Abstract. An important step in the verification of the reliability of portal films as in vivo dosemeters is the evaluation of the agreement between exit dose profiles ...
T he British Journal of Radiology, 70 (1997), 1283–1287 © 1997 The British Institute of Radiology
Short communication
Measurements of exit dose profiles in 60Co beams with a conventional portal film system 1M STASI, 1V CASANOVA BORCA and 2C FIORINO 1Divisione di Radioterapia, Ospedale degli Infermi, Biella, and 2Servizio di Fisica Sanitaria, Istituto Scientifico H San Raffaele, Milano, Italy Abstract. An important step in the verification of the reliability of portal films as in vivo dosemeters is the evaluation of the agreement between exit dose profiles and optical density profiles measured on the portal film. To test the possibilities of a conventional portal film system in 60Co beams suitable for head and neck irradiation, we verified the agreement between relative exit doses (measured by ionization chamber) and relative optical densities, on cubic homogeneous phantoms, on an homogeneous ‘‘step’’ phantom and on a cubic phantom including air and aluminium inhomogeneities. The optical density profiles were corrected with the appropriate sensitometric curves. For an homogeneous phantom 10.8 cm thick and with the film in contact with the phantom, the agreement was found to be excellent with a mean deviation of 0.8% and a maximum deviation of 1.5%. The agreement was worse when the air gap between the exit surface of the phantom and the portal film was increased (with an air gap equal to 15 cm the maximum deviation was 4%), and when the thickness of the phantom was increased (for a thickness of 14.4 cm the maximum deviation was 3.1%). The agreement was found to be acceptable for the ‘‘step’’ phantom too, with a mean deviation around 1% and a maximum deviation within 2% (air gap equal to zero). When air and aluminium inhomogeneities were incorporated into the phantom a maximum deviation of 6% and a mean deviation less than 3% were found. Furthermore, the relative optical density profiles show an underestimate of measured off-axis exit dose values under a high density inhomogeneity and a small overestimate under a low density inhomogeneity. Results suggest the possibility of using conventional portal films for exit relative dosimetry in head and neck irradiation with 60Co beams if the air gap is kept as small as possible.
Introduction The possibility of using portal films for exit dosimetry has been recently suggested in a number of papers [1–4]. Van Dam et al [3] verified the reliability of portal films for relative exit dose measurements by comparing measured dose profiles in cubic homogeneous and inhomogeneous phantoms with optical density profiles measured on portal films. Fiorino et al [4] tried to extend these results to situations closer to clinical conditions, by considering phantoms simulating a curved patient surface. Huyskens et al [5] and Fiorino et al [6] applied portal film transit dosimetry in combination with diodes to estimate midline dose distributions. Many groups are also investigating the possibilities of electron portal imaging systems and diode arrays for online portal dosimetry applications [7–10].
Some clinical applications of portal dosimetry for treatment planning verification [11–13] and for design and verification of compensators [14–16] have also been reported. Concerning conventional portal films, no data are available on the reliability of the method for measuring relative exit doses in 60Co beams. The aim of the present work was to investigate the possibilities of portal films in relative exit dosimetry by comparing optical density profiles derived from portal films with the corresponding exit dose profiles measured by an ionization chamber in the flattened region of the beam. Relative exit dose measurements must be considered to be only the first step toward the use of portal films in clinical in vivo dosimetry for head and neck irradiation with 60Co beams.
Materials and methods Received 29 July 1996 and in revised form 14 April 1997, accepted 24 July 1997.
Portal film system
Address correspondence to Dott. Stasi Michele, Servizio di Fisica Sanitaria, Ospedale Mauriziano, L. 90 Turati 62, 10128 Torino, Italia.
To check the validity of the method proposed in clinically realistic conditions, conventional portal films were used and processed as normal in
T he British Journal of Radiology, December 1997
1283
M Stasi, V Casanova Borca and C Fiorino
our radiotherapy department; Kodak X-Omat V films were used in Kodak localization cassettes. This combination of film and cassette was selected for clinical use on the basis of visual assessment of image quality [3]. Films were processed in a Kodak automatic processing unit and read by an automatic densitometer Scanditronix Therados.
Dosimetry system Exit dose measurements were performed with an NE 2571 cylindrical ionization chamber with a Keithley 35040 electrometer.
Phantoms We considered acrylic (PMMA) homogeneous phantoms (r=1.19 g cm−3 and area 25×25 cm2) with a variable thickness (7.2–14.4 cm) and the acrylic homogeneous ‘‘step’’ phantom shown in Figure 1. To simulate the neck region, we used the cubic inhomogeneous phantom shown in Figure 2, (7.8 cm thick), using acrylic, air and aluminium (r=2.7 g cm−3) simulating bone structures.
Figure 1. Acrylic ‘‘step’’ phantom (measurements in millimetres). The hatched square in the middle represents the irradiated field size (10 cm×10 cm at SSD=80 cm).
Figure 2. Inhomogeneous cubic phantom using acrylic, air and aluminium to simulate air cavities and bone structures (measurements in millimetres). The hatched square in the middle represents the irradiated field size (15 cm×15 cm at SSD=80 cm). 1284
Measurements Radiation beams were supplied by our Theratron 780 Cobalt unit. All the measurements were performed by positioning the gantry in the 90° position ( lateral field) and the phantoms at SSD=80 cm. The portal film in the cassette was positioned on a movable metallic support routinely used in our radiotherapy department. Film doses within 30–40 cGy were delivered, to keep the film in the quasi-linear region of the sensitometric curve (optical density within 1–1.2). However, optical density profiles were corrected with the appropriate sensitometric curve. The corrected optical density profiles were geometrically backprojected at the build down ‘‘exit’’ level, to compare them with the corresponding measured dose profiles. For all measurements, several portal films were collected and the optical density profiles (OAR , f defined by Weltens et al [11]) on the central axis were normalized at the point corresponding to the centre of the beam. Exit dose profiles were measured by inserting the ionization chamber at 0.5 cm from the exit surface into the phantom and moving it perpendicularly to the direction of the beam. Ionization values were normalized to the central axis ionization value (OAR defined by Weltens et al [11]). x This technique has been demonstrated to be sufficiently accurate for exit dose measurements [1–4]. The standard set-up was: field size 10×10 cm2, phantom thickness 10.8 cm, source-to-skin distance (SSD)=80 cm and portal film positioned at the level of the exit field for the homogeneous phantom. For the inhomogeneous cubic phantom, the thickness was 7.8 cm and, to include the inhomogeneities inside the field, the beam size was increased to 15×15 cm2. This thickness was chosen to simulate the neck region. We varied air gap g for the homogeneous cubic phantom (from 0 to 20 cm), field size (from 10×10 cm2 to 20×20 cm2) and thickness (from 7.2 cm to 14.4 cm), respectively. For the inhomogeneous phantom we varied only the air gap (0, 10, 15, 20 cm). To assess the reproducibility of the measurement of relative exit doses by portal films, 10 films have been irradiated under the same conditions (using the inhomogeneous cubic phantom) on different days. Films were processed at different times of the day in the automatic processor used for routine portal film processing [3, 11]. Optical densities were read at four fixed points (two in correspondence with acrylic, one of aluminium and one of air). A mean OAR for every f point was assessed. Deviations from the respective mean OAR are small with an average standard f deviation equal to 1.3%. This includes the possible variations in beam homogeneity, film processing T he British Journal of Radiology, December 1997
Short communication: Measurements of exit dose profiles in 60Co with portal film
and densitometer reproducibility. The location of measurement point ( behind air, aluminium or acrylic) has no influence on the reproducibility.
Results Homogeneous cubic phantom
and on the higher sensitivity of the portal film to low energy scattered photons. When the field size increases from 10×10 cm2 to 20×20 cm2, the agreement was found to be good (mean deviation is 0.8%).
‘‘Step’’ phantom
In standard set-up, the agreement between OAR f and OAR is found to be excellent with a maximum x deviation of 1.5% and a mean deviation (in absolute value) of 0.8%. Larger deviations are present near the penumbra region; they seem to depend mostly on small geometrical imprecisions. The agreement was worse with increasing distance between the exit surface of the phantom and the portal film, probably due to the differences between the scatter condition inside the phantom and the film [3]. For instance, with an air gap equal to 15 cm, the maximum deviation was 4% and the mean deviation was about 2% (Figure 3). When the thickness of the phantom increased from 7.2 cm to 14.4 cm, the agreement between OAR f and OAR was worse, as shown in Table 1. x Furthermore, the relative optical density values showed an overestimate of OAR , probably x depending on the increase of scattered radiation
In Figure 4 the exit dose profile measured with an ionization chamber is plotted with the optical density exit profile. The measurement was performed in the standard set-up (field size 10×10 cm2, SSD=80 cm and the portal film positioned with g=0). The agreement was quite good with a mean deviation of around 1% and a maximum deviation equal to 2.1%.
Inhomogeneous cubic phantom To include both inhomogeneities within the flattened region of the beam, the field size was increased up to 15×15 cm2 and the beam axis was adjusted at the centre of the phantom, crossing acrylic only. Results are shown in Figure 5 and Table 2. For air gaps up to 20 cm, the maximum deviation found was within 6.2% with mean deviations less than 3%. Larger deviations were found near the aluminium–acrylic and air–acrylic interfaces ( high gradient dose regions); so these could also depend on geometrical imprecisions. Under the aluminium inhomogeneity the dose was underestimated whereas under the air inhomogeneity it was overestimated.
Discussion
Figure 3. Homogeneous cubic phantom. The densitometric profile at 15 cm (PF_15) projected back to the exit side of the phantom taking beam divergence into account, is plotted together with the profiles obtained with the ionization chamber (IC) and with the portal film (PF) positioned in contact with the phantom. The field size was 10 cm×10 cm and the phantom thickness was 10.8 cm.
Good agreement was found between measured dose profiles and optical density profiles on portal films for homogeneous cubic phantoms, although for all measurements, the penumbra region, where
Table 1. Homogeneous cubic phantom. The agreement between optical density profiles and profiles measured with an ionization chamber is shown for different phantom thickness Thickness (cm)
Mean deviation (%)
Maximum deviation (%)
7.2 9 10.8 12.6 14.4
0.5 0.6 0.8 1.8 1.6
1.8 1.5 2 2.7 3.1
T he British Journal of Radiology, December 1997
Figure 4. ‘‘Step’’ phantom. The exit dose profile measured with an ionization chamber (IC) is plotted with the optical density exit profile measured with the portal film (PF) in contact with the phantom. The field size was 10 cm×10 cm. 1285
M Stasi, V Casanova Borca and C Fiorino
Figure 5. Inhomogeneous cubic phantom. The densitometric profile at 15 cm (PF_15) projected back to the exit side of the phantom taking beam divergence into account, is plotted together with the profiles obtained with the ionization chamber (IC) and with the portal film (PF) positioned in contact with the phantom. The field size was 15 cm×15 cm and the phantom thickness was 7.8 cm. Table 2. Inhomogeneous cubic phantom. The agreement between optical density profiles and profiles measured with an ionization chamber is shown for different air gap Air gap (cm)
Aluminium (%)
Air (%)
0 10 15 20
−3.7 −4.8 −5.3 −6.2
+0.3 +1 +1.7 +3.5
larger differences could be due to small geometrical imprecision [3, 17], was not taken into account. The agreement is worse when increasing the air gap and the thickness of the phantom. When the air gap increases, the contribution of the scattered radiation decreases so that the optical density profile at relatively large distances should be considered as mainly due to primary radiation. Regarding the ‘‘step’’ phantom, results suggest that the variable thicknesses do not significantly influence the measurement of relative exit doses by portal films. However, we have to remember that in clinical conditions (i.e. a curved patient contour at the exit side of the beam) one should correct the optical densities by inverse square correction factors [4, 5], or, alternatively, correct the measured/calculated exit doses [11] before comparing them with portal film data. The inhomogeneous phantom measurement results show that the relative optical density values of portal films are underestimated in areas corresponding to high density inhomogeneity and slightly overestimated for low density inhomogeneities. The agreement is worse when the air gap under both low density and high density inhomogeneities is increased. Similar considerations to those for the homogeneous phantom should be taken into account when g is increased. However, the agreement seems to be acceptable for clinical applications if the air gap is kept close to zero. 1286
Similar results have been reported for megavoltage beams when using as portal dosemeter conventional portal films [3, 5], electronic portal imaging devices [18] and diodes array [10]. However, with larger thicknesses and air gaps, the agreement between OAR and OAR is unx f acceptable. The accuracy of the technique should be worse with respect to megavoltage beams, because of the relatively larger influence of scattered radiation. The combination of an on-axis diode exit dose measurement with the relative exit dose distribution by portal film should permit the estimation of in vivo exit dose distributions [3, 4]. The combination of on-axis diode entrance and exit dose measurements with a portal film should permit the estimation of in vivo midline dose distributions with an acceptable accuracy [5, 6].
Conclusion In this paper we investigated the possibilities of a conventional portal film system (Kodak X-Omat V film and Kodak localization cassette) in portal dosimetry with 60Co beams. The reproducibility of the technique in measuring relative exit doses is acceptable (1 SD=1.3%). Good agreement was found between measured dose profiles and optical density profiles on portal films for homogeneous cubic phantoms (with air gap=0 and thickness=10.8 cm, the maximum deviation is 2%). The agreement gets worse when the air gap between the phantom and the portal film is increased (for g=15 cm, thickness=10.8 cm, maximum deviation 4%) and when the thickness of the phantom is increased (for 14.4 cm, g=0, maximum deviation 3.1%). Deviations of 6% with a mean value less than 3% between optical density profiles and measured dose profiles have been found for an inhomogeneous cubic phantom (air gap up to 20 cm). Furthermore, the relative optical density values show an underestimate of measured off-axis exit dose values under high density inhomogeneity and an overestimate under low density inhomogeneity. Relative exit dose measurements must be considered to be only the first step toward the use of portal films in clinical in vivo dosimetry for head and neck irradiation with 60Co beams. As suggested in this work and in other published papers, the use of portal films together with other absolute dosemeters (e.g. diodes) should permit the estimation of in vivo exit dose distributions [3, 4].
References 1. Wong JW, Slessinger ED, Hermes RE, et al. Portal dose images I: quantitative treatment plan verification. Int J Radiat Oncol Biol Phys 1990;18:1455–63. T he British Journal of Radiology, December 1997
Short communication: Measurements of exit dose profiles in 60Co with portal film 2. Ying X, Geer LY, Wong JW. Portal dose images II: patient dose estimation. Int J Radiat Oncol Biol Phys 1990;18:1465–75. 3. Van Dam J, Vaerman C, Blanckaert N et al. Are port films reliable for in vivo exit dose measurements? Radiother Oncol 1992;25:67–72. 4. Fiorino C, Del Vecchio A, Cattaneo GM, et al. Exit dose measurements by portal film dosimetry. Radiother Oncol 1993;29:336–40. 5. Huyskens D, Van Dam J, Dutreix A. Midplane dose determination using in vivo dose measurements in combination with portal imaging. Phys Med Biol 1994;39:1089–1101. 6. Fiorino C, Uleri C, Cattaneo GM, et al. Patient dose estimation in TBI using portal film transit dosimetry in combination with semiconductors detectors. Phys Med 1996;12:41–8. 7. Kirby MC, Williams PC. Measurement possibilities using an electronic portal imaging device. Radiother Oncol 1993;29:237–43. 8. Kirby MC, Williams PC. The use of an electronic portal imaging device for exit dosimetry and quality control measurements. Int J Radiat Oncol Biol Phys 1995;31:593–603. 9. Yin FF, Schell MC, Rubin P. Input/output characteristics of a matrix ion-chamber electronic portal imaging device. Med Phys 1994;21:1447–54. 10. Fiorino C, Uleri C, Cattaneo GM, et al. On-line exit dose profiles measurements by a diodes array. Phys Med Biol 1996;41:1291–1304.
T he British Journal of Radiology, December 1997
11. Weltens C, Van Dam J, Leunens G, et al. Reliability of clinical port films for measuring dose inhomogeneities in radiotherapy for head and neck tumours. Radiother Oncol 1994;30:167–70. 12. Fiorino C, Uleri C, Bolognesi A, et al. Treatment verification by portal films in chest and mantle field irradiation. Phys Med 1995;11:67–74. 13. Essers M, Lanson JH, Leunens G, et al. The accuracy of CT-based inhomogeneity corrections and in vivo dosimetry for the treatment of lung cancer. Radiother Oncol 1995;37:199–208. 14. Evans PM, Hansen VN, Mayles WPM, et al. Design of compensators for breast radiotherapy using electronic portal imaging. Radiother Oncol 1995;37:43–54. 15. Roback DM, Gerbi BJ. Evaluation of electronic portal imaging device for missing tissue compensator design and verification. Med Phys 1995;22:2029–34. 16. Yin FF, Schell MC, Rubin P. A technique of automatic compensator design for lung inhomogeneity correction using an electronic portal imaging device. Med Phys 1995;21:1729–32. 17. Sibata CH, Mota HC, Beddar AS, et al. Influence of detector size in photon beam profile measurements. Phys Med Biol 1991;36:621–31. 18. Essers M, Boellard R, Van Herk M, et al. Transmission dosimetry with a liquid-filled electronic portal imaging device. Int J Radiat Oncol Biol Phys 1996;34:931–41.
1287
