Effect of processing time delay on the dose response of Kodak EDR2 film
Nathan L. Childressa) and Isaac I. Rosen
Department of Radiation Physics, The University of Texas M. D. Anderson Cancer Center,
1515 Holcombe Boulevard, Unit 94, Houston, Texas 77030
共Received 6 April 2004; revised 19 May 2004; accepted for publication 28 May 2004;
published 23 July 2004兲
Kodak EDR2 film is a widely used two-dimensional dosimeter for intensity modulated radiotherapy
共IMRT兲 measurements. Our clinical use of EDR2 film for IMRT verifications revealed variations
and uncertainties in dose response that were larger than expected, given that we perform film
calibrations for every experimental measurement. We found that the length of time between film
exposure and processing can affect the absolute dose response of EDR2 film by as much as
4%– 6%. EDR2 films were exposed to 300 cGy using 6 and 18 MV 10⫻10 cm2 fields and then
processed after time delays ranging from 2 min to 24 h. An ion chamber measured the relative dose
for these film exposures. The ratio of optical density 共OD兲 to dose stabilized after 3 h. Compared to
its stable value, the film response was 4%– 6% lower at 2 min and 1% lower at 1 h. The results of
the 4 min and 1 h processing time delays were verified with a total of four different EDR2 film
batches. The OD/dose response for XV2 films was consistent for time periods of 4 min and 1 h
between exposure and processing. To investigate possible interactions of the processing time delay
effect with dose, single EDR2 films were irradiated to eight different dose levels between 45 and
330 cGy using smaller 3⫻3 cm2 areas. These films were processed after time delays of 1, 3, and 6
h, using 6 and 18 MV photon qualities. The results at all dose levels were consistent, indicating that
there is no change in the processing time delay effect for different doses. The difference in the time
delay effect between the 6 and 18 MV measurements was negligible for all experiments. To rule out
bias in selecting film regions for OD measurement, we compared the use of a specialized algorithm
that systematically determines regions of interest inside the 10⫻10 cm2 exposure areas to manually
selected regions of interest. There was a maximum difference of only 0.07% between the manually
and automatically selected regions, indicating that the use of a systematic algorithm to determine
regions of interest in large and fairly uniform areas is not necessary. Based on these results, we
recommend a minimum time of 1 h between exposure and processing for all EDR2 film
measurements. © 2004 American Association of Physicists in Medicine.
关DOI: 10.1118/1.1774111兴
Key words: film dosimetry, IMRT treatment plan verification, quality assurance, radiation therapy
INTRODUCTION
Radiographic film allows for fast and convenient twodimensional dose distribution measurements. However, its
high silver halide content overresponds to low energy photons because of an increase in photoelectric effect interactions. Kodak’s extended dose range 2 共EDR2兲 film has a
significantly lower silver halide content than earlier films,1
and is a commonly used film for intensity modulated radiotherapy verifications.2–7 Its altered chemical formulation and
smaller grain size allow it to measure higher doses without
becoming saturated and to produce results that exhibit good
agreement with other dosimeters.8 –14 However, these
changes have resulted in a film that has a much different
chemical structure than older, well-characterized15–21 radiographic films.
In our clinical use of EDR2 film for full patient plan intensity modulated radiotherapy 共IMRT兲 verifications, we noticed variations and uncertainties in dose response that were
larger than expected, given that we perform film calibrations
for every experimental measurement. We discovered that the
length of time between exposure and processing alters EDR2
2284
Med. Phys. 31 „8…, August 2004
film’s absolute dose response. Thus, an IMRT verification
may have varying absolute dose responses depending on the
time elapsed between exposure and processing of both patient film共s兲 and calibration film共s兲. We quantitatively characterized the relationship between the absolute dose response
of EDR2 film and the time between its exposure and processing.
METHODS AND MATERIALS
All films were processed using an RP X-Omat Model
M6B film processor 共Eastman Kodak Company, Rochester,
NY兲 with Polymax RT developer 共Eastman Kodak Company兲 at 33 °C. A 16-bit VXR-16 Dosimetry Pro film scanner 共VIDAR Systems Corporation, Herndon, VA兲 was used
to scan all films.
Optical density response versus time delay, single
dose
Kodak EDR2 films 共Eastman Kodak Company兲 were radiated in a solid water phantom at 95 cm source to surface
distance 共SSD兲 with 10⫻10 cm2 fields of 6 and 18 MV pho-
0094-2405Õ2004Õ31„8…Õ2284Õ5Õ$22.00
© 2004 Am. Assoc. Phys. Med.
2284
2285
N. L. Childress and I. I. Rosen: Effect of processing time delay on EDR2 film
tons from a Varian Clinac 2100EX 共Varian Medical Systems,
Inc., Palo Alto, CA兲. Every film was from the same batch.
An ion chamber placed 5 cm below the film recorded relative
doses at the center of the field to allow the variations in
machine output to be taken into consideration. The relative
dose to the film was computed based on this measurement.
Films were irradiated perpendicular to the beam’s central
axis at 5 cm depth in their original jackets to approximately
300 cGy. They were processed following time delays of 2, 4,
and 30 min and 1, 3, 6, 12, and 24 h. One film was used for
the 2 and 4 min intervals, two films were used for the 30 min
time interval, and three films were used for all other time
intervals. The average optical density 共OD兲 per ion chambermeasured dose of the film was calculated for each time interval. These values were normalized to the time period with
the highest OD/dose response for each energy. Assuming that
the uncertainties in OD measurement and ion chamber measurement are independent of the time delay, a single standard
deviation representing the overall reproducibility of the experiment was computed and applied to all time periods. The
overall standard deviation for each photon energy was calculated by combining the ratio of each individual OD/dose
measurement to its corresponding mean for all data with a
processing time delay of at least 1 h.
Optical density response versus time delay, multiple
doses
The single-dose experiment was designed to explore a
wide range of processing time delay effects and so it did not
determine whether the processing time delay effect is dose
dependent. A limited number of time delay values were studied at different clinically relevant dose levels to see if processing time delays affect all OD levels of the film equally.
Eight dose levels were delivered to a single film using a
multileaf collimator-generated eight box pattern.22 Experiments were done with 6 and 18 MV beams. Processing time
delays of 1, 3, and 6 h were used to investigate the dose
dependence of the processing time delay effect. Three films
in their original jackets were placed perpendicular to the
beam’s central axis at a depth of 5 cm in a solid water phantom at 95 cm SSD and were exposed at each time interval.
While this type of dose pattern is typically used for singlefilm calibrations, it was used in this experiment to deliver
many dose regions to a single piece of film. It was not possible to use an ion chamber to measure relative doses during
the delivery because multiple dose levels were delivered simultaneously to each piece of film. The average OD of each
dose region and the background OD were used for analysis.
Results for both photon energies were normalized to their
average OD at 6 h.
Optical density response versus time delay, multiple
film batches
The processing time delay effect was also explored using
three EDR2 film batches that differed from the batch used in
the previous two experiments. We also performed measurements with one batch of Kodak XV2 film for comparison
Medical Physics, Vol. 31, No. 8, August 2004
2285
purposes. For each batch of film, the average OD/dose of
three pieces of film with a 1 h processing time delay was
compared to the average OD/dose of three pieces of film
with 2, 4, and 6 min processing time delays. A one-tailed
nonpaired heteroscedastic t-test was performed to determine
whether the means of the two samples were statistically different for each batch of film. The order in which the films
were scanned and processed was randomized. The experimental methods of the single dose experiment were used to
expose the film and to measure relative doses with an ion
chamber. Only 6 MV beams were used, and only 100 cGy
was delivered to XV2 films.
Manual selection of OD region versus automatic
analysis
As the magnitude of the processing time delay effect is on
the order of a few percentage points, it was crucial to eliminate as many sources of error as possible. An in-house film
dosimetry program was used to calculate the average OD of
a manually selected rectangular region in each of the scanned
film images. While the 10⫻10 cm2 exposure areas were relatively uniform, differences in the sizes, shapes, and locations
of these user-defined areas could add uncertainty to the final
results. For comparison, we developed a program that automatically and systematically determined a region of interest
and computed its average value. The program detected the
edges of the radiation field in order to calculate a region of
interest that comprised the central 25% of the total area of
the square. Both methods were used to determine the average
values of the OD regions. Finally, the effect of rotational
errors was explored by analyzing films with and without a 2°
rotation.
RESULTS AND DISCUSSION
Optical density response versus time delay, single
dose
Figure 1 displays the processing time delay effect on OD/
dose response over 24 h for 6 and 18 MV photon beams. The
overall standard deviation of the measurements from each
processing time delay group’s average was 0.25% for all
points with at least 1 h between processing and exposure.
The maximum deviation in machine output was 1.1% for the
6 MV data and 0.6% for the 18 MV data. The 6 MV 2 min
through 6 h data points were measured a second time to
verify that these results were not anomalous 共Fig. 2兲. Both
experiments produced similar results. The normalized OD/
dose value of the 2 min data point is very dependent on the
exact length of time between exposure and processing, as it
lies in a high-gradient portion of the curve. While this value
was difficult to accurately determine, it was 4%– 6% lower
than the stable OD/dose value for both experiments. The
OD/dose response stabilized after 3 h. The film’s OD/dose
was within 1% of its stable value after 1 h, and within 2% of
its stable value after 30 min. It appears that there may be
latent image fading for 6 MV exposures 共lower OD/dose as
processing time delays increase兲 for delays greater than three
2286
N. L. Childress and I. I. Rosen: Effect of processing time delay on EDR2 film
2286
FIG. 2. EDR2 film OD per dose time delay effect for 6 MV. Each data point
represents a single film. The OD/dose values were normalized to the average
OD/dose at 6 h. This experiment was performed to verify the more complete
results displayed in Fig. 1.
measurements to correlate the OD results for each time period. These figures confirm that the average OD at 1 h is
approximately 1% less than the average OD at 3 and 6 h, and
there is no significant change in these percentages with dose
levels. Furthermore, the background of the film did not significantly vary with processing time delay. As EDR2 film’s
processing time delay effect does not show a dose depen-
FIG. 1. EDR2 film OD per dose response at various processing time delays
for 共a兲 6 MV and 共b兲 18 MV exposures. Each data set was normalized to its
highest OD/dose response. Processing EDR2 films immediately after exposure results in 4 – 6% less OD/dose compared to processing films after a 3– 6
hour time delay. Error bars represent one standard deviation.
hours. Although the difference between the 6 and 24 h data
points of 0.6% is marginally significant, the 18 MV data does
not show as clear of a trend. All 18 MV measurements past 3
h agree within one standard deviation. Overall, there was no
significant difference between the 6 and 18 MV time delay
responses.
Optical density response versus time delay, multiple
doses
Figure 3 shows the dose dependence of the processing
time delay for 6 and 18 MV exposures. Again, no significant
energy dependence was observed. The 18 MV exposures result in higher doses owing to the beam’s lower attenuation in
the solid water. The 210/230 cGy 共6/18 MV兲 data points
consistently had a higher standard deviation than the other
dose levels, thus their abnormal OD response is more likely
due to their lower reproducibility than to a trend in EDR2
film’s dose dependence. Although not shown in the figures,
the standard deviations of the reproducibility of these data
points are generally 0.4%. This standard deviation excludes
the possible machine output variation that could not be measured, as there were no unifying ion chamber absolute dose
Medical Physics, Vol. 31, No. 8, August 2004
FIG. 3. EDR2 film normalized OD response to 3⫻3 cm2 squares with
several different doses, for 共a兲 6 MV and 共b兲 18 MV exposures. There was
no clear processing time delay dose dependence. While not shown in the
figures, the standard deviation for each of the data points is approximately
0.4%.
2287
N. L. Childress and I. I. Rosen: Effect of processing time delay on EDR2 film
FIG. 4. Processing time delay results for different film batches. All EDR2
film batches were different than the batch used to generate Figs. 1–3. The
processing time delay effect was statistically significant for all three EDR2
batches, but not for XV2 film. Error bars indicate one standard deviation.
dence at 1 h, it is safe to use a 1 h period between exposure
and processing for clinical verifications without experiencing
additional discrepancies due to the range of dose levels.
Optical density response versus time delay, multiple
film batches
The differences in films with an average 4 min processing
time delay and 1 h delay for three batches of film are shown
in Fig. 4. All three batches of film exhibited similar behavior.
The maximum deviation in machine output was 0.2%. The
difference between the two processing time delay groups was
statistically significant in all batches, with a maximum
p-value of 0.007. The standard deviation of the 4 min delay
OD/dose values was slightly higher than the 1 h values, as
the 2– 6 min delay points are in a high gradient region of the
time delay effect curve. The mean of all 9 EDR2 films in the
2– 6 min processing time delay group was 0.962⫾0.009 共one
standard deviation兲, significantly lower than the 1 h group’s
1.000⫾0.002 OD/dose response (p⬍0.0001). The batch of
XV2 film did not have a statistically significant processing
time delay effect for the 4 min and 1 h time periods studied.
These results confirm this effect is fairly consistent among
different batches of film and is statistically significant for
EDR2 films.
Manual selection of OD region versus automatic
analysis
The maximum difference between the average OD/dose
of user-defined areas and automatically generated areas was
an insignificant 0.07%. The average difference was 0.02%
⫾0.02% 共one standard deviation兲. Changing the film’s rotation by 2° produced similar results. It can be concluded that
a specialized algorithm to determine region averages is not
needed for large exposure areas, the authors are remarkably
skilled at accurately and consistently selecting regions of interests, or both.
CONCLUSIONS
example, if calibration film共s兲 are exposed following several
patient-specific quality assurance film exposures and then all
films are processed together, the time delay for the calibration film共s兲 would be shorter than for the patient films. Consequently, the film calibration would show an OD/dose that
is too low compared to the patient films and the patient doses
would be overestimated. Since fluence verifications typically
require only a few minutes to deliver, their doses would be
uniformly high. Full-patient plan verifications, which require
between 10 and 45 min to deliver, could exhibit different
dose responses for individual beams. Thus, the beam that
was delivered first can have a 4%– 6% greater OD/dose response than the beam that was delivered last if the film is
processed immediately after exposure. This results in planning target volume dose normalizations that can differ from
one verification to another, depending on the length of time
elapsed between calibration and patient films and the order in
which they were exposed. This effect may partially explain
why some clinics routinely observe measured to calculated
dose normalizations that are not equal to 1.23 Clearly, not
compensating for the processing time delay effect in clinical
verifications could lead to confusing results for both individual verifications and analyses that compare the dose normalization for a large number of patients.
All films should be processed at least 1 h after exposure
when performing measurements with Kodak EDR2 film.
This guarantees that the OD/dose measurements of the films
have reached 99% of their stable values. The OD/dose response for XV2 films is consistent for time periods of 4 min
and 1 h between exposure and processing. It is not known if
the processing time delay effect varies with different film
processors or developing solutions. If a film calibration is
performed before an IMRT verification and both films are
processed immediately, selected beams from the IMRT film
may exhibit an OD/dose response up to 4%– 6% lower than
the same dose in the calibration film共s兲. Other beams may
show a 0%– 6% decrease in the expected OD/dose, resulting
in confusing patient-specific IMRT verification results. These
effects would be reversed if the calibration film共s兲 were exposed after the patient film共s兲. Further research focusing on
the chemical reaction rates in EDR2 film would be necessary
to determine the cause of this effect. As there was no observed dose dependence, any time period between exposure
and processing is acceptable if one does not need to correlate
the OD/dose response between films with short exposure durations.
ACKNOWLEDGMENTS
This research was partially supported by Philips Medical
Systems and the American Legion Auxiliary Fellowship. We
would also like to thank Jeff Byng from Kodak Canada for
his assistance and suggestion to research the processing time
delay effect.
a兲
Electronic mail: [email protected]
X. R. Zhu, S. Yoo, P. A. Jursinic, D. F. Grimm, F. Lopez, J. J. Rownd, and
M. T. Gillin, ‘‘Characteristics of sensitometric curves of radiographic
films,’’ Med. Phys. 30, 912–919 共2003兲.
1
The sequencing of film exposure and processing could
produce 4%– 6% errors in EDR2 film’s optical densities. For
Medical Physics, Vol. 31, No. 8, August 2004
2287
2288
N. L. Childress and I. I. Rosen: Effect of processing time delay on EDR2 film
2
P. Cadman, R. Bassalow, N. P. Sidhu, G. Ibbott, and A. Nelson, ‘‘Dosimetric considerations for validation of a sequential IMRT process with a
commercial treatment planning system,’’ Phys. Med. Biol. 47, 3001–
3010 共2002兲.
3
N. Dogan, L. B. Leybovich, A. Sethi, and B. Emami, ‘‘Automatic feathering of split fields for step-and-shoot intensity modulated radiation
therapy,’’ Phys. Med. Biol. 48, 1133–1140 共2003兲.
4
D. Georg, B. Kroupa, P. Winkler, and R. Potter, ‘‘Normalized sensitometric curves for the verification of hybrid IMRT treatment plans with multiple energies,’’ Med. Phys. 30, 1142–1150 共2003兲.
5
A. Kapulsky, E. Mullokandov, and G. Gejerman, ‘‘An automated
phantom-film QA procedure for intensity-modulated radiation therapy,’’
Med. Dosim 27, 201–207 共2002兲.
6
P. N. McDermott, T. He, and A. DeYoung, ‘‘Dose calculation accuracy of
lung planning with a commercial IMRT treatment planning system,’’ J.
Appl. Clin. Med. Phys. 4, 341–351 共2003兲.
7
S. L. Richardson, W. A. Tome, N. P. Orton, T. R. McNutt, and B. R.
Paliwal, ‘‘IMRT delivery verification using a spiral phantom,’’ Med.
Phys. 30, 2553–2558 共2003兲.
8
P. M. Charland, I. J. Chetty, S. Yokoyama, and B. A. Fraass, ‘‘Dosimetric
comparison of extended dose range film with ionization measurements in
water and lung equivalent heterogeneous media exposed to megavoltage
photons,’’ J. Appl. Clin. Med. Phys. 4, 25–39 共2003兲.
9
I. J. Chetty and P. M. Charland, ‘‘Investigation of Kodak extended dose
range 共EDR兲 film for megavoltage photon beam dosimetry,’’ Phys. Med.
Biol. 47, 3629–3641 共2002兲.
10
N. Dogan, L. B. Leybovich, and A. Sethi, ‘‘Comparative evaluation of
Kodak EDR2 and XV2 films for verification of intensity modulated radiation therapy,’’ Phys. Med. Biol. 47, 4121– 4130 共2002兲.
11
N. Dogan and G. P. Glasgow, ‘‘Surface and build-up region dosimetry for
obliquely incident intensity modulated radiotherapy 6 MV x rays,’’ Med.
Phys. 30, 3091–3096 共2003兲.
12
J. Esthappan, S. Mutic, W. B. Harms, J. F. Dempsey, and D. A. Low,
‘‘Dosimetry of therapeutic photon beams using an extended dose range
film,’’ Med. Phys. 29, 2438 –2445 共2002兲.
Medical Physics, Vol. 31, No. 8, August 2004
13
2288
A. J. Olch, ‘‘Dosimetric performance of an enhanced dose range radiographic film for intensity-modulated radiation therapy quality assurance,’’
Med. Phys. 29, 2159–2168 共2002兲.
14
X. R. Zhu, P. A. Jursinic, D. F. Grimm, F. Lopez, J. J. Rownd, and M. T.
Gillin, ‘‘Evaluation of Kodak EDR2 film for dose verification of intensity
modulated radiation therapy delivered by a static multileaf collimator,’’
Med. Phys. 29, 1687–1692 共2002兲.
15
M. Bucciolini, F. B. Buonamici, and M. Casati, ‘‘Verification of IMRT
fields by film dosimetry,’’ Med. Phys. 31, 161–168 共2004兲.
16
S. E. Burch, K. J. Kearfott, J. H. Trueblood, W. C. Sheils, J. I. Yeo, and
C. K. Wang, ‘‘A new approach to film dosimetry for high energy photon
beams: lateral scatter filtering,’’ Med. Phys. 24, 775–783 共1997兲.
17
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 共2001兲.
18
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 共2002兲.
19
C. Martens, I. Claeys, C. De Wagter, and W. De Neve, ‘‘The value of
radiographic film for the characterization of intensity-modulated beams,’’
Phys. Med. Biol. 47, 2221–2234 共2002兲.
20
J. L. Robar and B. G. Clark, ‘‘The use of radiographic film for linear
accelerator stereotactic radiosurgical dosimetry,’’ Med. Phys. 26, 2144 –
2150 共1999兲.
21
N. Suchowerska, P. Hoban, M. Butson, A. Davison, and P. Metcalfe,
‘‘Directional dependence in film dosimetry: radiographic and radiochromic film,’’ Phys. Med. Biol. 46, 1391–1397 共2001兲.
22
N. L. Childress, L. Dong, and I. I. Rosen, ‘‘Rapid radiographic film
calibration for IMRT verification using automated MLC fields,’’ Med.
Phys. 29, 2384 –2390 共2002兲.
23
A. J. Olch, ‘‘Dosimetric accuracy of the ITP inverse treatment planning
system,’’ Med. Phys. 29, 2484 –2488 共2002兲.