Medical J. Conway, J. Slayton,

Burton

Robert

MS MS

#{149} John #{149} Orhan

L. McCrohan, H. Suleiman,

MS PhD

#{149} Robert

G. Antonsen,

BS

#{149} Fred

G. Rueter,

Physics

DSc

Average Radiation Dose in Standard CT Examinations ofthe Head: Results ofthe 1990 NEXT Survey’ In 1990, as part of the Nationwide Evaluation of X-ray Trends (NEXT) program, 252 computed tomographic (CT) systems were evaluated to measure

radiation

standard tiple-scan used

doses

associated

with

head CT in adults. The average dose (MSAD)

as the

dose

descriptor.

For

mulwas most

of the systems, the MSAD at the midpoint on the central axis of a standard dosimetry phantom was between 34 and 55 mGy. Doses were as high as 140 mGy, and dose sometimes varied by a factor of two or more for identical CT units. This range indicates that dose can potentially be reduced by careful selection of standard CT techniques. Users of CT systems should be aware of radiation dose delivered with CT, dose ranges associated with different systems, and doses delivered with their particular unit, which requires that dose performance of CT systems be assessed by means of a protocol that allows cornpanison of data collected for identical and/or different units. Index terms: Computed tomography (CT), radiation exposure #{149} Head, CT, 10.1211 #{149} Radiations, exposure to patients and personnel Radiology

I

From

Health,

1992;

the

Center

of the

ECAUSE

potential

for high

radiation doses in computed tomography (CT), it is important for users of CT systems to know the doses they typically deliver during

routine CT procedures. be aware, in particular,

Users should of the range

doses with colleagues

techniques who use

and

of CT system (i). as a part of the Nationwide

model In 1990,

Evaluation

of X-ray

Trends

vided

data

from

domly selected ties on record

Although

a similar

ran-

of 250 sys-

tems 1985

in 26 states was conducted (2), those data were derived

from

voluntary

dose state

measurements surveys rather

reports

dom national our knowledge,

sample the

for Device

and

of Health

Radiological

and Human

Services, Public Health Service, Food and Drug Administration, CDRH HFZ-240, 5600 Fishers Lane, Rockville, MD 20857. Received October 29, 1991; revision requested December 16; revision received February 14, 1992; accepted March 5. Address reprint requests to B.J.C. The mention of commercial products, their sources, or use in connection with material reported herein is not to be construed as either actual or implied endorsement of such products by the Department of Health and Human Services. 0 RSNA, 1992

of the

MATERIALS

AND

The

CT facilities from

reports United

the

16-cm-diammethacrylate

To sur-

made

by using no.

with

chamber

axis

of rotation

no.

was

placed

and

carefully

test

so that

Radcal).

length

patient

scanning

of the

plane

was

ensure

phantom

gan-

at right

proper

with

scan

checked

supCT

the longitudinal To

a test was

head in the

plane

bisected

of 100

of 3 mm3. the phan-

centered

phantom.

plane, scan

in the

the

to and of the

scanning axis had

ionization

10X5-10.3;

an effective volume dose measurements,

port

moni-

Monrovia,

has a sensitive

tom

ning

x-ray

Radcal,

a pencil-shaped

The chamber and During

an MDH

1015;

(model

alignment

of

to the

was

to verify

bisected

of the phantom and a minimum of pitch

the

obtained.

scan-

The

that

the

the longitudinal that and

the yaw

phantom with

respect to the central axis of the gantry. If the phantom was properly aligned, then the image of the alignment rod pattern, for

section

thicknesses

greater

than

8 mm,

would look similar to the diagram in Figure 2. Once alignment was verified, the ionization chamber was inserted into the hole along the central axis of the phantom for subsequent exposure measurements.

were

assemblers’

of CT system installations in the States. The procedure used for the

sample

selection

tinely

used

sentative

was

the

same

in the NEXT

been previously (3). The sample

program

reported by was selected

of all CT facilities

States

and

tative

of any

type,

and

used was polymethyl

parallel

tor (model

axis

METHODS

of x-ray

were

angles

head.

to be surveyed

a list

each

try,

during a ran-

Sample selected

50

and 12.7 mm in diameter, were drilled into the phantom. Four of these holes, one in each quadrant of the phantom, were drilled 10 mm from the surface and contamed alignment rods (Fig 1). The fifth hole, located along the longitudinal axis of the phantom, was used for insertion of the dosimeter. The CT dose measurements

mm

CT

of facilities. 1990 NEXT

CT examination

Characteristics

in

of surface obtained than from

Phantom Alignment

Cahif)

all U.S. CT facilitime of the study.

study

surveyors).

holes,

(NEXT)

252 facilities

individual

radia-

(PMM) head phantom described in the Code of Federal Regulations (4). Five

of the head was states pro-

from at the

by CDRH-trained state personnel (approximately

The phantom eter, 15-cm-long

program (a cooperative program of the Conference of Radiation Control Program Directors and the Center for Device and Radiological Health [CDRH]), a survey of a representative national sample of facilities performing CT examinations conducted. Forty-eight

conducted tion control

of

employed by the same make

vey represents the first nationwide U.S. study to use a random sample facilities in analysis of CT dose and technique information for standard

184:135-140

U.S. Department

B

make,

was

not

or model.

has

Conway et al to be repre-

to be

state All

rou-

and

in the United

chosen

individual

as that

represen-

or scanner

surveys

were

Abbreviations: CDRH = Center for Device and Radiological Health, CTDI = CT dose index, MSAD = multiple-scan average dose, NEXT = Nationwide Evaluation of X-ray Trends, PMM = polymethyl methacrylate, T/S = ratio of section thickness to section spacing.

135

I.

____________

178a,

45#{149} I

0

:

I

IL

1. cm-diameter

Table 1 Average Dose

i 1111m

iii lii

In

CGR

to the central axis of the rod and at 45#{176} angles to each other, to a depth of 0.61 cm. The rod is 1.26 cm in diameter. The center hole of the sequence is located 7.6 cm from one end (the

right)

included

Scanner Model

Manufacturer*

Alignment insert rod. Nine 0.14holes are drilled perpendicular

Figure

for CT Systems

in the NEXT

Elscint

of the rod.

3

8800

3

Typical alignment rod the phantom is properly

the scanning

plane

(for section

greater of the

than 8 mm). “crow’s feet” central axis.

lar to the

This

procedure

ployed

pattern aligned

differs

from

varies

from

can

occur

from

model to model or from manufacturer to manufacturer, depending on the amount and angular location of the regions above and below the scanner. These problems were avoided by using an axial measurement position for the dosimeter.

4 29

49±11 49±11

Siemens

CK DR HQ

3 3 3

1 18 3

30 34±13 57±9

FL

3

3

45±10

Technicare

1440 2020 2060

4

5

68±8

4

1

37

4 3

8 2

68±16 62±24

3

3

62±14

3

1

3

4

4

1

TCT8OA

Average

MSAD

values

may

not represent

typical

of Radiation

estimate scan

dose

profile

ple-scan

in the

average

dose

value

it is most

thickness

single-section scan

dose

dose profile

along

phantom cedure

subjected (obtainment

#{149} Radiology

associated

profile). describes

of dose

136

value

the longitudinal

The the

with

a

multipledistribution

axis of a

to a complete CT proof a series of scans).

cen-

(MSAD).

dose

to section than

profile

over

degree

For

of ripple the

surveyed

is

in the procedure). This can be done only when the shape of the single-section dose profile has been determined (eg, with a thermoluminescent dosimeter array). There is no numerical correction that can be applied to derive an accurate estimate of the dose when the shape of the singlesection

dose

profile

is not

was the case in this cd-shaped ionization

profile,

sively

left

than

Under

such

known.

This

study, in which chamber was

sponding

dose 1), and

will become

there

of the

there profile the

dose section

progres-

as a dose

descriptor.

is the practical

problem

integration

to the gaps

than 1, estimathat certain redose profile be (regions

left between

corre-

sections

when

than

is other

(since the CTDI 14-section-width

work,

degree

value of the multiple-scan in the region of the central

when the T/S is less of the CTDI requires of the single-section out

conditions,

below

less meaningful

a T/S

linearly

increasing

decreases

In addition, that, tion gions

less

For

used

gration

between sec1, the CTDI is not

in the multiple-scan T/S

(T)

are over-

increases

of overlap

a 1/S

(5).

sections

and the dose the

thickness

spacing

1, the

of the procedure,

profile in the region of the of such a series (rather than

to

by the ratio of section

well defined.

(as

appro-

used

of the

by the section

ple-scan central

some

of systems

a penused.

descriptor

that

and multiplying

average

on

if the number

ing

priate to describe radiation doses from CT in terms of the average value of the multi-

to focus

42

The MSAD is also based on the integral of a single-section dose profile and is the

will be a progressively

dose section

50 70±26

an interval of 14 section widths centered on the peak of the profile and then divid-

doses from CT (5-9). Since routine cedures typically involve obtainment scans,

been

of the multiple-

region

a single-section

grating

with

CT proof a

have value

The CTDI, discussed in detail by Shope et al (9) and defined in the Code of Federal Regulations (4), is computed by inte-

lapped,

Dose

parameters

the average

tions.

of individual

performance

83±40

CGRM,

A number of different parameters have been used in the literature to describe

series

35±10

2

greater

Characterization

38

20

study

to procealso

33±10

1

em-

reported

procedure

variations

2

3

4 4

tral section of a CT procedure. These are the CT dose index (CTDI) and the multi-

Dose

3

Columbia, Md; Elscint, Boston; GE Medical Systems, Milwaukee; Philips Medical Systems, Shelton, Conn; Picker International, Highland Heights, Ohio; Siemens Medical Systems, Iselin, NJ; Technicare, Cleveland; Toshiba, Tustin, Calif. t Numbers represent third-generation (rotate-rotate) and fourth-generation (rotate-stationary) scanners. t MSAD values are the average center dose for a typical CT examination of the head in an adult. SD = standard deviation.

measurements at the top surface can change from exposure to exwith some units, as the angular of the over or under scanning

dure.

40 52±22

3

because

region

1 6

600 1200

Two related

posure position

47±17

3 3

3

on CT dose(s), in which the chamber was placed in the top “surface” hole (2). The axial position was chosen for this survey location

35±9

94

LX

*

thick-

that

42±25

Picker

small.

Notice that the are perpendicu-

in the previously

38±3

9

3

MSAD* ±1 SD 28 31 23 28

24

3

TCT400 TCT500 TCT600 TCT900

nesses images

= 252)

Average (mGy)

1 1 1 1 4

3 3 3 3

9000 9800 9800HiLt CT-M\P CT-T 310-350 60

Toshiba

with

(n

CT Survey

No. Surveyed

Generation

1200 10000 1800 2002 2400

GE

Philips

Figure 2. seen when

1990

mm

by

the length

on doses

a pencil-shaped

ber to integrate

widths

is legally restricted to a integration). In this

we are reporting

using

of inte-

14 section

measured

ionization

cham-

the single-section

dose

The integration length is equal to the active length of the ionization chamber, which in this case was 100 mm. The MSAD measured in this fashion will be virtually identical to the CTDI when the section thickness is approximately 7 mm. For larger section thicknesses, the 100-mm integration length of the chamber will be profile.

less ing

than MSAD

a small tion

14 section widths, will underestimate

amount

thicknesses

tion thicknesses chamber will

(about

and

10%-l5%

of 10 mm

smaller integrate

the the

for sec-

or so).

than

over

resultCTDI by

For

7 mm, more

sec-

the than

July

14

1992

Table

2

Standard

for CT Systems

Parameters

10 or More Units

with

in 1990

NEXT

Survey

N Scanner

Parameter

GE

GE

8800(n=24)

9800(n=94)

120

kVp

U

Philips

Picker

60(n=20)

Siemens DR(n=16)*

1200(n=29)

120

120(90)

130

5 (5)

5 (1)

(mm)

spacing

10(23)

1(23)

T/Sratio

in parentheses

4(2) 8(12) 10(1)

1(92)

1

0

S yS

2(1)

the number

represent

e

1(14)

1

1.25(1) 2(1)

2(1)

Note-Numbers cable. *Section spadng

5 (1) 8(8) 10(20)

10

8(1) 10(88)

b

120(1) 125(15)

140(4) Section

m

of CT systems

for each

value.

e m

NA = not appli-

S

and T/S ratio were reported

Table 3 Average MSAD NEXT Survey

for 15 units.

19

and

Values

mAs

for CT Systems

10 or More

with

Units

Manufacturer

Average (mGy)

Model

GE

MSAD

MSAD/

(mGy/100

±1 SD*

Figure

8800

42±25

9800

47±17

Philips

60

(3-16)

59

69

DR

in parentheses represent are the average center

15±5

333±78

in air and in PMM

(8-27) 8 ± 3

(160-451) 442±26

energy

to 100 mM

of

(4-15)

(410-500)

CT examination

of the head

ionization

per section.

where

which

chamber were

=

was 100 mm for the we used (CT chambers

calibrated

configuration

C

photon

the energy correction factor; where L = the length of the ionization chamber

response effective

in an adult.

for an effective

of 70 keV);

used

SD = standard deviation. t Values have been scaled

distribution

(200-800)

in millimeters,

ranges. dose for a typical

99

4

shows

348±77 (198-528)

34 ± 13 (19-66)

89

(mGy)

3

Histogram

10±2 (7-13)

(23-70) Siemens

3.

79

MSAD

fourth-generation CT systems. The average MSAD value for the former is 44 mGy ± 19 (4.4 rad ± 1.9); for the latter, the average MSAD value is 54 mGy ± 14 (5.4 rad ± 1.4).

385±105

(6-27)

(23-61) 49±11

±1 SD

586±152 (317-768)

12±3

35±10

1200

Picker

mAs

7±3

(20-140)

values

49

Head

central MSAD values for the typical CT examination of the head, for the third- and

Average

mM)

±1 SDt

(17-119)

MSAD

39

Center

100mM

*

29

in 1990

Generation

Average

Note.-Values

20

in the

as that

same

used

general

during

the

MSAD measurements, ie, the entire length of the chamber was exposed, which results in a whole-chamber reading); where E = section timate

widths, and the CTDI by

approach

twice

that

tion thicknesses

the

MSAD

will overes-

technician)

that can

voltage

an amount

for the smallest

sec-

included

(kilovolt

liamperes),

(ie, I mm).

the

values

peaks),

scanning

for tube

tube current

time

(or

(mil-

milliampere

Since CT procedures consist of obtainment of a series of sections, the CTDI and

seconds), and when available, the beam filter selection normally used by the facility for a routine head procedure. Prese-

the MSAD

lected

relevant

are the dose

descriptors

to the clinical

They

both

dose

delivered

most

use of CT systems.

give a simple during

estimate the

chamber

(rather

than

section

of the

entire

a thermolumines-

In

pulsed

and

The

facility

the head,

spacing

CT technique

set to duplicate

the

of sections,

section

for

and

those routine

three

CT

rou-

scans

factors

normally

used

scanning

of

were

then

reading in milliroentgens factor of 2x was applied

chamber

has

standard

chamber

and where Because

ob-

We chose

region.

a 14-section-width

Detailed

integration

estimates

gans

or of other

more

complex

sorbed

dose

risk

thus,

of doses

estimates

descriptions

to or-

require of the ab-

1,

for a T/S

from

values,

(2).

equal

the average

=

of the three

the following

by using MSAD

to or greater

(1.C

.

L . E/T)

.

PMM

exposure

equation (‘/5

ratio)

Technique information only for routine CT head cial procedures This information

Volume

184

was collected procedures (spe-

were explicitly excluded). (obtained from the CT

#{149} Number

1

wheref

=

0.0078

convert

exposure

mGy/mR,

(a combination

dose-in-air

conversion

energy

factor

in air to absorbed

in PMM

of the mass

the

to

dose

of the exposure factor

absorption

and

the

to ratio

coefficients

the

spacing

all measurements

were the

of PMM, dose

in

made

in

MSAD

to the

plastic

in water

at 70 keV).

the manufac-

nostic

X-ray

doses

(CTDIs)

with

does

not

PMM. The proximately

mal brain

of the in

[or tissue] are required

Performance

eter

(the ratio coefficients

not to do so, since Standard

Systems

a volume alter

the

by the

for Diag-

(4) to specify

in terms

their user information. a correction in effect

Procedures

section

of CT systems

Federal

= ff’CLE/S)mGy, Survey

and

turers

(2):

of 6 mm3);

a correction factor energy absorption

mass

than

and

not to water or tissue. The dose or tissue can be obtained by ap-

culated

example);

of 3 mm3 a volume

nominal

=

absorbed

material, to water

not provide

for

has

phantom

represents

plying

array,

S

a uniform

the dose descriptor we employed is the MSAD, since the chamber will in general

dosimeter

a volume

millimeters.

tamed. The dosimeter reading was recorded during each scanning procedure. The MSAD value in milligrays was cal-

cent

(a corto the

meter reading for the ionization chamber when it was used with the x-ray monitor we used, to correct for the fact that the CT

were also recorded, as were of pulses and pulse length for

systems.

were by

for the number

thickness,

tinely used the number

proce-

dure to the region of the central section. practice, measurements in the field are made most easily by using an ionization

values

exposure rection

of dose

CT

to PMM

Additionally,

in

such

“ replaces” the dosimof tissue or water, but

fact

that

the

phantom

is

physical density of PMM is ap20% greater than that of nor-

tissue

and is of a somewhat Radiology

dif#{149} 137

40 5

7

N

N

U4

U

m b

m

6

3O

in

::::

b e

:.:

b

e

e

r

r 5

3,

0

0

0

::#{149}

f

:::>:

f 4.

20

S

SI

y

y

S

S S

S

3

0

m2, ?2

S

:::

2

::;i

e m10

t e

:::::

m

.::

S

S

#{149}:.:!:#{149}:#{149}

::

,‘

6,

#{163}

:i