RPC physicist travels to the institution: ◇ Measurements, interviews,
documentation review. ◇ Exit Interview with Radiation Oncologist and Medical
Physicist.
The Radiological Physics Center Quality Assurance in Clinical Trials David Followill, Ph.D. Director Radiological Physics Center
Radiological Physics Center •
Formed in 1968 by agreement between AAPM and Committee for Radiotherapy Studies (CRTS) chaired by G. Fletcher to monitor institutions participation in clinical trials
•
Funded continuously by NCI for 43 years even as structure of cooperative group programs have changed
Mission •
Assure NCI and cooperative groups that institutions participating in clinical trials deliver prescribed doses that are comparable and consistent. (Minimize dose uncertainty)
•
Help institutions to make any corrections that might be needed.
•
Report findings to the community.
Ideal dose delivery per protocol
Mission of the RPC is to minimize the dose uncertainty
Add errorinina dose Results No we have a less than orflatter delivery dose desirable dose distribution response curve within the cohort and thus lower response than expected
Pettersen et al 2008
Clinical Trial Participants Number of Active Institutions – 1,842 ~3,400 megavoltage machines ~22,124 active megavoltage beams Active Machines
2000
# of M achines
1800 1600
4000 3000
1400
1000
Year
09 20
03
01 0
06 20
Year
20
2 19 007 94 20 19 08 97 20 20 09 00 2
19 1 709 9
19 2 0 85 04 19 2 0 0 88 5 2 19 0 0 91 6
19 2 7300 0 19 2 7600 1 19 2 0 79 02 19 2 0 82 03
0
98
1000
2000
9
1200
19
# of Institutions
Active Institutions
l a c i n i l C
Trials
1842 Participating Institutions
CALGB ECOG
NCI
NCCTG ACRIN
COG
GOG
ACOSOG
NSABP
RPC
SWOG
QARC
ATC
RTOG RTOG
ITC
Clinical
NCI
Trials
Components of a QA Program Remote audits of machine output 1,842 institutions, ~14,000 beams measured with TLD and OSLD in North America and Internationally Patient Treatment record reviews Review for GOG, NSABP, NCCTG, RTOG (brachy) On-site dosimetry reviews 30 institutions visited (~150 accelerators/450 beams measured) Credentialing - Phantoms > 500 Phantoms irradiations in 2010
RPC Verification of Institutions’ Delivery of Tumor Dose Reference calibration (NIST traceable)
Evaluated by RPC Dosimeters
Remote Audits • Verification of Reference Calibration for photon and electron beams • Units of special design (Gamma Knife, CyberKnife, Tomotherapy)
Remote Audits TLD: ThermoLuminescent Dosimetry OSLD: Optically Stimulated Luminescence Dosimetry
OSLDs TLD
TLD vs OSLD •
LiF:Mg,Ti (TLD-100)
•
(Al2O3:C)
•
Disposable
•
Reusable (dose limit ~ 10Gy)
• One reading
• Re-readable
•
Temperature and weight control • .
No temp/weight ctrl, light tightness
•
3 dosimeters per point
2 dosimeters per point,
• 6 min reading time
•
• ~ 2 min reading time
• Dosimeter cost per check $2.40 • Dosimeter cost per check $1.00
Remote Beam Output Audits (OSLD)
928 institutions monitored in 2010 (10,707 beams)
130 institutions (14%) had repeats
OSLD audit has been efficient and cost effective BEAM CHECKS 6/1/10-1/10/11
X-ray
OSLD
TLD
1200 Photon Electron 1000
0.995 (1.6%) 0.998 (1.7%)
Electron 0.996 (2%) 6 mths
0.998 (1.9%) 5 yr
800 BEAM COUNT
Modality
600
400
200
0 0.91 0.92 0.93 0.94 0.95 0.96 0.97 0.98 0.99
1
1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 1.1
OSLD/INST
Need for Dose Audits Are there errors in beam calibration? Figure 1
Visited Institutions contribute ~85% of clinical trial patients
35% Percent of Institutions with TLD Outside Criteria
30% 25% 20%
VISITED NON-VISITED
15% 10% 5% 0% 2000 2001 2002 2003 2004 2005 2006 2007 2008 Year
RPC Verification of Institutions’ Delivery of Tumor Dose Reference calibration (NIST traceable) Correction Factors: Field size & shape Depth of target Transmission factors Treatment time
Evaluated by RPC Dosimeters
Evaluated by RPC visits and chart review
On-Site Dosimetry Review Visit The only completely independent comprehensive radiotherapy quality audit in the USA and Canada •
Identify errors in dosimetry and QA and suggest improvements.
•
Collect and verify dosimetry data for chart review.
•
Improve quality of patient care. 15
15
Dosimetry Review Visit Study of 1474 institutions participating in clinical trials correlating RPC visit with patient accrual visited Institutions: 771 Prioritization
not visited yet
838 schema Patient accrual:focuses 20,878our visit 3,313 resources (85%)where the (15%) majority of the patients are treated!
On-Site Dosimetry Review Audit (overview)
Select an institution with a high prioritization score
Pre-travel preparation (NIST traceable)
RPC physicist travels to the institution:
Measurements, interviews, documentation review
Exit Interview with Radiation Oncologist and Medical Physicist
Recommendations for resolving dosimetry discrepancies and implementation of proper QA techniques given during visit and in a report before leaving the institution
On-Site Dosimetry Review Audit BEAM CALIBRATION RPC Onsite Visits
Percent within 3% Criterion
100%
95%
90%
Photon Electron
Reference Beam Calibration 85%
Percent of Inst. with ≥ 1TG-21 beam out of Criteria (since 2000) Implementation 80%
Photons
TLD (±5%) 7-11%
Electrons
TG-51 Implementation
6-12%
75% Visits1975 (±3%)1980 ~13% 1985 ~15% 1990
1995 YEAR
2000
2005
2010
On-Site Dosimetry Review Audit Discrepancies Discovered (Jan. ’05 – Mar. ’10) Number of Institutions Receiving rec. (n = 156) 115 (74%)
Discrepancies Regarding: Review QA Program Photon Field Size Dependence (FSD) Wedge Factor (WF) Off-axis Factors (OAF)/Beam symmetry Electron Calibration Photon Depth Dose Electron Depth Dose Photon Calibration Review Temp/Press Correction Change to TG-51 Electron Cone Ratios Using Multiple Sets of Data
Monitor Units =
62 (40%) 50 (32%) 46 (29%) 27 (17%) 25 (16%) 18 (12%) 13 (8%) 11 (7%) 9 (6%) 8 (5%) 8 (5%)
Prescription Dose
(calibration) • (FSD) • (WF) • (depth dose) • (OAF)
Purpose of Treatment Record Review Maintain a low uncertainty in doses delivered to protocol patients by discovering and correcting errors Provide study groups with accurate dose data
Improve Clinical Trials
RPC Treatment Record Review
Independent calculation of tumor dose Agree within 5% for external beam (15% for brachytherapy) Verify dose, time, fractionation per protocol Notify institution if major deviation seen during review to prevent further deviations
Treatment Record Review Process Review of treatment records & forms for accuracy
Independent dose recalculation
Resolve errors by contacting institution
Discuss results with Group and Study Chair
Facilitate clinical review at Group meetings, RPC, HQ
Errors > 5% (>15%) Found in Dose Review • 1% Systematic errors • 8% Individual errors • 27% Dose Reporting errors
Actual patient dose error Transcription error
Without RPC review 36% of the reported doses used by the study group would be incorrect
RPC Verification of Institutions’ Delivery of Tumor Dose Reference calibration (NIST traceable) Correction Factors: Field size & shape Depth of target Transmission factors Treatment time
Tumor Dose
Evaluated by RPC Dosimeters
Evaluated by RPC visits and chart review
Evaluated by RPC phantoms
Purpose of Credentialing • Educate • Improve understanding of protocol • Evaluate ability to deliver dose • Improve treatment delivery
• Reduce the deviation rate
Non-Credentialed Protocols (closed studies) Study
Disease Site
GOG 0122 Endometrial (1992) NSABP B14 Breast (1981) RTOG 9001 Cervical (1990) RTOG 9003 Head & (1991) Neck
Major Deviations
Minor Deviations
Number of Patients
29 (15%)
37 (19%)
197
135 (9%)
214 (15%)
1460
43 (14%)
76 (24%)
315
75 (7%)
188 (18%)
1073
Results of Credentialing Study
Disease Site
Major Deviations
Minor Deviations
Number of Patients
COMS
Eye
43 (5%)
47 (3.6%)
1166
GOG 165
Cervix
0
15 (21%)
70
RTOG 95-17 HDR & LDR
Breast
0
4 (4%)
100
RTOG 0019 LDR
Prostate
0
6 (5%)
117
3 (0.3%) 0 0
121 (10%) 42 (12%) 8 (9%)
1171 348 93
NSABP B39/RTOG 0413 3D MammoSite MultiCatheter
Breast
Use of Advanced Technologies in clinical trials? TRACKING
IMRT
MV R O KV
G N I GAT S
T R B
Re spi rat Co ory ntr ol
IGRT
TPS IMR T D O SE H ET PAI ER O NTI NG COR REC TIO N
Imaging, Planning and Delivery QA required at each step Positioning and Immobilization
Image Acquisition (CT, MR)
Structure Segmentation
treatment planning
Black Box File transfer and management
Plan validation
Position verification
treatment delivery
What You See Is NOT What You Always Get
RPC Phantom family 10 prostate phantoms (Photon IMRT/Protons)
8 Spine phantoms
31 H&N phantoms (IMRT)
14 thorax phantoms (Photon SBRT/IMRT) Proton head phantom and Protons
RPC Phantom Design • Anthropomorphic shape
Phantom
Patient
• Water filled • Plastic inserts containing targets and organs at risk (heterogeneity) • Point dose (TLD) and planar (radiochromic film) dosimeters • Purpose is to evaluate the complete treatment process from imaging to planning to setup to delivery
Patient
Phantom
Number of Phantom Mailings
400 Spine SRS Head
300
Liver Prostate Lung
200
H&N
100 0
2 0 0 1 2 0 0 2 2 0 0 3 2 0 04 2 0 05 2 0 0 6 2 0 0 7 2 0 0 8 2 0 0 9 2 0 10
Phantoms Mailed
500
Year
Plan vs. Treatment
Good Agreement Primary PTV
Secondary PTV
OAR
Measurement
Plan
Examples of Failures
Measurement vs. Monte Carlo Criteria 3%/2 mm 0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2.0
L
700 680 660
Dose DoseValue Value
640 640 620 620 600 600
L
R
580 580
LATERAL
560 560 540 540 520 520 0 0
Calc Calc Meas Meas 10 10
20 20
30 30
40 50 60 40 50 60 Distance along line Distance along line
70 70
80 80
90 90
P Axial
Varian 6 MV IMRT H&N 36
Good Agreement PTV OAR Cord/bone
Measurement
Plan
Bad Agreement
PTV
OAR Cord/bone
RPC Lung phantom
Convolution R-L Profile Right Left Profile Axial plane Left
Righ 30
Average displacement Left side: on: 3 mm off: 1 mm Prescribed D
Average displacement Rigth side: on: 1 mm off: 5 mm
Dose (Gy)
20
D2cm 10
PTV 0 -7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
Distance (cm)
RPC Film
ITC off
ITC on
RPC Regression
Institution Regression
7
Pencil-Beam Profile Right Left Profile Axial plane 30
Left Average displacement Left side: on: -2 mm off: -6 mm
Right Average displacement Right side: on: -5 mm off: -2 mm
Prescribed D
Dose (Gy)
20
D2cm 10
PTV
0 -7
-6
-5
-4
RPC Film
-3
-2
ITC off
-1
0
Distance (cm)
ITC on
1
2
RPC Regression
3
4
5
Institution Regression
6
7
CyberKnife ray-tracing Algorithm • Initial results were unsatisfactory • TLD measured doses in the center of the PTV to be ~13% low compared to calculated values. • Profiles were correct shape, but wrong absolute dose.
GTV PTV
Film 1 Film 2 Film 3 Film Avg TPS +5%/3mm -5%/3mm
11
Right-Left Profile Axial plane Left
CyberKnife Monte Carlo
Right
Average displacement Left side: Prescribed D 1 mm
Average displacement Right side: 2 mm
8
D o s e (G y )
6
D2cm
4
2
PTV
•
Average TLD RPC/Inst = 0.985
0 -7
-6
-5
-4
-3
-2
-1
0
Distance (cm)
1
2
3
4
5
6
7
Lung Phantom TLD results TLD results All algorithms 1.06
Average = 0.966 (sd 2.6%) 231 irradiations ( no outliers included)
1.04
RPC/Inst
1.02 1.00 0.98 0.96 0.94 0.92 0.90
0
50
100
Irradiation number
150
200
Lung Phantom TLD results TLD results (MC shown)
1.06
MC Average = 0.996 (sd 2.5%) 28 irradiations ( no outliers included)
1.04
RPC/Inst
1.02 1.00 0.98 0.96 0.94 0.92 0.90
0
50
100
150
Irradiation number
200
Phantom Results Comparison between institution’s plan and delivered dose. Phantom
H&N
Prostate
Spine
Lung
Irradiations
961
234
61
337
Pass Pass Fail Fail Criteria
686 (79%) 162 (82%) 22 (63%) 178 (75%) Static Motion 765 (80%) 187 196
196 (84%) 36 (59%) 268 (80%) 35 59 56 (84%) Pass 13221 (79%) 38 25 69 7%/4mm 7%/4mm Fail 5%/3mm 605%/5mm 13
Decrease RTOG Inst.the criteria to 5%/3mm for all phantoms and 440 (52%) 132 (15%) 36 (4%) 216 (25%) Acceptable Number of failures DOUBLES
>80% Pass Rate RPC H&N and Lung Phantoms
Percent of Passing Irradiations
100
80
H&N
60
lung 40
20 2003
2004
2005
2006
2007
Year
2008
2009
2010
Proton therapy facilities
Presently 9 clinically-active US proton therapy centers wanting to participating in clinical trials
new centers expected to open this coming year
Several trials in development to allow protons
NCI provided funds for a consultant to refine visit procedures and lead visits to 5 centers during 2010 and 4 more in 2011
Proton Facility Approval
Proposed RTOG/NCI Guidelines require: Questionnaire - Completed by 8 sites
TLD monitoring Mailed to 9 US centers + 1 Japanese center
On-site dosimetry review visits 3 visits (MDACC, UF and MGH) completed 2 visits (Penn, IN) completed, reports being generated
Anthropomorphic phantoms
Irradiated by 5 sites (3 passed, 1 failed, 1 under analysis)
Modifications to existing phantoms for other protocols
Proton Beam Monitoring Most measurements at mid-SOBP ± 2 cm
Phantoms HU RSP Calibration Curve
Pelvis phantom was modified2
1.8 Irradiated at 5 facilities 1.6
Lung phantom modified RSP
Irradiated at 2 facilities
head phantom developed
Tested at MDACC
1.4 1.2 1 0.8 0.6 0.4 0.2 0
-1500
-1000
-500
0
500
1000
1500
HU Eclipse PBT-poly B100 EBT2 Presage
Acrylic HI Poly C552 Solid Water
Polyethy/Wax PVC D400 Waxy*
Nylon Water Blue Water Balsa
Polyethylene A150 Epolene Cork
2000
Lung Phantom
Brain Phantom
Selected Phantom Lab “Alderson” phantom
Materials fall on CT#-RLSP curve
Contains realistic bony anatomy
Inserts with target and dosimetry will be constructed
•MR image to define CTV •MR image fused with CT for prescription and dose calc.
Phantom Accomplishments 1.
2.
3. 4.
Setting a standard for IMRT use in clinical trials Use of heterogeneity corrections for modern algorithms Test ability to hit a moving target(s) Testing proton therapy planning and dose calculations (in development)
The RPC, with its established infrastructure and experience, is a unique resource to the clinical trial setting as well as to the Radiotherapy community. We are confident that RPC QA audits through detection, education and correction have maintained if not reduced the uncertainty in the doses delivered to trial patients and improved the safe delivery of radiation doses to many other patients.
Systematic Errors >5%(>15%) Problem
Magnitude
TPS used wrong depth when head frame used
27%
Lung correction used, not allowed on protocol
9 – 13%
TPS wedge factor differs from clinical wedge factor
9%
TPS did heterogeneity corrections incorrectly
8.5%
Non-measured output with average TLD > 5%
7%
Point of calculation on edge of field
7%
Institution didn’t used effective field size when >50% of the field was blocked
5%
Individual Errors >5%(>15%) Problem Incorrect prescription points on brachytherapy Reported dose rates rather than dose for brachytherapy
Magnitude Up to 553% Up to 480%
Magnification error on brachytherapy Combined with incorrect prescription point
144% 208%
Hand written daily treatment record differed from Record and Verify for one field Brachytherapy shielding error Dose reported under block for parametrial boost
145%
Inhomogeneity corrections used (not allowed on protocol)
5 - 7%
23% 21%