PositronEmission Tomography forRadiation Treatment

J. Daniel Bourland, PhD

Department of RadiationOncologyBioanatomic ImagingandTreatment ProgramWake ForestUniversity Schoolof Medicine

Winston-Salem,North Carolina, USAbourland@wfubmc.edu

Images© of citedauthorsor corporations

Outline: Part1

• Physicsof PET andPET-CT

• OncologyImagingwith FDG PET

• PETandRadiationTreatmentPlanning

• Summary

Work supportedin partby research grantsfrom North CarolinaBaptist Hospital, VarianMedical Systems,andGEHealthcare

Positron Annihilation

ββββ+

ββββ-nucleus

β+ range

511 keV γ

511 keV γ180o

• Coincidencedetection of two0.511 MeV photons

• Annihilation radiation frompositron-electron pair

• Photondirectionsat180oatannihilationpoint -different fromdecay point� rangeof positron

• Positronemitters with biologicalcompatibility

• Low Z (typically), protonrich,short half-lives

• “Local” productionwith acyclotron

Physics of PETandPET-CT• CE-Imaging: The Physicsand Technology of Radionuclide

Imaging – IIIWednesday,7:30:00AM - 9:25:00AM, Room:L100J– 7:30 AM WE-SAMS-L100J-1BasicPhysicsof PET and PET/CT - F.

Fahey*– 8:25 AM WE-SAMS-L100J-2PET/CT Performance Evaluation

Techniquesand Quality Assurance- B. Kemp*

• PET-CT Workshop• Wednesday,4:00:00 PM - 5:30:00PM, Room:L100E

WE-E-L100E-1 PET CT Workshop - PET/CT ScannerOffering From Different M anufacturers - J. Anderson*, O.Mawlawi*, M. Georgiou*

PET Radionuclidesbiologically compatible,short half-life

Radionuclide Half-Life “ run” speed

Fluorine18 (18F) 110min fast

Carbon 11 (11C) 20 min

Nitrogen13 (13N) 10 min

Oxygen15 (15O) 122sec

Rubidium82 (82Ru) 75 sec very fast(cardiac) Non-corrected

PETCs-attenuatedcorrected

CT-attenuatedcorrected

Phantomwith‘normal’ uptakeexcept left lungthathasair andonespherical‘hottumor’

CT image

Courtesyof K Mah, Univ Toronto,Sunnybrook

With

Without

1. Intensities under-estimatedat depth2. Intensities over-estimated on surface3. Intensities over-estimated for low

density objects (lung)

With and WithoutAttenuation Correction

1.3.

2.

FromRohren, Turkington,Coleman: Radiology2004;231:305-332

FDG PETImaging

PET Imaging in Oncology• Diagnosis– lesscommon

• Staging- yes• TargetDefinition

– Radiationtreatment

– Other“ targeted” therapy

• Re-staging– yes

• TreatmentEvaluation

PET Imaging• Activity distribution is imaged

– Physiology, function, biology

• Complementary to(~anatomic)CT and MR

• Increasedsensitivitycomparedto CT alone

• Most indicationsapprovedforreimbursementareoncology

2-[18F]-fluoro-2-deoxy-D-glucose(FDG)

• 18Fluorine-FDG is mostcommontracer• Malignant tissuesmayhaveincreasedglycolysisandincreasedFDG uptake• FDG-PETimagesare mapsof glucosemetabolism – non-specificimaging

• Diff icultieswith smalltumors(< 3 – 10 mm diameter)• Use increasing– 200 PET-CT scanners in 2 years*• PET-CT hybrid scanners:registration solved

Courtesyof K Mah, Univ Toronto,Sunnybrook*FromBradley, Thorstad, Mutic, et al.,IJROBP 59(1):78-86,2004.

Example: Non-SmallCell LungCancerDiagnosis Staging Re-Staging

FromRohren, Turkington, Coleman: Radiology2004;231:305-332

Example: Head& Neck CancerDiagnosis Staging Re-Staging

FromRohren, Turkington, Coleman:Radiology 2004; 231:305-332

PET and Radiation TreatmentPlanning

• Main contribution – Staging: TargetLocalization– Very important:stagedeterminestreatmentapproach

– Binary results:presence/absenceof disease,metastasis

• Dramaticdifferencespossible– Treatmentmode:reamo, chemo,beamo(EGShaw)

• none,one,all three?10-30% NSCLCpatientsstagechanges

– Radiationtreatment fields: ie, inclusion of nodes• Estimation – bettercoverageof targetwith PET in 30 – 60%of

patients receiving definitive radiationtreatment*

*FromBradley, Thorstad, Mutic, et al., IJROBP59(1):78-86,2004.

ColonCancer:PossibleTreatmentFieldsNodeNegative NodePositive

Adaptedfrom Rohren, Turkington, Coleman:Radiology 2004; 231:305-332

No TreatmentSimpleField Field IncludesNodalRegion

FDG PETand Staging, Localization

Changein stagebeforeandafter PET-CT

FromKoshy, Paulino, Howell et al., Head& Neck,27:6494-502,2005.

Impacton sizeof PTV: CT vs PET/CT

0

200

400

600

800

1000

1200

0 200 400 600 800 1000 1200

PTVCT (cm 3)

PT

VC

T/F

DG

(cm

3 )

Observer 1

Observer 2

Observer 3

• Relativechangesrangedfrom 0.40to 1.86

• In 5 of 23 cases,newFDG nodesdetectedthatincreasedPTV

MahMah, Caldwell,, Caldwell,UngUng et al. IJROBP52,2002et al. IJROBP52,2002Courtesyof K Mah, Univ Toronto,Sunnybrook

FDG PETandRadiationTreatmentChangein PTV andprescribeddose

FromKoshy, Paulino, Howell et al., Head& Neck,27:6494-502,2005.

PETmaydecreasetargetvolumesvs CT

CT: Purple PET/CT: Green

Courtesyof K Mah, Univ Toronto,Sunnybrook

Changesin targetoutlinetranslateto reducedtreatmentfield size

PTV basedon PET/CT

PTV basedon CT only

GTV: CT v PET-CTUseof PET-CT mayreduceGTV/CTV

GTV-CT GTV-PET-CT

FromSchwartz, Ford,Rajendran, et al., Head& Neck27(6): 478-487, 2005.

Ex: FDG-PETincreasedthe targetvolume

Pre-PET: RLL lesion with rightparatracheal node

Post-PET: RLL lesionwith rightparatracheal; subcarinal; andhilarlymphadenopathy (some< 1 cmon CT)

GTV increased to encompassthesenodalregions.

Courtesyof K Mah, Univ Toronto,Sunnybrook

Impactof PET-CT on RT VolumeDelineationUsingCombinedPET-CT Scanner: WashU Group

• Prospective study : 26 NSCLC patients– 8/26(31%)PETchangedstaging– Of 24 still plannedradically

• In 3, PET/CTreducedvolumecomparedto CT alonedueto atelectasis• In 10,PET/CT increasedvolumeasunsuspected nodaldisease was

detected• In 1, newseparatetumorfocuswas foundin samelung

– Overall PET/CT resultedin alterationsin radiationplanninginover 50%of patients by comparisonto CT alone.

Bradley,J et al. IJROBP59,2004

Process: PETin Radiation Treatment

• Cancer diagnosis� biopsy,imaging

• Treatmentposition, immobilization

• PETimaging in treatmentposition

• Expert imagereview

• Imagetransfer– DICOM, other?

• Imageregistration (if needed)

• Targetlocalization anddefinition

• Treatmentplanning

PET-CTSimulator

• Adjacentcontrol, scanner,inject-wait, lab, and toilet

• 1/8 in Pb; control, scanner– adjacentwaiting, scannerbkg

• 1/2 in Pb; inject-wait, toilet• Isotopeprep near on-site• CT, PET-CT operationis

quality – future additionalQA testing

• SharedVirtual Simulation• Lasermarkingsystem• AutomatedPACSarchive,

selective pushto TPS• Normalaccesssecurity

MR

PET-CT

MR Control

VirtualSim

PET

-CT

Control

MRAnnex

Inject-Wait

Office

Office

Wait

MREquip

Wait

Target Definition

• Qualitative:Expert clinical review– Visual, inclusionof clinical historyanddata

– Abovebackground

• Quantitative:Voxel intensityvalues– 40-50%of peak intensity (abovebackground?)

– StandardizedUptakeValue(SUV) of ROI• ie, SUV > 2.5 indicatespositivefor cancer

– Regiondeterminedby PET,extentby CT

RadiationTreatmentPlanning with PETTargetDefinition

Dependingon image windowandlevel setting, targetvolumecanchangeby 50%.Also:a) SUV not apartof DICOM datab) SUV util ity unclear

Samepatientimage:different window andlevel!

TheDigital ContourA Threshold Process

Peak

ThresholdValue

Bkg

TargetWidth

TheDigital ContourA Threshold Process

Peak

ThresholdValue

Bkg

TargetWidth

StandardizedUptakeValue (SUV)

• Semi-quantitativemeasureof glucosemetabolism• Essentially:Averagevoxel valuewithin anROI

– normalizedby activity andbody weight

• Relative to anindividualpatient– Malignancyv benign– Tumorgrade– Treatmentresponse– Prognosis (?) – biological models

In part from Rohren, Turkington, Coleman:Radiology 2004; 231:305-332

SUV Limitations

• Semi-quantitativemeasureof glucosemetabolism• Definition of theROI (CT? self-referencing),and its

location overtime (ie, scanto scan)• Tumorheterogeneity: necrosis, variablegrade• Tumorvolumechangeswith time• Smalltumorsdifficult to image (size� resolution)• Glucoseload?

Consensus?Quantificationof FDG uptake?

In partfrom Greven: SemRadOnc14:2, 2004.

PET-BasedTargeting – H/N

K Greven, PI

• 7 methodsThre shold vs Method H&N

0

2

4

6

8

10

40% 45% 50% BG 1 BG 2 BG 3 BG 4

Thresh old Method

Th

resh

old

SU

V

Patient 1

Patient 2

Patient 3

Lesion Volume vs Method H&N

0

5

10

15

20

25

30

35

40% 45% 50% BG 1 BG 2 BG 3 BG 4

Thresh old Method

Vo

lum

e(c

c)

Patient 1

Patient 2

Patient 3

airway + signal

Unpublished:Lawrence, Greven, Bourland,et al. 2007

Target Definition: “a mess” (R Jeraj)

Nestle et al 2005, J Nucl Med 46 (8), 1342.

TargetDefinition: “a mess”

Nestle et al 2005, J Nucl Med 46 (8), 1342.

DeviceQAGeometricPhantoms

GTV-CT andGTV-PETOtherIssue:Inter-observerVariation

PET CT

FromCaldwell, Mah, Ung, et al., IJROBP51(4):923-931, 2001

Acknowledgements

• Work supportedin part by researchgrantsfromNorth CarolinaBaptistHospital,Varian MedicalSystems,andGE Healthcare

• KathyMah, Universityof Toronto,Sunnybrook• SasaMutic, Washington University in St. Louis• AssenKirov, MemorialSloanKettering,NY, NY

Summary: Part1

• Physicsof PET andPET-CT

• PETin oncology:Staging� TreatmentPath

• PETandRadiationTreatmentPlanning– Localization of regions to betreated

– Targeting,contourdelineation, with limits

RobertJeraj, PhD

PET in TreatmentAssessment