Shielding Design for PET Clinics

1
Shielding Design for PET Clinics

• Robert L. Metzger, Ph.D.

2
Introduction

• Positron Emission Tomography is enjoying
  explosive growth due to its ability to stage and
  track cancer lesions.
• Patients receive 0.55 GBq (15 mCi) of 18F
  labeled 2-Fluoro-2-Deoxy-D-Glucose (FDG) and rest
  in a Quiet Room for 45 minutes to allow the FDG
  to localize in the lesions.

3
A Typical PET Scan

• After voiding, the patients are then scanned for
  approximately 45 minutes in the PET scanner, or
  more commonly now, a combined PET/CT unit.
• The PET/CT unit overlays the PET image on the CT
  image to create a precise registration of the
  tumor location.

4
PET/CT Study
5
Malignant Melanoma
6
Malignant Melanoma
7
PET Clinic Layout
Patient Bathroom
Quiet Rooms
Hot Lab
PET/CT Control
PET/CT
8
PET Clinic Layout

• This clinic layout utilizes distance rather than
  shielding to protect the technologist. The quiet
  areas are gt20 feet distant from the tech at the
  PET/CT control.

9
Example Clinic Layout
10
Shielding Problems in PET

• There was no published shielding guide for PET
  clinics. The AAPM Guide (handout) just
  published.
• PET started as a research tool with low patient
  volumes and tightly controlled research
  environments. Little shielding was necessary.
• PET has now exploded into clinical practice where
  patient volumes are high, and facilities
  frequently crammed into existing spaces.

11
Shielding Problems in PET

• The 511 keV photons can easily penetrate
  shielding used for conventional Nuclear Medicine
  (140 keV) and diagnostic X-Ray (lt120 kVp).
• PET patients can fog film stored in dry laser
  printers and dark rooms. Loaded cassettes are
  quite vulnerable.

12
Shielding Problems in PET

• NCRP 49 has attenuation coefficients for 511 keV
  photons, but they are narrow beam coefficients
  and PET is a broad beam environment. There are
  no buildup factors provided in NCRP 49. This has
  led to undershielding of some facilities (see
  Shielding Factors on pg. 6 of the handout).

13
Shielding Problems in PET

• The hot lab L block, syringe shields, and
  shadow shields for 511 keV photons are much more
  costly than those available for 140 keV Tc-99m.
  Some facilities have substituted Tc syringe
  shields and L blocks to save money.
• These problems have resulted in doses gt5 rem in
  one month for a technologist and gt 2 rem per year
  for adjacent tenants in some facilities.
• These doses have also resulted in a regulatory
  crack-down in some states.

14
AAPM Shielding Guide

• Provides conservative shielding estimates for
  clinics using Monte Carlo derived broad beam
  shielding curves for lead, concrete, and iron.
• Considers patient self attenuation, decay of 18F
  throughout the scanning process.
• Uses conservative usage for the quiet rooms and
  scanner.
• Uses NCRP 147 occupancy factors for surrounding
  areas.

15
Design Dose Limits

• Effective Dose Equivalent in Unrestricted areas
  must not exceed 1 mSv/yr (100 mrem/year).
• The occupational limit is 50 mSv/yr, but AAPM
  uses a design guide of 5 mSv/yr (500 mrem/yr)
  based on ALARA and pregnant worker limits.
• NCRP 147 uses the same criteria.

16
Source Term

• AAPM Draft Guide uses a patient dose rate of
  0.092 ?Sv m2/MBq hr or 5.1 mrem per hour at 1
  meter for a freshly dosed patient with 15 mCi of
  FDG (pg. 8).
• Represents about 65 of the point source gamma
  constant and is consistent with anterior patient
  measurements.

17
Scanner Source Term

• After the patient rests for approximately 45
  minutes, they void, and are then scanned. The
  radioactive decay for one hour resting is 0.74
  and the patient voids 15 of the dose, and the
  reduction during the scan is 0.91, leaving about
  ½ of the delivered dose.
• AAPM uses mean doses to come to the same number
  (pg 10).
• 30 of the dose is in the patients head with the
  remainder distributed in the body.

18
Scanner Attenuation

• AAPM uses a 15 reduction in dose due to scanner
  attenuation, but does not use this factor in the
  example problems.
• AAPM Guide was written before PET/CT became
  prevalent.
• Attenuation provided by the PET/CT double gantry
  is significant.
• Use manufacturers data for scanner attenuation.

19
Usage

• AAPM Guide assumes 100 occupancy of the quiet
  room with a dosed patient.
• It does not consider multiple quiet rooms that
  are now common.
• If multiple rooms are used, the occupancy cannot
  be 100.
• Scanners can scan a patient every 30 to 45
  minutes, so, at maximum, the scanner cannot do
  more than 16 patients per 8 hour shift.
• Practically, 10 is the maximum given patient
  setup times.

20
Usage

• Quiet rooms can produce no more than one patient
  every 45 minutes or 10.7 (call it 10) per shift.
• When multiple rooms are in use usage factors of
  0.65 are common.
• Scanners are considered to be continuously
  occupied.

21
Occupancy Factors

• Occupancy factors for surrounding areas are drawn
  from NCRP 147 (not NCRP 49).
• NCRP 147 values are more realistic.
• Caution must be used when choosing the 1/40th
  occupancy factor.

22
Draft Guide Limitations

• The guide does not manage layered shields that
  typically comprise floor and ceiling shielding.
  Treating each layer individually and summing the
  attenuation causes overshielding as 511 keV
  photons are assumed to be incident on each layer.
• Does not discuss the hot lab much.

23
Shielding

• The 511 keV photons from 18F and the mobile
  nature of the source (patient) create some unique
  shielding design problems for a PET clinic.
• New clinics are commonly sandwiched into existing
  imaging centers that are densely populated.
  Areas above and below the clinics are routinely
  occupied by other offices.

24
PET Clinic Shielding

• Inadequate structural shielding in some
  facilities has led to high doses to
  non-occupationally exposed personnel both within
  the facility and adjacent to it.
• Improper hot lab shielding (L block and syringe
  shields for 99mTc) has led to high doses to the
  Nuclear Medicine technologists.

25
Hot Lab Shielding
26
NCRP 49

• Dont use it!
• NCRP 49 HVLs are narrow beam attenuation values,
  while the PET patient represents a broad beam
  condition.
• Buildup in concrete is high at 511 keV.

27
Attenuation Curves(from AAPM Draft Guide on PET
Shielding)
28
Attenuation Curves
Monte Carlo Simulation (Broad Parallel Beam)
Constant TVL 17.6 cm

29
Wall Shielding

• Wall shielding is commonly required for the hot
  lab, quiet rooms, and scanner room.
• Many designs use distance rather than shielding
  for the interior spaces as technologists dislike
  closing off their patients in the quiet rooms.
  Doors, when provided, are rarely closed.

30
Wall Shielding

• Wall shielding can be easily calculated using
  point kernel techniques with buildup factors or
  from the AAPM Draft Guide on Pet Shielding
  Design.
• Source terms and occupancy factors may be taken
  from the draft guide or from actual experience.

31
Wall Shielding

• The height of the wall shielding is
  controversial. Some references say the shielding
  should extend to the floor above rather than the
  typical 7 foot height.
• Not practical. AC, electrical, call button, CCTV
  hardware, intercom, etc. run in the interstitial
  space above the false ceiling.
• Streaming is not significant at 450 KeV.

32
Example 1 Wall Shielding

• See Page 9-10, Example 1 Quiet Room.
• 15 mCi FDG, 40 pts/wk (one shift), uptake time
  one hour, 4 m to fully occupied uncontrolled area
  (T1)
• Weekly dose is 105.9 ?Sv (Eq 3).
• Limit is 20 ?Sv, therefore required attenuation
  is 0.189.

33
Determining Pb Thickness

• Draft Guide Table 4, 1.2 cm lead.
• Point Kernel I/I0Be-?x
• ? 1.79 cm-1 for Pb
• B 1.35 at 1.2 cm
• I/I0 0.16 Good Agreement
• NCRP49 No Buildup Off by about 30.

34
Ceiling and Floor Shielding

• When occupied areas exist above and/or below the
  quiet rooms and the scanning room, it is
  sometimes necessary to add sheet lead to the
  concrete deck. The floor thickness alone may not
  be sufficient to meet the non-occupational limit
  of 1 mSv (100 mrem) per year (25 mrem in some
  European countries).

35
Detector Locations

• The Draft Guide recommends that the dose limits
  be applied at 0.5 meter above the floor above
  (height of a low chair), and 1.7 meters (5.6 ft)
  above the floor below.
• That is, low chair above, and tall standing
  person below.
• Very conservative, even unrealistic.

36
Example 4

• See Page 11 - 12.
• Quiet room, 15 mCi FDG, 40 pts per week, uptake
  time 1 hr, 4.3 m floor to floor clear height, 10
  cm concrete deck, uncontrolled area above, T1.
• D (4.3 1) 0.5 3.8 meters.
• Eq. 3 117 ?Sv (one week)

37
Determining Concrete x

• Draft Guide Table 4 17 cm concrete.
• Point Kernel I/I0Be-?x
• ? 0.204 cm-1 for Concrete
• B 7.5 at 17 cm
• I/I0 0.23 Reasonable Agreement
• NCRP49 No Buildup Gross overestimate of
  shielding provided by concrete deck.

38
Example 4

• Guide calls for 0.65 cm (slightly more than ¼
  inch) of lead to be applied to the ceiling above.
• Sums attenuation provided by layered lead and
  concrete.
• Approach commonly leads to ½ inch lead
  requirements.
• Inaccurate. Better approach described later.

39
Ceiling and Floor Shielding

• The ceiling and floor shield consists of lead
  suspended under the support trusses of the
  concrete deck, forming a layered shield.
• Much of the shielding cost of a PET clinic is
  driven by these layered shields as they
  frequently require structural reinforcement to
  support the weight.

40
Imaging Room

• Correctly assumes about ½ of the delivered dose
  remains in patient.
• Assumes a 15 reduction in dose by scanner
  gantry, but does not use this in example
  calculations.
• The guide was developed before the advent of
  PET/CT units with multi-slice CT scanners.
• Actual dual gantries of the modern PET/CT units
  provide substantial attenuation.

41
Example 2

• See page 11.
• 15 mCi FDG original dose, 1 hr uptake, imaging
  time 30 min, 40 pts per week, 3 m to uncontrolled
  area with occupancy of 1.0.
• Calculates 59.7 ?Sv and calls for 0.8 cm Pb (1/2
  inch Pb practically).
• Very conservative.

42
AAPM Guide Assumptions

• Occupancy of the quiet room(s) with a dosed
  patient is 1.0. This is physically impossible.
  The patient must have a blood sugar test, be made
  comfortable, have an IV started, have the
  procedure described, and only then given the
  dose.

43
AAPM Guide Assumptions

• Single Quiet Room. Virtually all facilities now
  have 2 to 3 quiet rooms. The objective is to
  keep the scanner scanning. With the resting
  period of 45 minutes and a scanning time of 30
  minutes for new units, multiple quiet rooms are
  necessary for efficient utilization of the
  scanner.

44
AAPM Guide Assumptions

• 511 KeV. The guide correctly considers patient
  self-attenuation. But, since the
  self-attenuation is due to Compton Scattering,
  the energy is also reduced. The shielding curves
  are for 511 KeV photons incident on a shield wall
  under broad beam conditions. MCNP models
  indicate the actual energy is 350 450 keV
  incident on the shielding.

45
AAPM Guide Assumptions

• Limited to No Scanner Attenuation. The double
  gantries of the modern PET/CT units provide
  significant attenuation to the sides, top, and
  bottom. The guide describes an attenuation
  factor of 15, but then does not use it for the
  example calculation of the wall shielding.

46
CT

• Where distances are large (e.g. control booth),
  the shielding needed for the CT may dominate.
  Common for new PET/CT rooms as the space
  requirements are large for the dual gantry
  scanners.
• Use NCRP 147 to calculate CT shielding
  requirements.
• Look at both PET and CT shielding requirements
  and pick.

47
Testing

• NCRP 147 and many regulatory agencies ask for
  tests to ensure that the erected shielding is
  adequate.
• May be performed with pressurized ion chambers
  (e.g. Victoreen 451P) or large volume ion
  chambers (e.g. 1800 cc Radcal chamber)
• Alternately, monitor badges may be used (very
  cheap).

48
Testing
LLD is an issue. 100 mrad/year is 0.05 mrad/hour
for full occupancy. The survey instrument must
be able to accurately measure at 0.05 mR/hr or
integrate to achieve this LLD.
49
Monitor Badges

• Place away from areas frequented by patients.
• Factor for area occupancy and shift change (if
  any).
• May stop after compliance is determined.
• Some regulatory agencies have been erecting their
  own monitor badges where shielding is suspect.

50
Ceiling and Floor Shielding

• Point Kernel methods that calculate the
  attenuation provided by each layer and then sum
  them to obtain the total attenuation, tend to
  overestimate the shielding requirements when the
  layers are thick (in mfp). They assume 511 keV
  photons are incident on each layer.
• This can dramatically increase the shielding
  cost, particularly when the second course of lead
  requires structural reinforcement of the ceiling.

51
Monte Carlo Model of the Quiet Room

• We developed an MCNP model of the PET quiet room
  consisting of a MIRD phantom in a reclining
  position centered in a quiet room where the room
  and ceiling dimensions are taken from facility
  plans.
• The 0.55 GBq (15 mCi) of 18F was equally split
  between the bladder and brain in the phantom.

52
MCNP Model
53
Mercurad Model

• Mercurad is a deterministic code developed by CEA
  specifically for layered shielding problems.
• We developed a second model for this code using
  the same room materials and dimensions but a
  water sphere for the source term as human
  phantoms have not been ported to this code.

54
MCNP Model of Scanner

• A second model was developed for the scanner room
  using a MIRD phantom in a double gantry of a
  PET/CT scanner.
• The phantom was loaded with 7.5 mCi of FDG. 30
  of the dose was in the phantoms head with the
  remainder in the body.
• The scanner was developed from data provided by
  GE Healthcare for their current PET/CT unit.

55
MCNP Scanner Model
56
Mercurad Scanner Model

• A second model of the scanner room was developed
  for Mercurad using water filled spheres and
  cylinders to simulate the patient in the scanner.
• The source term was 7.5 mCi, with 30 in the head
  and the rest distributed uniformly in the body.
• Arrays of detectors were placed above and below
  the scanner room.

57
Mercurad Scanner Model
58
Example

• A clinic design was chosen where the point kernel
  method indicated that two layers of ¼ inch lead
  would be needed to protect the office above the
  quiet rooms.
• The actual room dimensions and concrete floor
  deck thickness was used and a volume detector was
  set 61 cm (chair height) above the second floor
  deck. The actual floor deck was corrugated, but
  only the thinner section was used in the
  calculation.

59
Scanner Example

• A scanner room with a fully occupied OT clinic
  above and a cafeteria below was used to test the
  scanner model.
• This room had a 4 inch ceiling and six inch
  concrete floor deck.
• The areas above and below were uncontrolled and
  fully occupied.

60
MCNP Runs

• Moritz visualization was used to ensure that all
  of the source points were within the brain and
  bladder and that the volumes were adequately
  sampled.
• After verification, 5 x 107 photons were run and
  the volume detector above the quiet room
  converged and passed all statistical tests.

61
Mercurad Run

• A point detector set to read in exposure rate
  was set at the same location for the Mercurad
  model. An array of detectors were used for the
  scanner room as the maximum exposure is difficult
  to predict given the complex geometry.
• The code converged in less than a minute.
• Results are expected to be higher than MCNP for
  the quiet room as no patient self attenuation was
  considered in this model (source term was a ten
  cm sphere).

62
Moritz Source Verification
63
Scanner Room Results
64
Scanner Room Results
65
Mercurad Scanner Results
66
Testing

• Once construction was complete, the exposure
  rates in the adjacent areas were measured before
  the facility went into full operation.
• Exposure rates on the second floor were measured
  with a Radcal 1800 cc ion chamber.

67
Measurements

• Actual exposure rates were measured with an 1800
  cc ion chamber while the rooms were occupied with
  dosed patients.

68
Results (mR/wk)
Location MCNP Mercurad Observed
Office Above QR 1 1.96 2.28 0.29
Office Above QR 2 1.96 2.28 0.86
69
Discussion

• The actual measurements were lower than predicted
  by either code.
• The floor trusses, lead shielding support,
  electrical , plumbing, and lighting fixtures in
  the false ceiling all provided some attenuation
  and reduced the actual dose above the rooms.

70
Discussion

• The Mercurad model (as expected) predicted
  slightly higher doses than the MCNP model as
  patient self attenuation was not considered.
• Mercurad results should be factored for patient
  self attenuation.
• Mercurad model development and execution was
  remarkably easy.

71
Conclusion

• The AAPM Draft Guide provides a conservative and
  simple shielding guide for PET clinics.
• The guide overshields clinics with multiple quiet
  rooms, layered ceiling and floor shielding, and
  scanner rooms where modern PET/CT units are
  employed.