Shani G. Radiation Dosimetry: Instrumentation and Methods

Подождите немного. Документ загружается.

Ch-04.fm Page 300 Friday, November 10, 2000 12:01 PM

301

Film Dosimetry

CONTENTS

I. Introduction............................................................................................................................................................301

II. Examples of Photographic Film Dosimetry..........................................................................................................301

III. Radiochromic Film Dosimetry ..............................................................................................................................310

References .......................................................................................................................................................................329

I. INTRODUCTION

Film dosimetry is attractive due to its high spatial resolu-

tion, wide accessibility, and the ﬂexibility to place the ﬁlm

in humanoid phantoms. Also, the short measuring time and

the fact that the ﬁlm dosimetry is intrinsically two-dimen-

sional and integrating in time are appreciated. Film is

potentially the ideal detector for determining dose distribu-

tion for dynamic beams and for studying the combination

of stationary beams treated sequentially. Film dosimetry is

widely used to obtain the relative dose distribution of elec-

tron and photon beams in water, in plastics, and in inho-

mogeneous phantom. Film dosimetry in phantoms is

advantageous because of high spatial resolution, short treat-

ment unit immobilization time, and 2D information.

Modern ﬁlm processing units improve the reproduc-

ibility and reliability of ﬁlm dosimetry and make it an

attractive method for many applications. Fast ﬁlm digitiz-

ers connected to a computer, equipped with proper eval-

uation software, allow rapid and accurate analysis of large

ﬁlms in a short time.

The most common setup in relative dose measurements

with ﬁlms is to sandwich a ﬁlm within a phantom of water-

equivalent material with the ﬁlm plane-parallel to the cen-

tral axis of the radiation ﬁeld. With the parallel geometry,

two particular precautions must be taken: there must be

heavy pressure on the phantom to avoid any air gap on

either side of the ﬁlm, and there must be perfect alignment

of the ﬁlm edge with the surface of the phantom. In addi-

tion, the artifacts that result from a thin air layer trapped

between the packaging material, the paper spacer, and the

ﬁlm are responsible for the inaccurate dose measurement

in the build-up region of the electron depth dose.

Conventional silver halide ﬁlm has a highly nonlinear

photon energy response, especially at low energies. It has

radiation interaction properties markedly different from

those of tissue. Along with variations introduced by the

necessary post irradiation processing, this type of ﬁlm is

extremely difﬁcult to use for accurate analytical dosimetry.

Most radiochromic systems are chemical radiation sen-

sors consisting of solid or liquid solutions of colorless

leuco dyes; these become colored without the need for deve-

lopment when exposed to ionizing radiation. Various radi-

ochromic forms, such as thin ﬁlms, thick ﬁlms and gels,

liquid solutions, and liquid-core waveguides, have been in

routine use for dosimetry of ionizing radiation over a wide

range of absorbed doses (10



to 10

Gy). Radiochromic

ﬁlm is used for general dosimetry of ionizing radiation in

high-gradient areas of electron and photon beams in a wide

energy range. The ﬁlm allowing approximately tissue-

equivalent dosimetry has been applied to mapping of dose

distribution in brachytherapy.

Radiochromic ﬁlm (RCF) is of great interest as a planar

dosimeter for radiation oncology applications. It consists

of a thin, radiosensitive, 7–23-



m thick, colorless leuco

dye bonded to a 100-



m-thick mylar base. RCF turns deep

blue in color upon irradiation. RCF is approximately tis-

sue-equivalent. GafChromic MD-55 is usable at doses

from less than 1 Gy up to 12 Gy when measured at the

wavelength of maximum sensitivity (676 nm) and up to

about 500 Gy when measured at a wavelength of low

sensitivity. The absorption spectra of GafChromic (GC)

ﬁlm contain two peaks with wavelengths in the range

610–680 nm and, thus, the dose-response curve as mea-

sured by an optical densitometer or spectrophotometer will

be highly dependent on the light source spectrum and sen-

sor material.

II. EXAMPLES OF PHOTOGRAPHIC

FILM DOSIMETRY

The use of ﬁlm as a dosimeter is still limited, due to the

various difﬁculties associated with ﬁlms such as energy

dependence, ﬁlm orientation, and sensitometric nonlinearity.

On the other hand, ﬁlm is probably one of the best detectors

for studying spatial distribution of dose or energy imparted.

The dosimetric resolution is limited only by the grain size

Ch-05.fm Page 301 Friday, November 10, 2000 12:01 PM

302

Radiation Dosimetry: Instrumentation and Methods

and the size of the aperture of an optical densitometer.

Commercially available laser-scanning densitometers

afford ﬁlm dosimetry with a resolution of a few microns.

Unlike other detectors such as ionization chambers, diode

detectors, thermoluminescent detectors (TLDs), scintilla-

tion detectors, and diamond detectors where dose informa-

tion must be recorded before the readout is cleared for

another irradiation, radiographic ﬁlms allow repetitive

readouts and provide a permanent record of the dosimet-

ric measurements. Films may be customized in various

sizes and shapes to ﬁt any dosimetric application. Due to

the relatively small thickness, a ﬁlm can be treated very

close to a Bragg-Gray cavity. The physical ﬂexibility of

a ﬁlm is also suitable for the curved and cylindrical sur-

face dose mapping when other detectors are impractical.

In relative dose measurements, the optical density may

be taken as proportional to the dose without any correc-

tion, since the collisional stopping power ratio of emul-

sion to water varies slowly with electron energy.

High-energy photon and electron dosimetry was car-

ried out by Cheng and Das using CEA ﬁlm. [1] The

packaging of the CEA ﬁlms is distinctly different from

that of the other ﬁlms, in that each ﬁlm is vacuum- sealed

in a shiny polyester-made, waterproof packet about 130



thick, permitting ﬁlm dosimetry to be carried out even in

water phantom. Aside from the difference in packaging,

the CEA ﬁlms have a clear polyester ﬁlm base as opposed

to a bluish-dye ﬁlm base found in the Kodak Readypack

XV and XTL ﬁlms. It is not clear if the difference in ﬁlm

base may have any effect on image quality. The CEA

ﬁlms come in two types: TLF (localization) and TVS

(veriﬁcation), similar to the XTL and XV ﬁlms of the

Kodak Readypack.

Figure 5.1 shows the variation of net optical density with

dose for the TVS ﬁlm for a range of photon energies from

137

Cs to 18 MV. The base optical density of the CEA ﬁlms

is on the order of 0.06, compared to about 0.2 for the Kodak

films. As shown in Figure 5.1, the linear portion of the char-

acteristic curve covers a range of optical density up to 4.3.

This wide range of linearity offers a convenient means in

percent depth dose and isodose measurements. It is inter-

esting to note that the ﬁlm is faster to rays than to

bremsstrahlung

x-rays from a linear accelerator but is inde-

pendent of energy for each type of radiation. This feature

is particularly attractive for high-energy x-rays, as one

single sensitometric curve can be used for a wide range of

x-ray energies. A straight line is ﬁtted to each of the

characteristic curves for rays and x-rays with a regression

coefﬁcient of 0.999. For gamma rays, the line of regres-

sion is [1]

(OD)



, rays



0.054 07



dose, (5.1)

while for x-rays, the line of regression is

(OD)

X, rays



0.047 65



dose (5.2)

Figure 5.2 compares the characteristic curves for the

CEA TVS ﬁlm and the Kodak XV ﬁlm for two x-ray

energies, 4 and 10 MV. For the CEA TVS ﬁlm, the linear

portion extends over the dose range up to 90 cGy, followed

by an abrupt saturation above 90 cGy. The CEA TVS

curves have a linear relationship, with a coefﬁcient of

regression very close to 1. For the Kodak XV ﬁlm, on the

other hand, the optical density varies curvilinearly with

FIGURE 5.1

Characteristic curves of the TVS ﬁlms for photon beams. (From Reference [1]. With permission.)

Ch-05.fm Page 302 Friday, November 10, 2000 12:01 PM

Film Dosimetry

303

dose up to 200 cGy. The characteristic curve of the Kodak

XV ﬁlms exhibits a longer tail and a smaller gamma

compared to that of the CEA TVS ﬁlm.

The size of the silver halide crystals is generally smaller

in the CEA TVS ﬁlm than in the Kodak XV ﬁlm. Indeed,

the silver halide crystals in the CEA TVS ﬁlm are of fairly

uniform size and shape, resulting in a linear characteristic

curve. On the other hand, the silver halide crystals of the

Kodak XV ﬁlm are of different sizes and shapes, with the

largest more than 10 times bigger than the smallest, result-

ing in a nonlinear characteristic curve. Furthermore, the

CEA TVS ﬁlm has a higher silver concentration (about 42



) compared to regular x-ray ﬁlms (about 7 g



The characteristic curves for the TLF ﬁlm over the

photon energy range 4-18 MV are shown in Figure 5.3.

The ﬁlm saturates at about 30 cGy, which is considerably

higher than the Kodak Readypack XTL ﬁlm. Unlike the

CEA TVS ﬁlm, the characteristic curves for the CEA TLF

ﬁlms are slightly curvilinear over the range of dose up to

30 cGy. However, if the data is considered only in the

range of 0–15 cGy, the curves are all straight lines with

a coefﬁcient of regression near unity. The CEA TLF ﬁlm

FIGURE 5.2

Comparison of the characteristic curves of CEA TVS ﬁlm and Kodak XV ﬁlm for different photon energies. (From

Reference [1]. With permission.)

FIGURE 5.3

Characteristic curves of the CEA TLF ﬁlm for photon energies 4–18 MV. For comparison, characteristic curve of

Kodak XTL ﬁlm is also shown for a 6-MV beam. (From Reference [1]. With permission.)

Ch-05.fm Page 303 Friday, November 10, 2000 12:01 PM

304

Radiation Dosimetry: Instrumentation and Methods

is also energy-independent within the experimental

uncertainty.

A method of ﬁlm dosimetry for high-energy photon

beams was proposed by Burch et al. [2] which reduced the

required ﬁlm calibration exposures to a set of ﬁlms obtained

for a small radiation ﬁeld size and shallow depth (6 cm



6 cm at 5-cm depth). It involves modiﬁcation of a compres-

sion-type polystyrene ﬁlm phantom to include thin lead

foils parallel to the vertical ﬁlm plane at approximately 1 cm

from both sides of the ﬁlm emulsion. The foils act as high

atomic number ﬁlters which remove low-energy Compton

scatter photons that otherwise would cause the ﬁlm sensi-

tivity to change with ﬁeld size and depth.

High-energy photon beams used in radiation oncology

are considered to interact primarily by the Compton scat-

tering process with tissue. However, when ﬁlm is placed

in a tissue-equivalent material, photoelectric interactions

associated with the silver atoms in the emulsion cause the

ﬁlm to over-respond relative to the tissue-equivalent mate-

rial. For the low-energy photons, ﬁlm dose may be as

much as 25 times the tissue dose at the same physical

location.

The results for the 6-cm



6-cm and 25-cm



25-cm

ﬁelds shown in Figure 5.4 emphasize the importance of

the ﬁlm sensitivity as a function of ﬁeld size and depth.

As the ﬁeld size was increased, the dose required to

produce a given density on the ﬁlm was reduced. For

sizes smaller than 10 cm



10 cm, the sensitivity change

is small (not shown); and for sizes less than 6 cm



6 cm,

no change in sensitivity occurs. For small ﬁelds with

less scatter, the change in sensitivity with depth is not

apparent.

The effects of foil-to-ﬁlm separation distance and foil

thickness were investigated in order to obtain a single

optimum distance-thickness combination, and the results

are presented in Figures 5.5 and 5.6. In Figures 5.5 and

5.6, the dose was calculated using data from the 6-cm



6-cm calibration ﬁlms. Film/foil separation distances of

0, 0.6, 1.2, and 1.9 cm and lead foil thicknesses of 0, 0.15,

0.30, 0.46, and 0.76 mm were included in the investiga-

tion. At 0-cm ﬁlm/foil separation distance, the curve

shows the effect of electrons coming from the lead due to

interactions within the foil. This intensiﬁcation effect

exaggerates the shape of the depth-dose curve and the non-

uniform ﬁlm/foil contact is apparent in the data (Figure 5.5).

At 0.6 cm the scattered electrons are absorbed in the inter-

vening polystyrene. Further increase in ﬁlm/foil separation

distance produced only minor changes in the calculated

dose curve. Figure 5.6 shows the effect of changing foil

thickness. A single thickness of 0.15 mm causes a signif-

icant decrease in the calculated dose, with only subtle

changes as additional layers are added. The best match to

ion chamber percentage depth-dose data was observed for

lead foil thickness of 0.46 mm with a 1.2 cm ﬁlm/foil

separation distance.

The use of radiographic ﬁlm for the dosimetry of large

(i.e.,



10 cm



10 cm) photon radiotherapy beams is

complicated, due to a dependence of ﬁlm emulsion sen-

sitivity on depth within the phantom. This dependence is

caused by a relative increase in depth in the population

of lower-energy, scattered photons in the spectrum and

the subsequent photoelectric absorption of these photons

by the ﬁlm emulsion. This effect becomes most pro-

nounced with increases in the photon population in the

FIGURE 5.4

Film sensitivity dependence upon ﬁeld size and upon depth. (From Reference [2]. With permission.)

Ch-05.fm Page 304 Friday, November 10, 2000 12:01 PM

Film Dosimetry

305

energy region below approximately 400 keV, where the

mass attenuation coefﬁcient of emulsion increases rapidly,

approximately as 1



. As the low-energy spectral

component increases, the tissue equivalence of ﬁlm emul-

sion diverges from that of tissue-equivalent materials, due

to an approximate dependence of the photoelectric

interaction cross section. [3]

In order to examine possible dependencies of emul-

sion exposed by radiosurgical beams, a series of sensito-

metric curves was established for a range of depths and

ﬁeld sizes. For each exposure a 10-in



12-in sheet of

Kodak X-Omat V ﬁlm from a single batch (#194 05 2)

was sealed within a light-tight Solid Water (Gammex

RMI, Inc.) cassette. The cassette consists of two 2-cm-

thick slabs of solid water sealed around three edges by

nylon screws and a rubber O-ring. [3]

To minimize possible variations due to ﬁlm-processing

conditions, a Kodak X-Omat RP processor was used by

Robar and Clark for which the throughput is very high

and quality assurance is performed daily. Developer tem-

perature ﬂuctuated by less than



0.5°F between pro-

cessing sessions. Optical density was measured using a

FIGURE 5.5

Lead foils placed adjacent to the ﬁlm show an exaggerated response at shallow depth and a wavy appearance due to

undulation in their surface. Each curve is normalized to 5-cm depth. (From Reference [2]. With permission.)

FIGURE 5.6

The effect of foil thickness on the shape of the depth-dose curve. (From Reference [2]. With permission.)

hv()

Ch-05.fm Page 305 Friday, November 10, 2000 12:01 PM

306

Radiation Dosimetry: Instrumentation and Methods

scanning densitometer. Base-plus-fog optical density was

subtracted from scanned optical densities for each ﬁlm.

In order to relate optical densities to absolute doses,

the dose to water corresponding to each of these depths

and for each MU setting was calculated from

(5.3)

where

(

) is the dose at depth

for ﬁeld size

at the

phantom surface,

is the number of monitor units

given,

PDD

(

) is the percent depth dose and

(

) is

the total scatter factor. Both

(

) and

PDD

(

) were

measured in a water phantom using a

-type silicon elec-

tron diode .

Figure 5.7a shows the sensitometric curves obtained

for the large (20-cm



20-cm) photon beam at depths of

1.0 cm,10.0 cm, and 20.0 cm. For each depth a curve was

ﬁtted to the measured sensitometric data using the single-

target/single hit equation

(5.4)

where

and

are the measured optical density and

given dose, respectively. The saturation density of the ﬁlm,

sat

, was estimated by delivering a large dose (500 cGy)

to a ﬁlm in phantom, and it was held constant in the ﬁtting

algorithm. The ﬁtting parameter



represents emulsion

sensitivity and was allowed to vary. For this large ﬁeld,

the curves diverge markedly, indicating that using a sen-

sitometric curve that is not depth-speciﬁc would introduce

signiﬁcant error in converting optical density to dose. In

order to minimize this error to below approximately 10%,

for example, it is necessary to conﬁne the dose range to

less than 80 cGy. In contrast, Figure 5.7b shows sensito-

metric curves corresponding to the same depths for the

radiosurgical (2.5-cm-diameter) ﬁeld. For this small ﬁeld,

the curves agree to within the reproducibility of ﬁlm devel-

opment and scanning. By determining the value of



for

each curve, ﬁlm sensitivity was obtained as a function of

depth, as illustrated for the 20-cm



20-cm and 2.5-cm-

diameter ﬁelds (Figure 5.8). As expected from the dispar-

ity in the sensitometric curves in Figure 5.7a, the ﬁlm

sensitivity for the large (20-cm



20-cm) ﬁeld increases

systematically with depth. While small ﬂuctuation of the

values of



is apparent for the radiosurgical ﬁeld, no

systematic variation of sensitivity with depth is apparent.

Film dosimetry is most problematic for lower-energy

photon beams (e.g.,

Co and 4-MV) due to their lower

primary-to-scatter ratio. Figure 5.7 shows the importance

of corrections in converting scanned optical density to dose

even for the 6-MV beam.

Film is often used for dose measurements in hetero-

geneous phantoms. In those situations perturbations are

produced which are related to the density and atomic

number of the phantom material and the physical size and

orientation of the dosimeter. Signiﬁcant differences were

observed by El-Khatib et al. between the dose measure-

ments within the inhomogeneity. [4] These differences

were inﬂuenced by the type and orientation of the dosim-

eter in addition to the properties of the heterogeneity.

To investigate how the differences in dose measure-

ments are related to the dosimeter conﬁguration, the EGS4

Monte Carlo system, together with the PRESTA algorithm,

was used to model the experimental conditions. The per-

centage depth doses measured in the phantom containing

1-cm bone are shown in Figure 5.9 and are compared to

Dd, A()

MUPDD d, A()S

A()

100

---------------------------------------------------



OD OD

sat

110





()

FIGURE 5.7

The sensitometric curves established experimen-

tally for (a) the 20-cm



20-cm radiotherapy ﬁeld and (b) the

2.5-cm-diameter radiosurgical ﬁeld, shown for depth in solid

water phantom of 1.0 cm, 10.0 cm, and 20.0 cm. (From Refer-

ence [3]. With permission.)

Ch-05.fm Page 306 Friday, November 10, 2000 12:01 PM

Film Dosimetry

307

the depth doses measured in the homogeneous water phan-

tom. All ﬁlm measurements were done with the ﬁlm in the

ready pack with the inner paper removed. The doses were

normalized to the maximum dose in the homogeneous

polystyrene phantom, measured with whatever dosimeter

was used to measure the doses at depth. The doses within

the bone represent dose-to-unit density material within the

bone rather than dose to the bone itself, since a correction

by ratio of collisional stopping powers was not done. At

several cm beyond the bone, the PDD measured with all

dosimeter conﬁgurations are identical. The opposite effect

for the two ﬁlm orientations is observed for the phantom

containing the cork (Figure 5.10). The PDD is unaffected

in the initial 4 cm of polystyrene and there is greater

penetration of the beam in cork. Within the cork the dif-

ference measured with the ﬁlm in different orientations is

as much as 6%, and the TLD measurements are lower than

the measurements with ﬁlm. These greater discrepancies

are attributed to the fact that the inhomogeneity is much

larger and lies in a region of high dose gradient.

A method for the creation of a complete 3D dose

distribution from measured data was discussed by

FIGURE 5.8

The curve-ﬁtting parameter



, representing emulsion sensitivity, as a function of depth for the 20-cm



20-cm

radiotherapy ﬁeld and the 2.5-cm-diameter radiosurgical ﬁeld. (From Reference [3]. With permission.)

FIGURE 5.9

The percentage depth doses measured in a polystyrene phantom containing 1-cm hard bone at 1-cm depth using

(A) ﬁlm in parallel orientation to the beam, (B) ﬁlm in perpendicular orientation to the beam, and (C) TLD are shown and compared

to the PDD measured in a homogeneous water phantom for electron beams of nominal energy (a) 12 MeV and (b) 20 MeV. (From

Reference [4]. With permission.)

Ch-05.fm Page 307 Friday, November 10, 2000 12:01 PM

308

Radiation Dosimetry: Instrumentation and Methods

Stern et al. [5] It involved the measurement with ﬁlm of

the dose distribution in a series of “beam’s-eye-view”

(BEV) planes (planes perpendicular to the beam central

axis at different depths). Film measurements were made

using Kodak Readypack XV-2 ﬁlm (Eastman Kodak Co.,

Rochester, NY) sandwiched between sheets of water-

equivalent solid. Exposed ﬁlms were digitized using a

laser digitizer with spot size set to 0.42 mm and pixel

size 0.45 mm.

Film image data ﬁles were next converted into the

planning system’s standard grayscale image ﬁle format.

When necessary, the number of pixels was reduced using

nearest-neighbor sampling. The images were entered into

the planning system and aligned with the planning system

representation of the beam by translating and rotating the

coordinate system of the displayed ﬁlm image. Optical

density values were converted to dose values using the

appropriate measured ﬁlm sensitometric curve.

The sensitometric curves obtained at four different

depths ranging from

max

to 20 cm are plotted vs. absolute

dose in Figure 5.11. The largest variation in dose for a

given optical density determined from this set of curves is

FIGURE 5.10

The percentage depth doses measured in a polystyrene phantom containing 6-cm cork at 4-cm depth using (A) ﬁlm in

parallel orientation to the beam, (B) ﬁlm in perpendicular orientation to the beam, and (C) TLD are shown and compared to the PDD

measured in a homogeneous water phantom for electron beams of nominal energy (a) 12 MeV and (b) 20 MeV. (From Reference [4].

With permission.)

FIGURE 5.11

Film sensitometric curves derived from four sets of BEV ﬁlms placed at depths of 3, 5, 10, and 20 cm, respectively.

(From Reference [5]. With permission.)

Ch-05.fm Page 308 Friday, November 10, 2000 12:01 PM

Film Dosimetry

309

2.3%, at an optical density of approximately 2.3. Variation

is less than 1% over most of the measured dose range.

This demonstrates that a single sensitometric curve mea-

sured at one depth can be accurately used for optical

density-to-dose conversion of BEV ﬁlms at all depths for

the experimental situation used.

A feasibility study for mailed ﬁlm dosimetry has been

performed by Novotny et al. [6] The fading of the latent

image before ﬁlm processing is only 3% per month and

the normalized sensitometric curve is not modiﬁed after

a period of 51 days between irradiation and processing.

All ﬁlm experiments have been performed with Kodak

Readypack paper envelope ﬁlms (25.4 cm



30.5 cm).

All ﬁlms were processed with an automatic processing

unit, either an Agfa Gevamatic 1100 (processing time 90 s

and temperature 35



1°C) or a modern AGFA Curix 260

(processing time 90 s and temperature 35



1°C). No

dependence on the photon energy has been observed in

the shape of the normalized sensitometric curve for

Co,

6-MV, and 18-MV beams (Figure 5.12).

No increase is observed in the fog value for un-

irradiated ﬁlms after mailing or storing for more than

75 days; the average reading is 0.16



0.01 for the two

series of ﬁlms, either processed immediately after irradi-

ation or processed after mailing.

Fading is expressed as the ratio of the optical density

of a faded ﬁlm and the optical density of the reference ﬁlm,

as a function of the time

between ﬁlm irradiation and ﬁlm

processing. The interval

is equal to 0 for the reference

ﬁlm. The dependence of fading on time

, for ﬁlms stored

in the laboratory, is presented in Figure 5.13. A straight

line is ﬁtted through the measured data using the least-

squares method. A decrease of about 3% in optical density

is observed per month of delay between ﬁlm irradiation

and ﬁlm processing (corresponding to a given dose of 0.45

Gy and optical density of about 1.3). A smaller decrease

of about 1.8% per month in optical density is observed for

ﬁlms irradiated with dose of 1.5 Gy (corresponding to

optical density 2.7). These decreases correspond to esti-

mated variations on dose of 3.8% and 3.4%, respectively.

FIGURE 5.12

Relative sensitometric curve for three photon beam qualities (

Co, 6-MV x-rays, 18-MV x-rays). (From Reference [6].

With permission.)

Ch-05.fm Page 309 Friday, November 10, 2000 12:01 PM