Rapid radiographic film calibration for IMRT verification using ...

Rapid radiographic film calibration for IMRT verification using automated 

MLC fields 

Nathan L. Childress, a) Lei Dong, and Isaac I. Rosen 

Department of Radiation Physics, U.T. M.D. Anderson Cancer Center, 1515 Holcombe Blvd., Box 94, 

Houston, Texas 77030 

Received 28 March 2002; accepted for publication 1 August 2002; published 30 September 2002 

A method for measuring a film sensitometric curve using a single sheet of film exposed with a two 

field step-and-shoot MLC treatment was developed and tested with Kodak XV2 and EDR2 films. 

With this technique a film sensitometric curve can be completed in only 10 minutes, making it 

practical to generate new film calibrations daily. This method is applicable to film calibrations for 

all purposes, but is particularly useful in IMRT treatment verification due to the method’s use of 

small fields. This method agrees with the traditional large-field multifilm calibration within 0.5% 

and will produce sensitometric curves with errors less than 1% throughout the dose range, including 

uncertainties in dose delivery, film response, and optical density measurements. OD values for XV2 

and EDR2 films were consistent in the middle of exposure areas at high depths, but the XV2 film 

penumbra regions showed large amounts of over-response as the calibration depth increased. If 

XV2 film is used for IMRT treatment verification, it is necessary to reduce the fluence of low 

energy photons in areas around the film by using thin lead shields. EDR2 film was shown to have 

minimal energy dependence, as it accurately represented penumbra areas and yielded identical 

sensitometric curves generated with 6 and 18 MV photons. However, its darker tint may make it 

more sensitive to scanning laser film digitizers’ horizontal nonuniformities. This single film method 

proved to be superior to the traditional calibration method and allows fast daily calibrations of films 

for highly accurate IMRT delivery verifications. © 2002 American Association of Physicists in 

Medicine. DOI: 10.1118/1.1509441 

Key words: film dosimetry, IMRT treatment plan verification, film digitization 

I. INTRODUCTION 

Film dosimetry can be used to quickly obtain a two dimensional 

dose distribution of a radiation field. The purpose of 

this research was to develop a simple method to quickly 

produce a calibration curve of film optical density OD versus 

dose exposure, commonly called a sensitometric curve. 

This was accomplished by using step-and-shoot multileaf 

collimator sMLC files to generate regions of varying doses 

on a single sheet of film. This method was tested with Kodak 

XV2 and EDR2 films, and compared with the standard 

method of delivering different doses to 1010 cm fields in 

the centers of individual XV2 films. It is now in use at U.T. 

M.D. Anderson Cancer Center to produce daily film calibrations 

for quantitative dose verifications of intensity modulated 

radiotherapy IMRT treatments. 

In a developed film, the light transmission through any 

point is inversely related to the amount of metallic silver, 

which is a direct but nonlinear function of dose deposited 

at the point. For radiation dosimetry, the film is scanned with 

a densitometer. The OD is measured at each point in the film 

and converted to dose using a sensitometric curve. 

While the process of film dosimetry appears to be simple, 

it has many sources of error. Table I summarizes the sources 

of uncertainty in film dosimetry. Their magnitudes can vary 

greatly among different machines and institutions. Analytical 

methods to combine these errors are unclear. Because of error 

and measurement uncertainties in film, it is common to 

find conflicting results in film research publications. For example, 

there has been a study that examined the wide variations 

of published research containing film OD measurements 

at different depths. 3 

The primary source of error in film dosimetry is the large 

amount of high-Z silver halides in films, resulting in overresponses 

to low energy photons primarily due to their increased 

photoelectric attenuation values when compared to 

biological tissue. Because small IMRT fields create many 

penumbra regions with low energy photons, IMRT film dosimetry 

can have many areas of film over-response. Traditional 

methods of dealing with film over-response, such as 

mathematical modeling, 4 are not applicable to multi-field 

treatments. For this reason, lateral scatter filtering 5 has been 

recently formulated. The purpose of this filtering is to remove 

low-energy scattered photons. The thin lead sheets 

preferentially attenuate low-energy photons, while allowing 

most higher-energy photons to be transmitted. It has been 

found that 0.15 mm thick sheets of lead placed parallel to the 

film at distances of 6–12 mm results in optimal filtering. 6 

While lateral scatter filtering was originally intended for 

films placed parallel to the photon beam axis, it may also be 

useful for use in nonparallel orientations. Even though the 

lead shields attenuate the primary beam, their low-energy 

photon absorption can result in an overall improvement in 

film response. 7 Most photon interactions are Compton 

events, which result in average electron energies of around 

2384 Med. Phys. 29 „10…, October 2002 0094-2405Õ2002Õ29„10…Õ2384Õ7Õ$19.00 © 2002 Am. Assoc. Phys. Med. 2384

2385 Childress, Dong, and Rosen: Rapid radiographic film calibration 2385 

TABLE I. Possible sources of error in IMRT absolute film dosimetry Refs. 

1–3. 

Error source 

Air gaps in either side of film may exist 

Spectral variations due to different phantom depths may 

exist 

The beam orientation may be changed throughout 

treatment verification, thus changing both 

depths and photon spectrum along the film 

Low-energy photons originating from penumbra region or 

edge of MLC treatment fields may 

cause film to overrespond 

Daily linac output variations exist 

Film processor temperature, solvent, and miscellaneous 

variations may exist 

The film response may be changed by storage and 

irradiation environmental conditions 

Differing radiation responses from different film batches 

may exist 

There are variations in film densitometer OD 

measurements 

The calibration field will measure a slightly different 

spectrum than the spectra that are 

present in larger or smaller treatment fields 

Avoidable? 

Yes 

No 

half the incident photon’s energy. With ranges of approximately 

5 mm per MeV, some of the secondary electrons created 

by high-energy photon interactions in lead can still expose 

the film. Lateral scatter filtering can cause film to 

slightly under-respond when compared to biological tissue. 

However, it improves results and can yield agreement to 

within 3% of ion chamber measurements in the penumbra 

region. 7 Results are similar for parallel and perpendicular 

film orientations, though the primary beam attenuation is a 

function of the angle of incidence. 

Film over-response can also be minimized by the use of 

high dose films. Kodak’s EDR2 film contains only 50% as 

many silver halide particles as their XV2 film, thus expanding 

its upper dose limit by a factor of four. It is also less 

sensitive to low doses in general. EDR2 film can handle 

doses large enough to simulate full patient treatments up to 

400 cGy, rather than requiring the halving of the patient 

plan’s monitor units MUs as with XV2 film. However, film 

sensitivity is sacrificed due to the more sparse disbursement 

of crystals. With fewer high-Z silver halides, the effective Z 

of EDR2 film as a whole is lower. The reduced effective Z 

lowers the photoelectric attenuation coefficient of the film 

for low photon energies, and the film responds to photons in 

a manner more similar to tissue. 

Due to the film’s high sensitivity to its environment during 

developing and scanning, it is essential that routine 

planned maintenance be performed on all film processors 

and densitometers. 8 During the simultaneous development of 

multiple films with low-duty processors, the temperature of 

solutions can change by a few degrees if not properly filled 

and maintained. Appropriate phantom selection is required 

when measuring the high photon energies found in linear 

accelerators, as most photon interactions that lead to film 

exposure will occur in the material surrounding the film. The 

No 

Yes 

No 

Yes 

Yes 

Yes 

No 

No 

proper compression in phantoms is also a necessity to minimize 

air packet formation between the film and its outer 

light–tight packaging. Reasonable error estimates due to 

variations in the film uniformity and response, processor, and 

densitometer would be 6–15% if no method was used to 

compensate for the standard film overresponse, or 3–8% otherwise. 

One solution to reduce many film errors is to create a 

calibration curve each day, rather than relying on old calibrations 

or attempting to scale a standard curve. This method 

minimizes errors due to film storage, exposure conditions, 

film developer, and scanner variations assuming all films are 

simultaneously processed. The daily production of a sensitometric 

curve requires a quick way to accurately generate 

different known exposures on film so that labor costs can be 

minimized and dosimetrists’ tasks can be focused on treatment 

verifications rather than film calibrations. 

The sMLC calibration method can be used to calibrate 

films for any purpose. However, it is especially accurate for 

IMRT treatment verifications due to the small field sizes used 

in both techniques. The beam’s energy spectrum varies with 

field size, and film’s OD response is dependent on the energy 

spectrum. Thus, it is beneficial to calibrate film using field 

sizes that are similar to the application in which film dosimetry 

will be used. This allows for a more accurate calibration, 

which is particularly important when using film as the sole 

method to verify complex IMRT treatment deliveries. While 

this calibration method could easily be used for non-IMRT 

treatment verifications, 2D confirmations of conventional 

treatments are not a standard practice. 

II. METHODS AND MATERIALS 

As shown in Fig. 1, step-and-shoot MLC leaf motion files 

were created to automatically deliver eight 33 cm squares 

of different doses spaced 4 cm apart vertically and 7 cm 

apart laterally to one sheet of film as a two-field treatment. 

Each MLC file corresponds to one side of the film exposure 

and requires asymmetric jaw settings. The y-axis jaws were 

set to have a symmetric 24.4 cm opening, while the x-axis 

inner jaw was placed 2 cm across the beam centerline and 

the outer jaw was pulled 7 cm from the beam centerline. This 

utilized two sMLC files that controlled leaf motions as functions 

of dose fractions. The dose is initially delivered to four 

open areas, then the MLC closes to deliver doses to fewer 

regions until the four areas are irradiated to different levels. 

The MLC leaves are parked 1 cm away from the jaw openings 

to ensure that minimal leakage from the MLC closing 

gaps is allowed to penetrate to the film. 

Ion chamber measurements were taken of the individual 

doses delivered to the eight regions of the film using EDR2 

MUs. The ion chamber used, a Wellhofer CC04 with a4mm 

diameter and 4 mm length, was calibrated on Machine 1 

Varian Clinac 2100EX with a Varian Millenium 120 leaf 

MLC to obtain a dose per charge value. A Wellhofer electrometer 

was used with frequent redefinition of its background 

compensation. Calibration films were placed at 95 

cm SSD, with 5 cm solid water attenuation and 3 cm solid 

Medical Physics, Vol. 29, No. 10, October 2002


depths of 2, 5, 11, and 19 cm while changing the MU delivered 

to closely match the doses obtained at 5 cm. These 

additional depths were generated by adding acrylic blocks to 

the attenuation layer while keeping the source to film distance 

the same. Ion chamber and film measurements of 6 

MV photons were taken on four occasions using Machine 1. 

One set of ion chamber measurements and a film calibration 

of 18 MV photons was acquired. Four sequential sMLC film 

calibrations were performed for repeatability testing. The 

above runs were performed using both Kodak XV2 and 

EDR2 films taken from the same batches. Daily machine 

output fluctuation was recorded by measuring the middle of a 

1010 cm field with an ion chamber. Traditional static film 

calibrations were performed by irradiating 1010 cm sections 

on individual XV2 films at a depth of 5 cm. Ion chamber 

measurements were taken once on Machines 2–4 for 

comparison purposes. Machine 2 was identical to Machine 1, 

while Machines 3 and 4 were Varian 2100 accelerators with 

Varian 80 leaf MLCs. Interleaf leakage was explored by determining 

the OD in the regions of the sMLC calibration 

films between exposed boxes. 

A Kodak film processor that is maintained by physicists 

developed all films, and a laser film digitizer Kodak LS75 

was used to measure ODs with Wellhofer Dosimetrie software. 

The horizontal uniformity was previously calibrated 

for use with XV2 films, but was not altered in this study to 

allow for the darker tint of EDR2 films. In the weeks following 

the initial measurements, daily film calibrations were 

performed and recorded that used different batches of EDR2 

film. The slopes of these calibration curves were estimated 

by using numerical three point derivatives. Errors were calculated 

by finding the standard error in the mean estimation 

using 95% t-distribution confidence levels. To estimate the 

total calibration error, standard errors for daily dose variations 

and single date OD repeatability measurements were 

added in quadrature. Standard deviations for same day measurements 

were compared to each other to assess dose variations. 

These deviations were not used in computing the total 

error, as the film repeatability study includes these fluctuations. 

FIG. 1. Visual representation of the sMLC calibration method. a The exposed 

film pattern and its collimator settings dashed lines. Isocenter is 

represented by the crosshairs. b The MLC movements and their exposures 

in MU used for file 2 and MLCs with 1 cm leaves. 

water backscatter layers. The ion chamber was inserted in the 

middle of a custom drilled 2 cm thick acrylic slab and placed 

at 95 SSD with 4 cm solid water attenuation and 2 cm solid 

water backscatter layers. Ion chamber readings were corrected 

for temperature and pressure. Doses delivered to XV2 

films ranged from 15–120 MU in 15 MU increments, while 

EDR2 films received 30–240 MU doses in 30 MU increments. 

Film calibrations using the new method were done at 

III. RESULTS 

Table II shows ion chamber measurement data for Machine 

1 on the four dates on which data were collected. 

These numbers can be halved exclusive of error for XV2 

film runs. The standard error in the table represents the standard 

error in the four averages measured on four separate 

dates, and is mainly due to machine output fluctuation and 

ion chamber inaccuracies. While the dose imparted to the 

film is fairly constant for regions not blocked by a collimator, 

sequential same day measurements showed the dose to have 

slightly higher standard deviations for areas that receive high 

amounts of MLC and main collimator scatter. The MLC remains 

open the entire time the 240 and 120 MU doses are 

delivered, but the 30 MU box is blocked by the MLC 75% of 

the time the beam delivers the dose to its side. In addition, all 

areas receive small doses while protected by the low trans- 



TABLE II. Data from the four daily sets of ion chamber measurements collected 

on Machine 1 for the sMLC calibration method with EDR2 film 

doses. 

MU 

Interdate 

average 

dose cGy 

95% confidence 

level in interdate 

average dose 

Intradate daily 

dose standard 

deviations 

240 219.7 0.29% 0.05% 

210 192.2 0.32% 0.04% 

180 164.1 0.42% 0.10% 

150 137.9 0.61% 0.06% 

120 110.8 0.40% 0.06% 

90 84.18 0.65% 0.17% 

60 56.69 0.92% 0.18% 

30 29.83 0.78% 0.54% 

mission jaws during exposure of the opposite side. Ion chamber 

measurements found that this dose was approximately 

0.57 cGy for the low column and 0.28 cGy for the high 

column at EDR2 doses. Table III shows comparative data for 

other Varian machines that were collected on only one date. 

These data show the amounts of machine variations that 

could be expected for this method of calibration. 

When comparing initial film calibrations produced by the 

various machines, it was noted that both machines with 

Varian 120 leaf MLCs had a more even interleaf leakage 

pattern than the 80 leaf MLC machines. Figure 2 shows 

scans of XV2 films that visually compare the leakages. The 

ODs ranged from 0.27 to 0.35 for the 80 leaf MLC, while the 

120 leaf MLC film scan produced a fairly constant OD near 

0.31. Though it may seem that the 120 leaf MLCs exhibit 

less interleaf leakage, the average OD in the leakage areas is 

nearly the same for both MLCs, indicating that the leakage is 

merely more evenly dispersed in Varian 120 leaf MLCs. The 

leakage is much less prominent on EDR2 film, which does 

not exhibit as high of sensitivity to the low-energy photons 

that escape between the MLC leaves. 

Figure 3 shows sensitometric curves produced on four 

separate dates by Machine 1. The final EDR2 run had ODs 

6% higher than the other calibrations at high doses. This 

demonstrates the need for a daily calibration, regardless of 

previous performance. All films used to prepare this graph 

were taken from the same batch. A two dimensional polynomial 

provided excellent fits to one set of daily data for each 

FIG. 2. Scans of interleaf leakage areas in XV2 films for two sizes of Varian 

MLCs. While the 80 leaf scans produced darker stripes up to an OD of 

0.35, the average OD of the entire interleaf area was 0.31 for both films. 

film. High agreement less than 0.5% difference between 

traditional static calibration and the single sheet IMRT 

method was found. While data from only one date are 

shown, they are representative of data collected on all four 

dates. The largest discrepancy between the two methods was 

1.0%. Calibrations with doses ranging up to 360 MU 330 

cGy for EDR2 film showed a continued linear OD response. 

Repeatability errors were low for both films. Figure 4 

shows the variation of four sequential calibrations. In this 

study we give an estimate of the maximum precision that can 

be obtained with this method of absolute film dosimetry. The 

deviation from the average value of all points was less than 

1%. The average difference between high and low values for 

all eight dose regions was 1.44% for XV2 and 1.22% for 

EDR2 film. The maximum difference for these points was 

the high dose region for both films, 2.15% for XV2 and 

1.58% for EDR2. 

The behavior of the OD of the film at different depths was 

explored to predict deviations from calibrated behavior when 

film is used in phantoms. Figure 5 shows results for sensitometric 

curve generation at different depths using the sMLC 

TABLE III. EDR2 single data set ion chamber measurements for other Varian accelerators using the sMLC calibration method. 

Machine 2 Machine 3 Machine 4 

MU 

Dose 

cGy 

% diff. from 

Machine 1 

Dose 

cGy 

% diff. from 

Machine 1 

Dose 

cGy 

% diff. from 

Machine 1 

240 219.4 0.15% 220.2 0.22% 220.4 0.29% 

210 191.9 0.16% 191.7 0.26% 193.9 0.87% 

180 163.9 0.10% 165.6 0.92% 168.5 2.71% 

150 137.5 0.25% 139.8 1.41% 141.2 2.41% 

120 111.2 0.41% 110.7 0.05% 110.7 0.10% 

90 84.63 0.53% 83.61 0.68% 84.64 0.54% 

60 56.96 0.47% 56.95 0.45% 57.96 2.24% 

30 29.97 0.46% 30.17 1.13% 30.73 3.01% 



FIG. 3. Sensitometric curves generated on four dates, using a XV2 film and 

b EDR2 film. The lines are second order polynomial fits of Day 1 data 

whose equations are printed in the graphs. 

FIG. 4. Single date repeatability sensitometric curves for a XV2 film and 

b EDR2 film. These data were taken sequentially on the same machine. 

calibration method. The middle 11 cm portion of the film 

exposure areas were measured, and varied by less than 2% 

for either film at any depth. While the middle portions of the 

dose boxes gave desired results for these small field sizes, 

Fig. 6 shows that the penumbra areas are overrepresented by 

XV2 film, since they contain progressively more low-energy 

photons as the depth is increased. It should also be noted that 

the visibly exposed portions of the EDR2 film at a depth of 5 

cm measure nearly 33 cm, while at any depth XV2 film 

shows them to be much larger. 

When the difference in the horizontal OD uniformity in 

the laser film scanner was explored, it was found that EDR2 

film is more affected by the brightness difference from the 

lighter center to the darker edges than XV2 film. Figure 7 

demonstrates that the ODs determined for low doses will 

have more error than high doses. 

Figure 8 compares sensitometric curves between 6 and 18 

MV photons for both types of film. EDR2 film again demonstrates 

that it has less energy dependence than XV2 film, 

as the calibration curves are nearly identical for both energies. 

When viewing the film scans, it was noted that 18 MV 

photons exhibited a significantly more varied dose distribution 

in the small 33 cm field than 6 MV photons. The 

EDR2 film again displayed a more uniform and compact 

dose delivery than XV2 film due to its lower-energy sensitivity. 

The total calibration error can be thought of as the minimum 

amount of error in dose values that are determined 

from the derived sensitometric curve. The average error 

across the calibration dose range was 0.68% for EDR2 film 

and 0.75% for XV2 film. Errors for XV2 film are slightly 

higher due to its higher variations in the OD repeatability 

study. Machine 1’s daily output never fluctuated more than 

0.3% from its calibrated value. Other errors will be present in 

the clinical implementation of film dosimetry, as previously 

discussed in the Introduction section. Potential errors generated 

by the horizontal inconsistency of the film scanner were 

not incorporated into these error estimates. 

In the initial weeks of using the sMLC method to generate 

film calibration curves, daily calibration data was saved and 

compared. Figure 9 shows 21 calibration curves taken for 

clinical usage. The two curves with significantly greater OD 

than the others were created after the film processor underwent 

extensive repairs. The standard deviations for all points 

were consistently around 7.5% of the OD. This collection of 

curves emphasizes the day-to-day variations that occur in 

film dosimetry. Since the EDR2 film sensitometric curve is 

linear, a 7.5% OD daily standard deviation translates into a 

7.5% uncertainty in dose calculations. The difference in film 

OD response to dose can be even larger if changes were 

made to the processor or its chemicals. The large magnitude 

of this variation highlights the need to complete daily film 

calibrations for accurate IMRT film-based quality assurance, 

rather than the relying on the average response curves. 



FIG. 7. Percent difference between OD measurements taken from the edge 

and middle of the scanning laser film densitometer. The differences in the 

measurements are due to the Vignetting effect, as the laser must pass 

through a greater distance of film along the edges than the central area 

measurements. 

FIG. 5. Generation of sensitometric curves using the sMLC method at different 

film depths, measuring only the center of exposed portions of film. 

Curves show responses of a XV2 film and b EDR2 film. 

IV. DISCUSSION AND CONCLUSIONS 

The more even distribution of interleaf leakage from 

Varian 120 leaf MLCs as compared to 80 leaf MLCs produced 

better film calibrations. While 120 leaf MLCs do not 

reduce the amount of interleaf leakage, ion chamber measurements 

will account for the added dose and the dose measurements 

will be more independent of the ion chamber’s 

placement within the fields. Homogeneous dose distributions 

within the exposed regions also reduce the repeatability error 

in measuring average ODs. 

EDR2 film’s darker base tint causes it to have a greater 

response to the film scanner’s horizontal nonuniformity 

FIG. 6. Scans of high dose box of EDR2 film left and XV2 film right for 

different depths. The increased area of the XV2 boxes at high depths demonstrates 

its low-energy photon overresponse. 

FIG. 8. A comparison of 6 MV and 18 MV sensitometric curves for a XV2 

film and b EDR2 film. It can be seen that EDR2 film is not dependent on 

beam energy for calibrations. 



FIG. 9. 21 calibration curves taken over a period of 6 weeks that show 

variations in film batches and the effects of processor maintenance and 

repair. The line represents the average of all data. 

when compared to XV2 film. The densitometer has a single 

laser directed towards a mirror that pivots to scan the laser 

beam in a horizontal path across the film, leading to a range 

of laser pathlengths through the film. The laser light will be 

more attenuated at outer regions due to its increase in pathlength 

known as the vignetting effect. Film densitometers 

are usually calibrated to adjust laser intensity as a function of 

pivot angle to maintain a uniform image for a range of OD 

values, but EDR2 film’s dark bluish tint exhibits increased 

attenuation of the red HeNe laser. Film densitometers that 

use a pivoting laser and were previously calibrated to maintain 

horizontal uniformity with XV2 film may need to be 

recalibrated for use with EDR2 film. This can introduce 

problems when the same densitometer is used for both types 

of film. The calibration is further complicated by the fact that 

the non-uniform attenuation is a function of OD; thus software 

corrections may need to be applied. 

The error due to the vignetting effect will affect clinical 

results for large film measurements that exhibit nonuniformity 

across their surface or for small films that are measured 

in the center of the scanner. Small fields are affected due to 

the calibration areas being scanned in a different location 

than the film measurements. The vignetting effect essentially 

makes the calibration curve specific to the horizontally 

scanned region of the pivoting laser densitometer. While this 

error is small enough to ignore with lightly tinted XV2 film, 

it cannot be assumed that EDR2 film will exhibit similar 

performance. At clinically relevant doses above 100 cGy, 

EDR2 film exhibits a 3% horizontal uniformity uncertainty 

error. When this error is combined with the uncertainties in 

the calibration method and other film errors, it is possible to 

have total verification errors exceeding 4%. As EDR2’s OD 

is linear with dose, this becomes a 4% dose uncertainty. 

When making a clinical transition from XV2 to EDR2 film 

measured on a pivoting laser densitometer, it is necessary to 

analyze the horizontal OD uniformity in order to obtain accurate 

IMRT verifications. 

The two field treatment method described earlier yielded 

minimal XV2 heterogeneities and virtually no detectable 

leakage with EDR2 films. The use of a single sMLC file to 

deliver the full pattern in one field was also studied. Although 

no quantitative measurements were taken to compare 

the increased dose fluctuations, interleaf leakage using one 

field was visible with XV2 film and noticeable when EDR2 

film was read with a densitometer. Additionally, the use of 

two fields only adds around a minute to the delivery time and 

allows the calibration to be scaled to different doses simply 

by inputting different MU settings. The use of the secondary 

collimator jaws to create this exposure pattern Fig. 1a was 

found to be too slow, difficult to scale for different dose 

ranges, and yielded minimal scatter improvements over the 

two field sMLC method. 

The high reproducibility, low error, quick delivery time, 

ease of use, and agreement with the traditional static film 

calibration method shows that the two field IMRT calibration 

method is superior to previous procedures. It is further recommended 

that EDR2 film be used clinically due to its near 

tissue equivalent response to low-energy photons, its linear 

response for large doses, and its better repeatability for highenergy 

doses compared to XV2 film. However, scanning laser 

film densitometers may have to have their horizontal uniformity 

recalibrated for use with EDR2 film. The depth study 

revealed inconsistencies in XV2’s response at central and 

penumbra areas, perhaps partially explaining the inconsistent 

reporting of XV2’s OD depth dependence in previous research. 

The use of thin lead shields and proper phantom 

placement should improve the accuracy of XV2 film dosimetry 

for IMRT verification but additional research is needed. 

Accurate average doses for each exposure region were found 

by taking only four sets of ion chamber measurements on 

different days. Large variations in both the amplitude and 

slopes of daily sensitometric curve generation were found 

and emphasized the need for full dose range daily calibrations 

which can be completed easily and accurately with the 

proper use of MLCs. The entire two field sMLC calibration 

process will require around ten minutes, and its success 

would negate the need to perform an ion chamber measurement 

on IMRT plans in addition to film verifications. 

a Corresponding author. Electronic mail: nchildre@mdanderson.org 

1 I. J. Das and C. W. Cheng, Basic Film Dosimetry, AAPM Annual Meeting 

Presentation Salt Lake City, Utah, 2001. 

2 F. H. Attix et al., Radiation Dosimetry Academic, New York, 1966, Vol. 

2, pp. 344–357. 

3 C. Danciu, B. S. Proimos, J. C. Rosenwald, and B. J. Mijnheer, ‘‘Variation 

of sensitometric curves of radiographic films in high energy photon 

beams,’’ Med. Phys. 28, 966–974 2001. 

4 J. F. Williamson, F. M. Khan, and S. C. Sharma, ‘‘Film dosimetry of 

megavoltage photon beams: A practical method of isodensity-to-isodose 

curve conversion,’’ Med. Phys. 8, 94–981981. 

5 S. E. Burch et al., ‘‘A new approach to film dosimetry for high energy 

photon beams: Lateral scatter filtering,’’ Med. Phys. 24, 775–783 1997. 

6 I. J. Yeo, C. K. C. Wang, and S. E. Burch, ‘‘A filtration method for 

improving film dosimetry in photon radiation therapy,’’ Med. Phys. 24, 

1943–1953 1997. 

7 S. G. Ju, Y. C. Ahn, S. J. Huh, and I. J. Yeo, ‘‘Film dosimetry for intensity 

modulated radiation therapy: Dosimetric evaluation,’’ Med. Phys. 29, 

351–355 2002. 

8 T. W. Holmes and E. C. McCullough, ‘‘Acceptance testing and quality 

assurance of automated scanning film densitometers used in the dosimetry 

of electron and photon therapy beams,’’ Med. Phys. 10, 698–700 

1983.

Rapid radiographic film calibration for IMRT verification using ...

Create successful ePaper yourself

Delete template?

Save as template?