Therapy Shielding Calculations

Melissa C. Martin, M.S., FACR, FACMP

American College of Medical Physics
21st Annual Meeting & Workshops
Scottsdale, AZ
June 13, 2004
 Therapy Shielding Design Traditionally
Relies on NCRP Reports
NCRP Report 49
Primary and secondary barrier calculation methodology
Applicable up to 60Cobalt and linacs up to 10 MV

NCRP Report 51
Extended NCRP 49 methodology up to 100 MV
Empirical shielding requirements for maze doors

NCRP Report 79
Improved neutron shielding methodology

NCRP Report 144

Update of NCRP 51 primarily aimed at non-medical facilities
Reports reflect progress in linac design and shielding research
 Revised NCRP Report in Drafting Stage by
AAPM Task Group 57, NCRP SC 46-13
Design of Facilities for Medical Radiation Therapy
4 MV - 50 MV (including 60Co)

Calculation scheme generally follows NCRP 49

All shielding data (TVLs) reviewed and updated
Updated for intensity modulated radiation therapy (IMRT)
Improved accuracy of entrance requirements
Both with and without the use of maze

Laminated barriers for high energy x-rays

Photoneutron generation due to metal in primary barrier
Goal: Improved accuracy
 Linear Accelerator Energy and Workload

BJR #11 megavoltage (MV) definition used here

British Journal of Radiology (BJR) Supplement No. 11

Comparison of BJR #11 and BJR #17 MV definitions

BJR #11 MV 4 6 10 15 18 20 24
BJR #17 MV 4 6 10 16 23 25 30

Workload assumptions typically used for shielding design

Workload identified by symbol W in calculations
For MV 10 MV: W = 1000 Gy/wk at 1 meter from the target
Based on NCRP 49 Appendix C Table 2
For MV > 10: W = 500 Gy/wk
Based on NCRP 51 Appendix B Table 5
 Radiation Protection Limits for People

Structural shielding is designed to limit exposure to people

Exposure must not exceed a specific dose equivalent limit
Limiting exposure to unoccupied locations is not the goal

NCRP 116 design dose limit (P)

0.10 mSv/week for occupational exposure
0.02 mSv/week for the general public

Typical international design dose limits

0.12 mSv/week for controlled areas
0.004 mSv/week for uncontrolled areas

NCRP 116 dose limit is a factor of 5 lower than NCRP 49 value

 Radiation Protection Limits for Locations

Permissible dose outside vault depends on occupancy

Occupancy factor (T):
Fraction of time a particular location may be occupied
Maximum shielded dose (Smax) at protected location

S max = P
T
Assuming occupancy factor T for protected location

Maximum shielded dose is traditionally referred to simply as P/T

 Occupancy Values from NCRP 49

Full occupancy for controlled areas by convention (T=1)

Full occupancy uncontrolled areas (T=1)
Offices, laboratories, shops, wards, nurses stations, living
quarters, childrens play areas, and occupied space in nearby
buildings

Partial occupancy for uncontrolled areas (T=1/4)

Corridors, rest rooms, elevators with operators, unattended
parking lots

Occasional for uncontrolled areas (T=1/16)

Waiting rooms, toilets, stairways, unattended elevators, janitors
closets, outside areas used only for pedestrian or vehicular traffic
 Hourly Limit for Uncontrolled Areas

0.02 mSv hourly limit for uncontrolled areas

20 Gy/hr common assumption for calculation
Implies a lower limit for occupancy factor
T 20 / ( U W )
T 0.16 for higher energy accelerators (500 Gy / wk workload)
T 0.08 for lower energy accelerators (1000 Gy wk workload)

Not applied to low occupancy locations with no public

access
e.g., unoccupied roof, machinery room

T = 1/10 rather than 1/16 typically used for exterior walls

 NCRP 134 Impact on Linac Shielding

NCRP 134 distinguishes general employees from public

NCRP 134 maintains NCRP 116 limit of 0.02 mSv/wk for both
Limit 25% of 0.02 mSv/wk from individual facility for general public

Occupancy assumptions proposed for general public

T=1/40 for occasional occupancy

Equivalent to T=1/10 occasional for general employees

Similar to P/T required by hourly limit for primary barriers
Slightly increase from T = 1/16 used for secondary barriers
T=1/16 still appropriate for locations with no public occupancy
e.g., machine rooms, unoccupied roofs, etc.

Impact increases if higher occupancy than T=1/40 adopted

 Basic Primary Barrier Calculation
Unchanged from NCRP 49
Unshielded dose calculation

WU
S pri = 2
A'
A
d pri Door

Target
Attenuation in tenth-value layers Rotational
D' Plane

S pri D
Maze * Target
Isocenter B
n = log10
P /T dpri

Barrier thickness (tc) calculation tC C

C'
1 ft
tC = TVL1 + (n - 1) TVLe

Margin in primary barrier thickness is recommended to

compensate for potential concrete density variation
 Primary Barrier Photon Tenth-Value Layers
(mm) Come from a Variety of Sources
Lead Concrete Steel Earth Borated Poly
MV TVL1 TVLe TVL1 TVLe TVL1 TVLe TVL1 TVLe TVL1 TVLe
0.2 1.7 1.7 84 84 15 15 135 135 84 84
0.25 2.9 2.9 94 94 19 19 151 151 94 94
0.3 4.8 4.8 104 104 22 22 167 167 104 104
0.4 8.3 8.3 109 109 29 29 175 175 109 109
0.5 11.9 11.9 117 117 33 33 188 188 117 117
1 26 26 147 147 54 51 236 236 147 147
2 42 42 210 210 76 69 336 336 210 210
4 53 53 292 292 91 91 468 468 292 292
6 56 56 367 323 100 100 572 572 343 343
10 56 56 410 377 104 104 648 648 379 379
15 56 56 445 416 108 108 720 720 379 379
18 56 56 462 432 109 109 740 740 379 379
20 56 56 470 442 110 110 752 752 390 390
24 56 56 483 457 110 110 773 773 401 401
NCRP 49 NCRP 51 Nelson & LaRiviere McGinley Estimated from Concrete

Anticipate upcoming NCRP report to review and update TVL data

 Primary Barrier Width

0.3 meter margin on each side of beam rotated 45 degrees

Barrier width required assuming 40 cm x 40 cm field size
wC = 0.4 2 d C ' + 1.0 ft
Field typically not perfectly square (corners are clipped)
35 cm x 35 cm field size typically used to account for this

Target to
* Target
Isocenter Target to
* Target
Isocenter Target to
* Target
Isocenter
Narrow Point Narrow Point Narrow Point
Distance Distance Distance
(d C') (d C') (d C')

1 ft 1 ft 1 ft 1 ft

C' wC wC Metal
C C' C
wC
1 ft 1 ft
 Slant Factor and Obliquity Factor

Slant Factor
Path from target to protected location diagonally through barrier
Incident angle q of line with respect to perpendicular
Required barrier thickness reduced by cos(q)
Same total distance through barrier to protected location

Scatter causes slant factor to underestimate exit dose

Multiplying thickness by obliquity factor compensates for this
Lead Concrete Steel
Angle 4 MV 10 MV 18 MV 4 MV 10 MV 18 MV 4 MV 10 MV 18 MV
0 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
30 1.03 1.02 1.03 1.02 1.00 1.00 1.02 1.02 1.04
45 1.07 1.07 1.10 1.07 1.04 1.04 1.07 1.07 1.08
60 1.21 1.21 1.22 1.20 1.14 1.08 1.20 1.17 1.20
70 1.44 1.47 1.52 1.47 1.28 1.22 1.48 1.42 1.45
 Photoneutron Generation Due to Metal in
Primary Barrier (Linacs 10 MV)
Dose-equivalent 0.3 m beyond barrier (McGinley)
WU NF - t / TVL -t / TVL
SN = 10 1 P
10 3 N

t2
+ t + 0.305
3
2
N is neutron production constant (Sv neutron per Gy workload)
1.9 x 10-3 for lead, 1.7 x 10-4 for steel at 18 MV (from McGinley)
Recent safety survey indicated somewhat higher 3.8 x 10-4
value for steel at 18 MV is appropriate
N adjusted versus MV based on neutron leakage fraction vs MV
F is field size (conventionally 0.16 m2), t2 is metal thickness (m)
X-Ray attenuation prior to metal layer: 10^(-t1 / TVLp)
Neutron attenuation after metal layer: 10^(-t3 / TVLN)
 Patient Photonuclear Dose Due to Metal in
Primary Barrier for MV > 10
Metal in primary barrier can increase patient total body
dose if MV > 10
Lead inside layer approximately doubles patient total body dose
Increases risk of secondary cancer

Concrete or borated polyethylene inside metal in primary

barrier is recommended if MV >10
Each inch of borated poly decreases patient dose from metal
barrier photoneutron by approximately factor of 2

Impact of IMRT on patient photonuclear dose is addressed

later

Avoid metal as inside layer of primary barrier if MV > 10

 Secondary Barrier

Patient scatter unshielded dose

a W ( F / 400)
Sp = 2 2 A'
d sca d sec Door A

Target
F is field size in cm2 Rotational
D' Plane
typically 1600
D
Maze dsca * Target
Isocenter B
a = scatter fraction
dsec
for 20 x 20 cm beam
tB 1 ft
Leakage unshielded dose
C
Assumes 0.1% leakage fraction C'
-3
W 10
SL = 2
d sec
Leakage Photon Tenth-Value Layers (mm)
Also Come from a Variety of Sources

Lead Concrete Steel Earth Borated Poly

MV TVL1 TVLe TVL1 TVLe TVL1 TVLe TVL1 TVLe TVL1 TVLe
4 53 53 292 292 91 91 468 468 292 292
6 56 56 341 284 96 96 546 455 341 284
10 56 56 351 320 96 96 562 512 351 320
15 56 56 361 338 96 96 578 541 361 338
18 56 56 363 343 96 96 581 549 363 343
20 56 56 366 345 96 96 586 552 366 345
24 56 56 371 351 96 96 594 562 371 351

Kleck & Varian Estimated

NCRP 49 Nelson & LaRiviere
Average from Concrete
 Neutron Leakage

Same form as photon leakage calculation

Based on dose-equivalent neutron leakage fraction vs MV
0.002%, 0.04%, 0.10%, 0.15% and 0.20% for 10, 15, 18, 20 and 24 MV
Based on Varian and Siemens neutron leakage data
Assumes quality factor of 10 for absorbed dose

Shielded dose equivalent based on leakage neutron TVLs

211 mm for concrete
96 mm for borated polyethylene
 Intensity Modulated Radiation Therapy
(IMRT)
IMRT requires increased monitor units per cGy at isocenter
Typical IMRT ratio is 5 MU per cGy, as high as 10 for some systems

Percent workload with IMRT impacts shielding

50% typically assumed; 100% if vault is dedicated to IMRT

Account for IMRT by multiplying x-ray leakage by IMRT

factor
IMRT Factor = % IMRT x IMRT ratio + (1 - % IMRT)
3 is typical IMRT factor (50% workload with IMRT ratio of 5)

IMRT factor lower for neutrons if machine is dual energy

e.g., 1.5 if dual energy linac with 50% of treatments below 10 MV
Pessimistic since most IMRT is performed at 6 MV (next chart)
 IMRT above 10 MV Significantly Increases
Patient Photonuclear Dose
Neutrons dominate patient total body dose for high energy
linacs
Neutron dose equivalent as high as ten times photon dose
Potentially 1% of workload vs 0.1% photon leakage
0.05% required absorbed neutron dose x 20 quality factor
Typical neutron dose equivalent is lower than requirement
0.1 to 0.2% of workload

IMRT factor of 5 increases patient incidental dose 5X

Results in typical neutron total body exposure of 0.5 to 1.0% of WL
Significantly increases risk of secondary cancer

Most IMRT is performed at 6 MV to mitigate increased secondary

cancer risk from photoneutrons
 Patient Scatter Significant Adjacent to
Primary Barrier
Scatter traditionally neglected
for lateral barriers
A'
Generally a good assumption Door A

90 degree scatter has low Target

energy Rotational
D' Plane
Target
Scatter is significant adjacent to D
Maze dsca * Isocenter B
primary barrier
q Scatter
Angle
Calculations indicate Slant thickness dsec
comparable to leakage used to calculate
attenuation

Slant thickness through barrier C

C'
compensates for the increase in 1 ft
unshielded dose due to scatter Actual
barrier
thickness
Barrier thickness
comparable to lateral is
adequate for same P/T
 Patient Scatter Fraction for 400 cm2 Field

Based on recent simulation work by Taylor et.al.

Scatter fraction increases as angle decreases
Scatter fraction vs MV may increase or decrease
Tends to increase with MV at small scatter angles
Decreases with increasing MV at large scatter angles

Angle (degrees)
MV 10 20 30 45 60 90 135 150
4 1.04E-02 6.73E-03 2.77E-03 2.09E-03 1.24E-03 6.39E-04 4.50E-04 4.31E-04
6 1.04E-02 6.73E-03 2.77E-03 1.39E-03 8.24E-04 4.26E-04 3.00E-04 2.87E-04
10 1.66E-02 5.79E-03 3.18E-03 1.35E-03 7.46E-04 3.81E-04 3.02E-04 2.74E-04
15 1.51E-02 5.54E-03 2.77E-03 1.05E-03 5.45E-04 2.61E-04 1.91E-04 1.78E-04
18 1.42E-02 5.39E-03 2.53E-03 8.64E-04 4.24E-04 1.89E-04 1.24E-04 1.20E-04
20 1.52E-02 5.66E-03 2.59E-03 8.54E-04 4.13E-04 1.85E-04 1.23E-04 1.18E-04
24 1.73E-02 6.19E-03 2.71E-03 8.35E-04 3.91E-04 1.76E-04 1.21E-04 1.14E-04
 Patient Scatter Energy

Mean Scatter Energy

Scatter Angle (degrees)
MV 0 20 45 90
6 1.7 1.2 0.6 0.25
10 2.8 1.4 0.6 0.25
18 5.0 2.2 0.7 0.3
24 5.7 2.7 0.9 0.3

No standardized scatter Tenth-Value Layer

Primary MV rating based on peak MV in spectrum, not mean energy
Primary TVL at slightly higher MV (e.g, 50%) appears reasonable
% increase little more than wild guess; more research is needed

Ambiguity remains as to TVL to use for scatter

 Maze Calculation Likely Revised in
Upcoming NCRP Report
New method identifies and evaluates specific mechanisms
Patient Scatter, Wall Scatter, Leakage scatter
Direct leakage
Neutrons, capture gammas

Mechanisms calculated at most stressing orientation

Scatter calculations multiplied by 2/3 to compensate for this

Scatter energy relatively low at maze door

Primary 0.3 MV TVLs used for patient and wall scatter (2 bounces)
Primary 0.5 MV TVLs used for leakage scatter (1 bounce)
Scatter is significant typically only for low energy linacs

Goal: More-precise calculation avoiding over or under-shielding

 Maze: Patient Scatter

Unshielded dose

a W ( F / 400) a 0.5 AC A'

= Door
A
Sp
d P21 d P2 2 d P2 3 Target
Rotational
D' Plane
Target
where D dP3
dP1 * Isocenter B
a0.5 is 0.5 MV scatter fraction
dP2
Second bounce fraction
0.02 per m2 typically used
AC = wC h
C
Other constants as before, e.g., wC
a = patient scatter fraction
F = field size in cm^2
h = room height
 Maze: Wall Scatter

Unshielded dose

f W a 1 A1 a 0.5 AM A'
= Door A
SS
d S21 d S22 d S23 Target
Rotational
dS3 D' Plane

where D
* Target
Isocenter

f = patient transmission dS1

a1 = first reflection coefficient dS2
dM
0.005 per m2 for 6 MV
C
0.004 per m2 for 10 MV
A1 = beam area (m2) at wall
AM = Maze cross section (m2)
dM x room height
 Maze: Leakage Scatter

Unshielded dose

A'
Door A
W 10 3 a 1 AC
-

S LS = Target

d L21 d L22
Rotational
D' Plane

where D dL2 * Target

Isocenter B
Constants as previously dL1
defined
AC = wC h
C
wC
 Maze: Direct Leakage

Unshielded dose

A'
Door A
- - t D ' / TVL
W 10 3 10
SL = tD' q
Target
Rotational
d L2 dL Plane

Same as standard secondary D

D' * Target
Isocenter
B
photon leakage calculation
Standard neutron leakage not
typically used C
C'
Use only if it exceeds the maze
neutron calculation
e.g., if maze wall not
sufficiently thick
 Maze Neutron Calculation Based on
Modified Kersey Method
Unshielded dose equivalent

A'
W Ln Door A
H NT = 2 [1+ ( d N 2 -3) / 5 ]
d N 1 10 Target
Rotational
D' Plane

where D dN2 * Target

Isocenter B
Ln is neutron leakage fraction dN1

Same as used for secondary

neutron leakage calculation
C
Modification to Kersey is C'
assuming first tenth-value
distance is 3 m instead of 5 m

Upcoming NCRP report may recommend a more-complex

approach than this
 Maze Neutron Shielding

Modeled as 50% thermal neutrons and 50% fast neutrons

1 inch borated poly effectively eliminates all thermal
neutrons
Fast neutron TVL is 2.4 inches for the first 4 inches
Fast neutron TVL is 3.6 inches beyond 4 inches thickness
 Maze Capture Gammas from Concrete

Gamma rays generated by neutron capture in the maze

Very significant for high energy linacs

Unshielded dose is a factor of 0.2 to 0.5 of the neutron

dose equivalent at the treatment room door
Use the conservative factor (0.5)

Capture gammas have moderate energy (3.6 MeV)

TVL of 61 mm for lead
Limited attenuation also provided by polyethylene (278 mm TVL)

Dominates X-Ray dose at maze entrance for high energy linacs

 Direct-Shielded Door

Neutron Door is simply a secondary barrier

Typically more layers and different materials than a wall
Lead to attenuate leakage photons
Borated polyethylene to attenuate leakage neutrons
Typically sandwiched between layers of lead
Steel covers

Specialized shielding procedure adjacent to door

Compensates for relatively small slant thickness in this location
Vault entry toward isocenter similar to maze
Vault entry away from isocenter is secondary barrier
But with specialized geometry
 Direct-Shielded Door: Far Side of Entrance

Extra material added to corner

Lead to entrance wall
Protected Point
Borated polyethylene or (1 ft beyond
concrete beyond wall door enclosure)

Door Overlap
Uses standard secondary barrier Isocenter to Isocenter to Door Beyond Far Side
Far Side of
Secondary
of Entrance
calculation Entrance
Distance Distance 7.5"
Isocenter Overlap
Typical
Goal: provide same protection Target Typical
as wall or door for path through Rotational
Plane
Gap
0.5"
corner
 Direct-Shielded Door: Near Side of Entrance

Geometry similar to short maze

Maze calculation can be used
Typical
but is likely pessimistic Gap
0.5"
Requires less material than far 7.5"
side of entrance Typical
Door
Overlap
Lower unshielded dose dN1
Isocenter dN2
Lower energy Protected
Target
*
Target
Rotational
Point
(1 ft beyond
Plane door
enclosure)
 Shielding for Heating, Ventilation, and Air
Conditioning (HVAC) Ducts
HVAC penetration is located at ceiling level in the vault
For vaults with maze, typically located immediately above door
For direct-shielded doors, located in a lateral wall as far away from
isocenter as possible

Ducts shielded with material similar to the door at entrance

Material thickness 1/2 to 1/3 that required of the door
Path through material is at a very oblique angle due to penetration
location with slant factor between 2 and 3
Factor of at least 5 reduction in dose at head level (the protected
location) vs. at the HVAC duct opening

NCRP 49 recommends that shielding extend at least a

factor of three times the width of the HVAC penetration
 Photon Skyshine

Unshielded dose

0.0249 W U W1.3
S sky = W
d Y21 d Y22 2 meters

where
dY1
W (steradians) = 0.122
h
for 40 x 40 cm beam
Isocenter

Multiplying by additional factor

of two is recommended * Target

Floor
Primary TVLs used to calculate h
dY2
attenuation

New construction seldom shields solely for skyshine due to

vigilance required to prevent unauthorized roof access
 Neutron Skyshine

Unshielded dose
-4
5.4 10 H pri W
H sky =
2p W

where
W = 2.71 (steradians) typical * Target

(target above isocenter) Isocenter

Hpri is neutron dose-eq in beam

(0.00013, 0.002, 0.0039, 0.0043,
and 0.014 times W for 10, 15, 18, Floor
20, and 24 MV, respectively)
Up to 20 meters lateral distance
Use factor is not applied since
neutrons in all orientations
Multiplying by additional factor
of two is recommended
 Primary Goal of Upcoming NCRP Report is
Improved Shielding Calculation Accuracy
Very little impact for low energy accelerators
Primary and secondary barrier calculation method unchanged
Very little impact to calculated shielding for given protection limit

Improved accuracy for high-energy accelerators

Avoids extra cost of over design due to pessimistic calculations
Avoid extra cost of retrofitting if inaccurate calculations
underestimate required shielding
 References

Biggs, Peter J. Obliquity factors for 60Co and 4, 10, 18 MV

X rays for concrete, steel, and lead and angles of incidence
between 0 and 70, Health Physics. Vol. 70, No 4, 527-536,
1996.
British Journal of Radiology (BJR) Supplement No. 11.
Central axis depth dose data for use in radiotherapy, 1972.
Chibani, Omar and C.C. Ma. Photonuclear dose
calculations for high-energy beams from Siemens and
Varian linacs, Medical Physics, Vol 30, No. 8:1990-2000,
August 2003.
Kleck, J. Radiation therapy facility shielding design.
1998 AAPM Annual Meeting
 References (Continued)

McGinley, P.H. Shielding Techniques for Radiation

Oncology Facilities, 2nd ed. Madison, WI: Medical Physics
Publishing, 2002.
National Council on Radiation Protection and
Measurements. Structural shielding design and evaluation
for medical use of x-ray and gamma rays of energies up to
10 MeV. Washington, DC: NCRP, NCRP Report 49, 1976.
National Council on Radiation Protection and
Measurements. Radiation protection design guidelines for
0.1-100 MeV particle accelerator facilities. Washington, DC:
NCRP, NCRP Report 51, 1977.
 References (Continued)

National Council on Radiation Protection and

Measurements. Neutron Contamination from Medical
Accelerators. Bethesda, MD: NCRP, NCRP Report 79, 1984.
Nelson, W.R., and P.D. LaRiviere. Primary and leakage
radiation calculations at 6, 10, and 25 MeV, Health
Physics. Vol. 47, No. 6: 811-818, 1984.
Rodgers, James E. IMRT Shielding Symposium AAPM
Annual Meeting, 2001.
Shobe, J., J.E. Rodgers, and P.L. Taylor. Scattered
fractions of dose from 6, 10, 18, and 25 MV linear
accelerator X rays in radiotherapy facilities, Health
Physics, Vol. 76, No. 1, 27-35, 1999.
 References (Continued)

Taylor, P.L., J.E. Rodgers, and J. Shobe. Scatter fractions

from linear accelerators with x-ray energies from 6 to 24
MV," Medical Physics, Vol. 26, No. 8, 1442-46, 1999.