– Primary NSCLC stage III lesions are ideal candidates for Primary NSCLC stage III lesions are ideal candidates for IMRTIMRT
[Methodology]
Patient Selection[Methodology]
Patient Selection
• Patient selection based on organ motion– Immobile tumors are preferred for IMRT (
tumor motion < 0.5 cm)– For mobile tumors, effects of tumor motion
needs to be considered with adequate margins and dosimetric impact on other structures
•• Patient selection based on organ motionPatient selection based on organ motion–– Immobile tumors are preferred for IMRT ( Immobile tumors are preferred for IMRT (
tumor motion < 0.5 cm)tumor motion < 0.5 cm)–– For mobile tumors, effects of tumor motion For mobile tumors, effects of tumor motion
needs to be considered with adequate needs to be considered with adequate margins and margins and dosimetricdosimetric impact on other impact on other structures structures
[Methodology]
Treatment Simulation[Methodology]
Treatment Simulation
• CT simulation– Patient preparation
• Immobilization device• Marking skin• Breathing training
(optional)– PET/CT
• CT attenuation correction• Regular PET
– CT• Freebreathing CT (to be
obselete)• 4DCT
– Data transfer
•• CT simulationCT simulation–– Patient preparationPatient preparation
•• Immobilization deviceImmobilization device•• Marking skinMarking skin•• Breathing training Breathing training
(optional)(optional)
–– PET/CTPET/CT•• CT attenuation correctionCT attenuation correction•• Regular PETRegular PET
–– CTCT•• FreebreathingFreebreathing CT (to be CT (to be
obseleteobselete))•• 4DCT4DCT
–– Data transferData transfer
PET/CT suite with in-room laser
5
[Methodology]
Treatment Simulation[Methodology]
Treatment Simulation• Patient Setups and Immobilization
– Alpha cradle• Stereotactic bodybag maybe preferred
for improved precision
– Wing board– Head holder (optional)– T-bar and arm up position– Reference markers are placed near
carina with relatively stable anatomy– Isocenter shift based on planning CT– Position variations
• Arm down or setup similar to head&neck cases can be customized for special situations
•• Patient Setups and ImmobilizationPatient Setups and Immobilization–– Alpha cradleAlpha cradle
•• StereotacticStereotactic bodybagbodybag maybe preferred maybe preferred for improved precisionfor improved precision
–– Wing boardWing board–– Head holder (optional)Head holder (optional)–– TT--bar and arm up positionbar and arm up position–– Reference markers are placed near Reference markers are placed near
carina with relatively stable anatomycarina with relatively stable anatomy–– IsocenterIsocenter shift based on planning CTshift based on planning CT–– Position variationsPosition variations
•• Arm down or setup similar to head&neck Arm down or setup similar to head&neck cases can be customized for special cases can be customized for special situationssituations
[Methodology]
Target Delineation[Methodology]
Target Delineation– 4DCT : Assess tumor/anatomy motion– PET/CT: Assess target extension and
nodal involvement
–– 4DCT : Assess tumor/anatomy motion4DCT : Assess tumor/anatomy motion–– PET/CT: Assess target extension and PET/CT: Assess target extension and
nodal involvementnodal involvement
[Methodology]
Target Delineation[Methodology]
Target Delineation
• Margins of target volumes– GTV– iGTV = U[GTVi] (from all respiratory phases
and PET/CT)– ITV = iCTV = iGTV + microscopic expansion– PTV = ITV + setup uncertainty
•• Margins of target volumesMargins of target volumes–– GTVGTV–– iGTViGTV = = U[GTVU[GTVii] (from all respiratory phases ] (from all respiratory phases
and PET/CT)and PET/CT)–– ITV = ITV = iCTViCTV = = iGTViGTV + microscopic expansion+ microscopic expansion–– PTV = ITV + setup uncertaintyPTV = ITV + setup uncertainty
[Methodology]
Treatment Planning[Methodology]
Treatment Planning
• Inverse planning for lung IMRT– IMRT Inverse Planning
• Region-of-interests (ROIs)• Fluence optimization• MLC sequences
– Plan Evaluation– Beam configuration
•• Inverse planning for lung IMRTInverse planning for lung IMRT–– IMRT Inverse PlanningIMRT Inverse Planning
•• RegionRegion--ofof--interests (interests (ROIsROIs))•• FluenceFluence optimizationoptimization•• MLC sequencesMLC sequences
–– Plan EvaluationPlan Evaluation–– Beam configurationBeam configuration
6
[Methodology]
Inverse Planning[Methodology]
Inverse Planning
• Optimization engine– Specification of objective functions (costs)
• Distance between current solution to the desired one
– Free parameters for optimization• beamlet intensity
– Search engine• Deterministic approaches (Gradient based)• Stochastic approaches
• Interface with optimization engine– Objective functions/constraints– Solution output and evaluation of results
•• Optimization engineOptimization engine–– Specification of objective functions (costs)Specification of objective functions (costs)
•• Distance between current solution to the desired oneDistance between current solution to the desired one
–– Free parameters for optimizationFree parameters for optimization•• beamletbeamlet intensityintensity
–– Search engineSearch engine•• Deterministic approaches Deterministic approaches (Gradient based)(Gradient based)•• Stochastic approachesStochastic approaches
•• Interface with optimization engineInterface with optimization engine–– Objective functions/constraintsObjective functions/constraints–– Solution output and evaluation of resultsSolution output and evaluation of results
[Methodology]
Inverse Planning[Methodology]
Inverse Planning• Secrets of inverse planning for lung IMRT
– Objective functions are the steering wheels for the optimization engine
– Planners need to know the behavior of the optimization engine and effects of choosing objective functions
– Planners need to know how to make compromises among conflicting goals
• Tumor vs. lung• Different OARs: lung, heart, cord, esophagus, tissue
•• Secrets of inverse planning for lung IMRTSecrets of inverse planning for lung IMRT–– Objective functions are the steering wheels for the Objective functions are the steering wheels for the
optimization engineoptimization engine–– Planners need to know the behavior of the Planners need to know the behavior of the
optimization engine and effects of choosing objective optimization engine and effects of choosing objective functionsfunctions
–– Planners need to know how to make compromises Planners need to know how to make compromises among conflicting goalsamong conflicting goals
•• Tumor vs. lungTumor vs. lung•• Different OARs: lung, heart, cord, esophagus, tissueDifferent OARs: lung, heart, cord, esophagus, tissue
*Method is more specific to Pinnacle and similar systems
• Problem– Each patient is unique– Appropriate objectives are difficult to foresee– Inverse planning involve many trial-errors
• Solution– Feedback guided, stepwise progressive,
iterative planning
•• ProblemProblem–– Each patient is uniqueEach patient is unique–– Appropriate objectives are difficult to foreseeAppropriate objectives are difficult to foresee–– Inverse planning involve many trialInverse planning involve many trial--errorserrors
•• SolutionSolution–– Feedback guided, stepwise progressive, Feedback guided, stepwise progressive,
iterative planningiterative planning
[Inverse Planning]
Inverse Planning by Iterative Planning
[Inverse Planning]
Inverse Planning by Iterative Planning
[Inverse Planning]
ROIs specific for Lung IMRT[Inverse Planning]
ROIs specific for Lung IMRT
• In addition to ROIs that are needed for regular 3D planning, it will be helpful to have the following ROIs to drive the inverse planning:– PTV_Expanded: PTV + 1~2 cm margin – PTV_Moat: 1~2 cm moat outside PTV_Expanded– Normal tissue: skin contracted until PTV_moat– Cord_Expanded: cord + 1cm margin– Esophagus_Expanded: esophagus + 1cm margin– Other hot spots (at the end of the planning)
•• In addition to In addition to ROIsROIs that are needed for regular that are needed for regular 3D planning, it will be helpful to have the 3D planning, it will be helpful to have the following following ROIsROIs to drive the inverse planning:to drive the inverse planning:–– PTV_Expanded: PTV + 1~2 cm margin PTV_Expanded: PTV + 1~2 cm margin –– PTV_Moat: 1~2 cm moat outside PTV_ExpandedPTV_Moat: 1~2 cm moat outside PTV_Expanded–– Normal tissue: skin contracted until PTV_moatNormal tissue: skin contracted until PTV_moat–– Cord_Expanded: cord + 1cm marginCord_Expanded: cord + 1cm margin–– Esophagus_Expanded: esophagus + 1cm marginEsophagus_Expanded: esophagus + 1cm margin–– Other hot spots (at the end of the planning) Other hot spots (at the end of the planning)
7
[Inverse Planning]
ROIs specific for IMRT[Inverse Planning]
ROIs specific for IMRT
PTV PTV_moat Tissue
[Inverse Planning]
Inverse Planning by Iterative Planning
[Inverse Planning]
Inverse Planning by Iterative Planning
Step 1. Start by using default objective function templates that include:
• CTV: min dose• PTV: min dose, max dose, (uniform dose)• PTV moat: max dose• Total lung:
– V5 (60%)– V10 (45%)– V20 (35%)– Mean lung dose (15 Gy)
• Cord_Exp: max dose (45Gy)• Heart: V45 (30%)• Esophagus_Exp: V45 (30%)• Normal tissue: max dose or V20
– Disadvantage: increase of parameter space
Step 1. Start by using default objective function templates Step 1. Start by using default objective function templates that include:that include:
•• CTV: min doseCTV: min dose•• PTV: min dose, max dose, (uniform dose)PTV: min dose, max dose, (uniform dose)•• PTV moat: max dosePTV moat: max dose•• Total lung: Total lung:
–– V5 (60%)V5 (60%)–– V10 (45%)V10 (45%)–– V20 (35%)V20 (35%)–– Mean lung dose (15 Mean lung dose (15 GyGy))
•• Cord_Exp: max dose (45Gy)Cord_Exp: max dose (45Gy)•• Heart: V45 (30%)Heart: V45 (30%)•• Esophagus_Exp: V45 (30%)Esophagus_Exp: V45 (30%)•• Normal tissue: max dose or V20Normal tissue: max dose or V20
–– Disadvantage: increase of parameter spaceDisadvantage: increase of parameter space
[Inverse Planning]
Demo of an Example Case[Inverse Planning]
Demo of an Example Case
• insert•• insertinsertStep 2:Assign equal weighting to all objectives.
Step 3:•Run 1 search iteration only;•Evaluate current solutionwhich is similar to a 3D plan;•Adjust the objectives accordingly and their associated costs.
[Inverse Planning]
Inverse Planning by Iterative Planning
[Inverse Planning]
Inverse Planning by Iterative Planning
Step 4 and beyond: to balance the priorities and conflicting goals– Evaluate optimization solution– Re-adjust objective functions and their costs– Objectives with the highest costs will be pushed
down first during the next iteration loop– Follow the sequence of priority and organ sensitivity
A. Target coverageB. Lung dose/volumeC. Heart, esophagus, cordD. Normal tissues, hot spot
– Continue based on existing solution
Step 4 and beyond: to balance the priorities and Step 4 and beyond: to balance the priorities and conflicting goalsconflicting goals–– Evaluate optimization solutionEvaluate optimization solution–– ReRe--adjust objective functions and their costsadjust objective functions and their costs–– Objectives with the highest costs will be pushed Objectives with the highest costs will be pushed
down first during the next iteration loopdown first during the next iteration loop–– Follow the sequence of priority and organ sensitivityFollow the sequence of priority and organ sensitivity
A. Target coverageA. Target coverageB. Lung dose/volumeB. Lung dose/volumeC. Heart, esophagus, cordC. Heart, esophagus, cordD. Normal tissues, hot spot D. Normal tissues, hot spot
–– Continue based on existing solutionContinue based on existing solution
8
[Inverse Planning]
Demo of an Example Case[Inverse Planning]
Demo of an Example Case
• insert•• insertinsert
Iterative feed-back guided inverse planning
•Number of optimization iterations does not have to exceed 5 ~ 8 for gradient algorithms
•Choose the battle wisely, the key issue is to set appropriate objectives
•Upon completion of each run, critically assess the results and re-adjust objectives and costs, and rerun upon existing results
[Inverse Planning]
Plan Evaluation[Inverse Planning]
Plan Evaluation• Isodoses
– Target conformity vs. hot spots – Dose avoidance to ROIs, particularly lung– Spread of low-dose volume to lung and normal tissue
• DVHs– Evaluate whether objectives/constraints being placed
properly– Adjust objectives if reoptimization is required
• Other biological parameters– Mean dose or EUD for lung– NTCP
•• IsodosesIsodoses–– Target conformity vs. hot spots Target conformity vs. hot spots –– Dose avoidance to Dose avoidance to ROIsROIs, particularly lung, particularly lung–– Spread of lowSpread of low--dose volume to lung and normal tissuedose volume to lung and normal tissue
•• DVHsDVHs–– Evaluate whether objectives/constraints being placed Evaluate whether objectives/constraints being placed
properlyproperly–– Adjust objectives if Adjust objectives if reoptimizationreoptimization is requiredis required
•• Other biological parametersOther biological parameters–– Mean dose or EUD for lungMean dose or EUD for lung–– NTCPNTCP
[Inverse Planning]
Plan Evaluation[Inverse Planning]
Plan Evaluation[Inverse Planning]
MLC Sequence Conversion[Inverse Planning]
MLC Sequence Conversion• Deliverable plans are often degraded from
fluence-optimized plan• May have to reoptimize plan due to degradation of
leaf conversion• On Pinnacle: May perform direct segment
optimization for converted plan. If this is necessary, adjust objective functions first before reoptimization
• Minor manual adjustment to segments can be helpful to reduce cold/hot spots
• Direct leaf sequence optimization is another option
•• Deliverable plans are often degraded from Deliverable plans are often degraded from fluencefluence--optimized planoptimized plan
•• May have to May have to reoptimizereoptimize plan due to degradation of plan due to degradation of leaf conversionleaf conversion
•• On Pinnacle: May perform direct segment On Pinnacle: May perform direct segment optimization for converted plan. If this is optimization for converted plan. If this is necessary, adjust objective functions first before necessary, adjust objective functions first before reoptimizationreoptimization
•• Minor manual adjustment to segments can be Minor manual adjustment to segments can be helpful to reduce cold/hot spotshelpful to reduce cold/hot spots
•• Direct leaf sequence optimization is another Direct leaf sequence optimization is another optionoption
9
[Inverse Planning]
MLC Sequence Conversion[Inverse Planning]
MLC Sequence Conversion
• Balance between delivery efficiency (#segments and MUs) vs. dose gradient (conformity and avoidance)
• In general, #segment/beam should be < 20 for lung plans, it is not necessary to exceed more than 30 segments/beam
• Treatment planning system may not be adequate to compute dose accurately for plans with MU efficiency < 25%
•• Balance between delivery efficiency (#segments and Balance between delivery efficiency (#segments and MUsMUs) ) vs. dose gradient (conformity and avoidance)vs. dose gradient (conformity and avoidance)
•• In general, #segment/beam should be < 20 for lung plans, it In general, #segment/beam should be < 20 for lung plans, it is not necessary to exceed more than 30 segments/beamis not necessary to exceed more than 30 segments/beam
•• Treatment planning system may not be adequate to Treatment planning system may not be adequate to compute dose accurately for plans with MU efficiency < compute dose accurately for plans with MU efficiency < 25%25%
average_MU_per_angle =Total_MU_per_angle
Num_beam_per_angle
MU_efficiency = Fractional_prescription_dose (cGy)
sum(average_MU_per_angle)
[Inverse Planning]
Beam Configuration[Inverse Planning]
Beam Configuration• 6MV photon beams are preferred choice• 18MV beams should be avoided if possible
– Electron disequilibrium– Neutron production
• Coplanar beams are more practical and easy for planning
• Noncoplanar beams may offer additional choices for beam angle optimization
•• 6MV photon beams are preferred choice6MV photon beams are preferred choice•• 18MV beams should be avoided if possible18MV beams should be avoided if possible
–– Electron disequilibriumElectron disequilibrium–– Neutron productionNeutron production
•• Coplanar beams are more practical and easy for Coplanar beams are more practical and easy for planningplanning
•• NoncoplanarNoncoplanar beams may offer additional beams may offer additional choices for beam angle optimizationchoices for beam angle optimization
[Inverse Planning]
Beam Configuration[Inverse Planning]
Beam Configuration• Placing of beam angles should carefully consider
planning priorities for normal structures– PTV is not sensitive to the beam angles– Lung is the determining factor for selecting beam angles– Heart is more sensitive to angle selection than esophagus and
cord
• 4-6 beams should be sufficient for lung IMRT• Excessive beams will reduce MU efficiency and
delivery complexity/time• Experience from 3DCRT on optimal angles can be
extended to IMRT
•• Placing of beam angles should carefully consider Placing of beam angles should carefully consider planning priorities for normal structuresplanning priorities for normal structures–– PTV is PTV is notnot sensitive to the beam anglessensitive to the beam angles–– Lung is the determining factor for selecting beam anglesLung is the determining factor for selecting beam angles–– Heart is more sensitive to angle selection than esophagus and Heart is more sensitive to angle selection than esophagus and
cordcord
•• 44--6 beams should be sufficient for lung IMRT6 beams should be sufficient for lung IMRT•• Excessive beams will reduce MU efficiency and Excessive beams will reduce MU efficiency and
delivery complexity/timedelivery complexity/time•• Experience from 3DCRT on optimal angles can be Experience from 3DCRT on optimal angles can be
extended to IMRT extended to IMRT
[Inverse Planning]
Beam Configuration[Inverse Planning]
Beam Configuration5B-IMRT
9B-IMRT
5B-IMRT
9B-IMRT
•Use of 4-6 beams can achieve essentially equivalent plan quality compared to 9 beams
•Use of fewer beams require beam angle optimization that minimize lung dose-volume
10
[Inverse Planning]
Beam Configuration[Inverse Planning]
Beam Configuration
0
200
400
600
800
1000
1200
1400
3D IMRT 5 IMRT 7 IMRT 9
MU
s
0
50
100
150
200
3D IMRT 5 IMRT 7 IMRT 9
Num
Seg
men
ts
•MUs and #segments increase with #beams
•Reduction of #beams improves delivery efficiency and low-dose leakage
•Compromise between #beams and likelihood of hotspots
[Methodology]
Treatment Delivery and QA[Methodology]
Treatment Delivery and QA• QA procedure should be similar to other sites• Frequent imaging maybe needed to ensure
accuracy and precision of patient positioning• Dosimetry issues specific to lung cancers
– More significant tissue inhomogeneities– Large field sizes and high degree of intensity
modulation– Low doses in lung and normal tissues maybe more
difficult to compute accurately by conventional treatment planning systems
•• QA procedure should be similar to other sitesQA procedure should be similar to other sites•• Frequent imaging maybe needed to ensure Frequent imaging maybe needed to ensure
accuracy and precision of patient positioningaccuracy and precision of patient positioning•• DosimetryDosimetry issues specific to lung cancersissues specific to lung cancers
–– More significant tissue More significant tissue inhomogeneitiesinhomogeneities–– Large field sizes and high degree of intensity Large field sizes and high degree of intensity
modulationmodulation–– Low doses in lung and normal tissues maybe more Low doses in lung and normal tissues maybe more
difficult to compute accurately by conventional difficult to compute accurately by conventional treatment planning systemstreatment planning systems
[Treatment Verification]
Dosimetry QA[Treatment Verification]
Dosimetry QA
• Sources of dose calculation uncertainties– Tissue inhomogeneities– Beam modeling– MLC modeling– Dose calculation algorithms
• Pencil-beam algorithms• Convolution algorithms
•• Sources of dose calculation uncertaintiesSources of dose calculation uncertainties–– Tissue Tissue inhomogeneitiesinhomogeneities–– Beam modelingBeam modeling–– MLC modelingMLC modeling–– Dose calculation algorithmsDose calculation algorithms
•• PencilPencil--beam algorithmsbeam algorithms•• Convolution algorithmsConvolution algorithms
[Treatment Verification]
Dosimetry QA[Treatment Verification]
Dosimetry QA
• Commissioning/Implementation of lung IMRT procedure – Intensity verification
• Ion chamber in water phantom• Film in solid water phantom
– In-vitro dosimetry• TLDs in anthropomorphic phantoms
– Monte Carlo calculations
•• Commissioning/Implementation of lung Commissioning/Implementation of lung IMRT procedure IMRT procedure –– Intensity verificationIntensity verification
•• Ion chamber in water phantomIon chamber in water phantom•• Film in solid water phantom Film in solid water phantom
–– InIn--vitro vitro dosimetrydosimetry•• TLDsTLDs in anthropomorphic phantomsin anthropomorphic phantoms
–– Monte Carlo calculationsMonte Carlo calculations
11
[Treatment Verification]
Phantom Measurements[Treatment Verification]
Phantom Measurements[Treatment Verification]
Phantom Measurements[Treatment Verification]
Phantom Measurements
•Comparison of Pinnacle calculations (v6.2) vs. TLD measurements for lung IMRT cases from high to low dose regions
[Treatment Verification]
Monte Carlo Based QA[Treatment Verification]
Monte Carlo Based QA•Comparison of Corvus calculations (v4; v5) with Monte Carlo simulation for mesothelioma cases
MCS – Total Dose 50 Gy
5500 5000 4000 3000 2000 1000 cGy< -500 - 250 250 > 500 cGy
-10% -5% +5% +10%
Diff = MCS - Corvus
[Methodology]
Dosimetry Verification
[Methodology]
Dosimetry Verification
• Ensure dose calculation accuracy• in high-medium dose region • using the types of leaf sequences generated within the
planning system itself
• Treatment planning systems may underestimate dose in low dose region• Strongly depends on beam modeling and MLC
modeling (leaf transmission, leakage)• Effects is more prominent for beams with low MU
efficiency, I.e. greater leakage
• Ensure dose calculation accuracy• in high-medium dose region • using the types of leaf sequences generated within the
planning system itself
• Treatment planning systems may underestimate dose in low dose region• Strongly depends on beam modeling and MLC
modeling (leaf transmission, leakage)• Effects is more prominent for beams with low MU
efficiency, I.e. greater leakage
12
[Treatment Verification]
Dosimetry Verifications
[Treatment Verification]
Dosimetry Verifications
• Tissue inhomogeneity may not be a significant cause of error for lung IMRT, even using Pencil-beam algorithms (based on Corvus results)
• QA for single IMRT beam may not be adequate, composite dose distribution is more sensitive to dose errors
• Monte Carlo simulation is a powerful/effective tool for IMRT QA• Provide independent MU and dose distribution
verification• However, MCS can also be subjective to beam
parameters used for IMRT• Also requires rigorous commissioning process
• Tissue inhomogeneity may not be a significant cause of error for lung IMRT, even using Pencil-beam algorithms (based on Corvus results)
• QA for single IMRT beam may not be adequate, composite dose distribution is more sensitive to dose errors
• Monte Carlo simulation is a powerful/effective tool for IMRT QA• Provide independent MU and dose distribution
verification• However, MCS can also be subjective to beam
parameters used for IMRT• Also requires rigorous commissioning process
SummarySummary1. IMRT can be an effective treatment modality for managing
advanced stage NSCLC and other suitable lung cancers (superior sulcus, meso, etc)
2. Patient candidates need to be identified to maximize benefits of IMRT
3. Target delineation and organ motion need to be carefully considered during simulation
4. Low-dose volume of lung and normal tissue need to be reduced when planning for beam angles and dose distributions
5. Dosimetry accuracy should be validated for each treatment planning system
1.1. IMRT can be an effective treatment modality for managing IMRT can be an effective treatment modality for managing advanced stage NSCLC and other suitable lung cancers advanced stage NSCLC and other suitable lung cancers (superior (superior sulcussulcus, , mesomeso, etc), etc)
2.2. Patient candidates need to be identified to maximize Patient candidates need to be identified to maximize benefits of IMRTbenefits of IMRT
3.3. Target delineation and organ motion need to be carefully Target delineation and organ motion need to be carefully considered during simulationconsidered during simulation
4.4. LowLow--dose volume of lung and normal tissue need to be dose volume of lung and normal tissue need to be reduced when planning for beam angles and dose reduced when planning for beam angles and dose distributionsdistributions
5.5. DosimetryDosimetry accuracy should be validated for each treatment accuracy should be validated for each treatment planning systemplanning system
Questions & DiscussionsQuestions & Discussions
Contact: Helen Liu, Unit 94, Radiation Physics,UT-MDACC, 1515 Holcombe BlvdHouston, TX [email protected]