quality assurance for image-guided radiotherapysource and detector position. marinello, aapm, prague...
TRANSCRIPT
MARINELLO, AAPM, Prague 2008
QUALITY ASSURANCE FOR
IMAGE-GUIDED RADIOTHERAPY
Ginette MARINELLO, Ph D
Unité de Radiophysique et Radioprotection du patient
Centre hospitalo-universitaire Henri Mondor51 avenue de Lattre de Tassigny
94010 CréteilFRANCE
MARINELLO, AAPM, Prague 2008
• A comprehensive QA program ismandatory at the timeof commissionningandto maintain the requirements ofthe system
QUALITY ASSURANCE PROGRAM
• Quality control (QC) tests should :
- be reasonable, simple and fast to perform
- include safety and functionalities, precision ofthe OBI and treatment couch movements, image quality and for CBCT or MV-CT, HU accuracy
- involve algorithms allowing mechanical couchcorrection or image reconstruction and matching, etc
- include patient dose evaluation
MARINELLO, AAPM, Prague 2008
IMAGE GUIDED RADIOTHERAPYInvolves multiple imaging procedures:
- during planning process:CT images sometimesfused with PET, SPECT, ultrasound or MR imaging
- to optimise patient set-up: film or electronicimaging using a kV simulator, or use of calculated digitallyreconstructed radiographs (DRRs) from the planning CT used as the simulation images
- to check patient set-up: use of EPID , diagnostic X-ray or kV or MV cone beam CT imaging
- to characterize organ or tumor motion due to therespiration or other influences: breathing correlated CT (referred as 4D CT) or linac, etc
MARINELLO, AAPM, Prague 2008
OVERVIEW OF EXISTING COMMERCIALLY AVAILABLE TECHNOLOGIES IN IMAGE GUIDANCE
• Planar imaging, EPID
• kV-CBCT , only availableon Elekta and Varian
MARINELLO, AAPM, Prague 2008
MV-CBCT , onlyavailable on Siemens
Paco hernandez 2007
MV-CT , only avalaibleon helical Tomotherapy
• Stereoscopic X-ray imaging, ultrasound imagingindependent of linac manufacturer...
MARINELLO, AAPM, Prague 2008
COMMISSIONNING AND QUALITY ASSURANCE
Irrespective of the imaging equipment used, the followingfeatures must be verified:
• mechanical and electrical safety
• geometrical performance and reproducibility
• calibration
• image quality
• sofware performance, etc
The controls must be undertaken only if mechanicalperformances and radiation stability of the linac arecorrect
MARINELLO, AAPM, Prague 2008
ANTI-COLLISION AND MECHANICAL VERIFICATIONS
After M. Djordjevic, 2007
MARINELLO, AAPM, Prague 2008
must be made and stored ...Example:
AAPM report 75 (TG58),2001
MARINELLO, AAPM, Prague 2008
CHECK OF AGREEMENT BETWEEN VISUAL AND DISPLAYED SCALES
Source and detector position
MARINELLO, AAPM, Prague 2008
MV -aSiEPID(AS 500)
kV-aSi detector(Pax Scan 4030 CB)
kV X-ray source
VARIAN- Clinac 2100 C
MARINELLO, AAPM, Prague 2008
EPID GENERALLY BASED ON:• matrix ion chambers• combination of camera and phosphor screen
• active matrix flat panel (photodiode array made of amorphoussilicium, or photoconductor array made of selenium)...
MARINELLO, AAPM, Prague 2008
EPID DETECTION PRINCIPLE
• Compton electrons produced in copper plate : energy released in the scintillator
=> fluorescence photons detected by photodiodes
VARIAN
aS 500 or
aS 1000
cover
copper plate(1 mm)
glass (0.1 mm)
a-Si photodiodes
Beam
cover
scintillator
MARINELLO, AAPM, Prague 2008
• copper platereplaced by aluminium plate
• anti-scatter grid addedto avoid scattered radiation frompatient and table
kV-X-RAY DETECTORS
cover
aluminium plate (2mm)
thin lead plate
a-Si photodiodes
Beam
cover
scintillator
antiscatter gridVARIANPax Scan
4030CB
RECEPTOR
MARINELLO, AAPM, Prague 2008
Photodiodes signal read out line by line and sent to a converter=> digitalized images through pulses continously processed
and displayed
Courtesy of VARIAN
IMAGE CONSTRUCTION
MARINELLO, AAPM, Prague 2008
QUALITY CONTROLS (QC)
1. ELECTRONIC PORTAL IMAGING DEVICE
(EPID)
MARINELLO, AAPM, Prague 2008
CALIBRATION OF EPID
• Method:
Should be carried out for all energies, dose-rates and geometrical conditions used for patients,following the recommendations madeby the vendor, and using the tools provided withthe system
• Purpose:
Check thatdark current, noiseandhomogeneityare acceptable
MARINELLO, AAPM, Prague 2008
HOMOGENEITY CHECK
NOISESD
A
B
B
A
RX 6MV20 x 26 cm2
maximal field size
MARINELLO, AAPM, Prague 2008
CALIBRATION OF EPID
Pixel defect can be detected and repaired through software interpolation
MARINELLO, AAPM, Prague 2008
IMAGE QUALITY : CONTRAST
whereΦP1, ΦP2, andΦS are theprimary and scatter photon fluences
( )/22ΦΦΦ
ΦΦ
signalmeansignal
CSP1P2
P1P2
++−==
( )SF1
2SFe1
e-12 C
∆-
-∆
−++
=
withwaterµboneµXL∆ −=
( )PSS ΦΦΦSF +=
AAPM, Report 58, 2001
MARINELLO, AAPM, Prague 2008EPID QC Phantom - PTW
minimum distancebetween 2 adjacent features, or
minimum size of a featurethatcan be detected by EPID.
IMAGE QUALITY : SPATIAL RESOLUTION
MARINELLO, AAPM, Prague 2008
Tolerances for EPID
• acceptable contrast :1%,
• spatial resolution 2-3 mm
IMAGE QUALITY
AAPM TG 58, Report N° 75
Med. Phys. 28, 712-737, 2001
⇓⇓⇓⇓
MARINELLO, AAPM, Prague 2008
Spatial resolution given by thevisible columnsandcontrast
deduced from thevisible lines
ALUMINIUM LAS VEGAS PHANTOM
6 MV
side to place opposite to X-ray source
MARINELLO, AAPM, Prague 2008
EPID: on line or off line 2 D tissue and bone imaging => patient set-up and treatment verification by comparison with
DRR or reference EPID image
DRR
EPID image matched with DRR
= > softwares check mandatory
MARINELLO, AAPM, Prague 2008
Alderson phantom
EXAMPLE OF SOFWARE CHECKS
CT Scan of the phantom
in "treatment" position
⇓
DDR PRODUCING
using TPS
⇓
Transfert to linac
through network
STEP 1
MARINELLO, AAPM, Prague 2008
EPID image
taken without
phantom
⇓
check of the field shape
using
sofware
⇓
check of geometric
reproduction of fieldshape
STEP 2
MARINELLO, AAPM, Prague 2008
Anterior field
EPID Image DRR
REFERENCE
VISUAL COMPARISON
STEP 3
MARINELLO, AAPM, Prague 2008
Lateral field
EPID Image DRR
ANATOMY
CHECK
WITH
CORRELATION
POINTS
REFERENCE
STEP 4
MARINELLO, AAPM, Prague 2008
2. MV-kV & kV-kV
QUALITY CONTROLS
MARINELLO, AAPM, Prague 2008After Bissonnette, 2007
MARINELLO, AAPM, Prague 2008
• safety and anticollision checks
• mechanical checks such as coincidence between rotation and beam axis,etc
note: no light field => need to use special toolsadapted to X-ray energyandmodeused.
kV-kV2 portal images
taken at0° and 90°either in
radiographic or fluoroscopic modes
MARINELLO, AAPM, Prague 2008
CHECK OF CENTRAL AXIS
1. Check lasers2. Align cube phantom with light and lasers3. Acquire AP and lateral images4. Check rotation axis
lead ball(ΦΦΦΦ = 2mm)
at the center ofthe cube
MARINELLO, AAPM, Prague 2008
CHECK OF IMAGE QUALITY
http://www.leedstestobjects.com
Same method as for EPID but with tools adapted to diagnostic energy*...
* AAPM Report N°74QA in Diagnostic Radiology,2002
MARINELLO, AAPM, Prague 2008
50kV – small focus spot
It is important to evaluatecontrastandspatial resolutionwith the sameparameters as those used for patients
MARINELLO, AAPM, Prague 2008CHECK OF IMAGE GEOMETRY ......
kV-kV
MARINELLO, AAPM, Prague 2008
Printed Board : Set-up
Source surface distance : 1m
Printed Board : Acquisition
Source surface distance : 1m
Various surface detector distance :
20cm to 50cm
MARINELLO, AAPM, Prague 2008
angles
44.9°
45°
distances 20.1 cm
20 cm
PRINTED BOARD
QA OF IMAGE GEOMETRY
To checked for bothperpendicular directions
MARINELLO, AAPM, Prague 2008
CHECK OF TABLE DISPLACEMENT
Example of tool used to checkautomatic correction of table displacement
by software
⇓
It can be used for:• kV-kV• kV- MV• CB- CT• MV- CT
MARINELLO, AAPM, Prague 2008
CHECK OF TABLE DISPLACEMENT
2. Check lasers3. Setup the phantom with light and lasers according to white
crosses4. Move the couch to set up the phantom according to the red or
green crosses5. Check the deviation founded by the software6. Repeat points 3 to 4 ; then check and apply automatic correction
of table displacement in order to check it
1. Acquire CT scan with phantom. Treatment isocenter aligned according to white crosses
accuracy and repeatability⇓⇓⇓⇓
MARINELLO, AAPM, Prague 2008
3. CONE-BEAM CT(CBCT)
QUALITY CONTROLS
MARINELLO, AAPM, Prague 2008
CONE BEAM CT: 3 D on linebone and soft tissue imaging=> patient set-up corrections and evaluation of organ movements
through comparison withCT used for planification
available on ELEKTA and
VARIAN linacs...
MARINELLO, AAPM, Prague 2008
CHARACTERISTICS OF VARIAN – CBCT(model VZW2940FH2-02)
• a kV x-ray sourceandan a-Si flat panel detectormountedorthogonally to the EPID (aSi500 PortalVision) on linacgantry via 2 robotically controlled arms
• source- axis distance 100 cm and axis - detector distance 50 cm
• detector area 40 x 30 cm2 => maximum beam size at isocentre:20 cm (longitudinal axis) and 26 cm (transversal axis)
• CBCT acquired by a rotation of the linac gantry over 370° (10° overlap on left side) and using x rays in radiographic or pulsed-fluoro mode (focal spots 0.4mm or 0.8 mm)
• 2 acquisition modes depending on the size of anatomical site ...
MARINELLO, AAPM, Prague 2008
CBCT ACQUISITION MODES
FULL FAN MODEused to scan site where the
size is less than 24 cm(head andneck), with detector centered
HALF FAN MODEused for sites larger than 24 cm(mostlyfor body scans), with detector shifted to
1 side to cover more than the halfpatient volume
SA
D =
100
cm
AD
D =
50
cm
KVS
KVD
KVS
KVD
Half Bow-Tie
BladesBlades
Full Bow-Tie
axis
MARINELLO, AAPM, Prague 2008
Dark grey = scanner, Cleargrey = CBCT
TOOLS TO CHECK THE QUALITY OF THE CORRECTION
MARINELLO, AAPM, Prague 2008
image avec couleurs
TOOLS TO CHECK THE QUALITY OF THE CORRECTION
Reference : CT imagesCB CT images
CB CT imagesCB CT images
CT window
CT window CT window
MARINELLO, AAPM, Prague 2008
lead ball(ΦΦΦΦ = 2mm)
at cube center
Coincidence between the image of lead ball and axis < 2mm (After Bissonnette, 2007)
CHECK OF ROTATION AXIS
MARINELLO, AAPM, Prague 2008
EXAMPLES OF ARTEFACTS
A B
A : Ring artefactB: Streaking
Truncation artefactAfter ZHANG, IJROBP,2005
MARINELLO, AAPM, Prague 2008
FLEXMAP
• ELEKTA: Flexes are corrected for in reconstruction algorithm
• VARIANFlexes are compensated for by roboticarm motion
If artefacts remain once the calibration of thedevice made again => defective componentsor sofware problems
MARINELLO, AAPM, Prague 2008
QC OF THE CBCT IMAGE
CATPHAN
phantoms 500 or 600
(or equivalent) allow
most of the checks at
the same timeusing
the same protocols as
for patientsphantom made ofdifferent modules
MARINELLO, AAPM, Prague 2008
MARINELLO, AAPM, Prague 2008
QUALITY CONTROL OF CBCT IMAGE
CBCT Varian (half-fan- max field-125 kV- mA- 25 s)
Uniformity ResolutionDetectability &slice thickness
Results obtained highly dependent of the equipment,sofware versions... and parameters of irradiation(beam hardening, scatter, energy, dose-rate...)
MARINELLO, AAPM, Prague 2008
IMAGES OF THE HOMOGENEOUS MODULE
Full fan mode Half-fan mode
VARIAN- CBCT
M. DJORDJEVICMaster of sciences
thesis, 2007
MARINELLO, AAPM, Prague 2008 After Bissonnette, 2007
MARINELLO, AAPM, Prague 2008
20 cm* (longitudinal axis) x 26 cm *maximal size
5 cm (longitudinal axis) x 26 cm at the axis
INFLUENCE OF THE SCATTERING VOLUME ONCONTRAST
MARINELLO, AAPM, Prague 2008After Bissonnette, 2007
MARINELLO, AAPM, Prague 2008
CBCT Varian
M. DJORDJEVICMaster of sciences
thesis, 2007
MARINELLO, AAPM, Prague 2008
IMAGES CAN BE VISUALLY
OR AUTOMATICALLY ANALYSED USING
SPECIAL SOFTWARES
MARINELLO, AAPM, Prague 2008
The The The The ARTiScanARTiScanARTiScanARTiScan* Solution* Solution* Solution* Solution****http:// http:// http:// http:// www.aquilab.comwww.aquilab.comwww.aquilab.comwww.aquilab.com
Integrated solution for Quality Controlin Diagnostic or Radiotherapy Imagingincluding:
• Equipmentindependent PC connected to the PACSor network
• Software ARTiScanDICOM, Image gestion
• Modules for automaticallyanalysing the images
Specific module for each test object and imaging device
MARINELLO, AAPM, Prague 2008
MARINELLO, AAPM, Prague 2008
Results can automatically becompared to :
- reference data,either published in national or international recommendations (IEC...) or establishedby users (if possible acceptance testsof equipment) when no recommendations are existing
reference
measuredvalues
- previous checksin order to point out possible changes andundertake preventive actions
INTEREST :
MARINELLO, AAPM, Prague 2008
scannerreferencecontour
CBCT
CBCT CBCT
shift
shift
shift
USE OF CBCT TO CHECK PATIENT SET-UP
MARINELLO, AAPM, Prague 2008
Automatic (version 2.0) ormanual correction
of couch position and angle
END OF CORRECTION
MARINELLO, AAPM, Prague 2008
EVALUATION OF PATIENT DOSE
MARINELLO, AAPM, Prague 2008
Howeverquestions are raised
• What is the dose delivered by IGRT ?
• Does daily use of IGRT for patient setup significantlyincreases the already high level of therapeutic radiation and the doses to the normal tissues of the patient ?
• How taking it into account?
Benefits of IGRT are obvious => possibility to seethe target and organ at risk (rectum) before the treatment
PATIENT DOSE
MARINELLO, AAPM, Prague 2008
IMAGING DOSE DURING IGRT
Although different in modality and method, allradiographic techniques have in common the fact thatthey can delivera significant radiation dose to the patient(especially when they are used for accurate day-to-dayset-up)
⇓⇓⇓⇓
REPORT OF THE AAPM TG 75
The management of imaging dose duringimage-guided radiotherapy
Medical Physics, Vol. 34, N°10,4041-4063
MARINELLO, AAPM, Prague 2008
EVALUATION OF IMAGING DOSE
Fundamental distinction between:
• dose: dose deposited locally per unit of mass by ionizing radiation (J/kg or Gy, mGy...).It depends on thefluence(rate and duration)
• integral dose: total dose integrated over the exposedmass (volume) of the anatomy (J).It depends on fluence and irradiated volume
⇓⇓⇓⇓
Absorbed dose is the meaningful quantity for estimating the biological effect of irradiationand integral dose to quantify the risk ofsecond cancers in future
MARINELLO, AAPM, Prague 2008
• Planar imaging(portal imaging, kV-kV, MV-kV) :doseto the patient /image greateast at the skin surfacenearest to the source and falls off progressively as the radiation transits the body to the image detector.
• Axial imaging (CT, CBCT, MV-CT):dose distributed nearly uniformly throughout the imagevolume to produce 3D image of uniform cross-sectionalquality
GENERALITIES
MARINELLO, AAPM, Prague 2008
DOSE DELIVERED BY EPID
• The image being created with the megavoltage beamwhich is used to treat the patient
• Generally source/detector distances are variable from120 to 150 cm => cumulative doseto acquire single image using aSi panels ranges from 1 to 5 MUfor most users(AAPM, 2007)
• For scenarios that involve daily or numerous portal images, it is recommended to substract the numberof MU used for imaging from the number of MUnecessary to treat the patient (possibility to do itautomatically with some sofware).
MARINELLO, AAPM, Prague 2008
DOSE DELIVERED BY kV- kV
When used in radiographic mode to acquireAP and lateral alignment radiographs, thedose to the patient is approximately1-3 mGyper image,depending on the technique (Perkins et al. 2006)
=> in most cases, it can be neglected.
MARINELLO, AAPM, Prague 2008
DOSE DELIVERED BY CBCT
kV CBCT
2.3 cGy
~ 5 cGy
2.3 cGy
Dose
(lateral)
68 cGy1.7 cGySiemens
(Prototype)3
120 cGy~3 cGyOBI (120 kV)2
64 cGy1.6 cGySynergy (120kV)1
80Gy
Daily image
Dose
(center)
Phantom
(Φ= 30 cm)
MV CBCTProstate (center) 6.4 cGy 256 cGy
1 M.K. Islam et al, Med Phys 33(6) 1573-1582, 2 N. Wen et al Phys Med Biol 2007 52 2267-2276, 3 U. Oelfke Comm. DKFZ Jan 2007
L.C. Peng et al, J. Appl Clin Med Phy 2007 8(1) 10-20
MARINELLO, AAPM, Prague 2008
The dose delivered by kV-CBCT can be optimalizedby adapting the collimation aperture to the size of thevolume to be explored:• by sofware (Varian linacs)• by inserting a special accessory(Elekta)
MARINELLO, AAPM, Prague 2008
• Irradiation conditions
identical to those used for
patient CBCT (prostate):
- 125 kV
- Half bow-tie
- 80 mA
- 25 ms/pulse
PREPARATION AND CHECK OF THE IVD METHOD
• Use of TLD 700H for point dose measurements…
• Phantom Rando Alderson
MARINELLO, AAPM, Prague 2008
... anduse of EBT GAFCHROMIC FILMfor evaluating the dose distribution
• A-P thickness 23 cm• Table height 11.5 cm• 15 CBCT =>
100% ~ 0.7 GyDpost
110 %
100 %
120 %130 %
90 %
xDleft
Dant
Dright
140 %
Dpost
Rectum
axis100% =
(Dant+Dpost) / 2
MARINELLO, AAPM, Prague 2008
• For geometrical and antomical reasons , maximal skin dosesare measured along theantero-posterior axisof the phantom (smallestthickness), and theirmean is close or higherto the dose at mid-thickness of the phantom or at therectum position (organ at risk)
=> mean = REFERENCE FOR PATIENT DOSE CORRECTION
DOSES (cGy)
6.25.44.7
--
4.9
--
4.9
6.66.86
86.55.5
1412.311.5*
Dmid-thick nessDrightDleftDpostDant
TH (cm)
RESULTS - 1 CBCT (Alderson phantom)
* mid-thickness
• Dose heterogeneous within the irradiated volumeand verydependent on the patient position/axis (TH)
MARINELLO, AAPM, Prague 2008
Posterior dosimeter under the tablesheet (centered with light beam)=> measured correction factor1.24
MARINELLO, AAPM, Prague 2008
/ 26 CBCT
MARINELLO, AAPM, Prague 2008
CONCLUSION
• In vivo evaluation of maximal dose susceptible to bedelivered to organ at risk (rectum dose for prostate irradiation) possible with TLD
• Important number of CBCT => contribution of dose to betaken into account... but how to do it and what EBR for low X-rays and high MV X-ray beams)*
*Wambersie Ph D recommends to use a correction factor of 1.13 for X-rays of 200 kV compare to 60Co but for higher doses
• Waiting for a general consensus, we have decided to subtract thephysical dosedelivered by CBCT from theprescribed dose (without correction for biological effects)