- Email: [email protected]
A Simple Oral Cone Attachment for the Varian Linear Accelerator
Medical Dosimetry, Vol. 23, No. 1, pp. 47–50, 1998 Copyright © 1998 American Association of Medical Dosimetrists Printed in the USA. All rights reserved 0958-3947/98 $19.00 1 .00
PII S0958-3947(97)00119-2
A SIMPLE ORAL CONE ATTACHMENT FOR THE VARIAN LINEAR ACCELERATOR M. CEDERBAUM, M. ZALIK, E. ROSENBLATT, R. BAR-DEROMA, and A. KUTEN Northern Israel Oncology Center, Rambam Medical Center, 31096 Haifa, Israel Abstract—Lesions in the oral cavity are often treated with two opposed lateral fields. These include a significant amount of normal healthy tissue whose radiation tolerance is dose-limiting. The tumor dose can be boosted to tumorcidal levels by brachytherapy or by small electron fields directed straight on the lesion. We have developed a simple attachment to the standard electron applicator of the Varian Clinac 1800 that allows irradiation of small electron fields through acrylic tubes—the oral cones. These tubes have been evaluated in terms of depth dose and field profiles for 6, 9, 12, 16, and 20 MeV electrons using film for relative dosimetry. At these small field sizes there are significant changes in output factors, in the depth dose as well as in the effective size of the field, and a thorough dosimetric evaluation is imperative prior to treatment. The attachment can be manufactured locally at low cost. For reasons of patient safety the assembly is collapsible. In clinical practice the cone is directed directly on the tumor. For deep-seated lesions we use a penlight and a mirror for positioning. © 1998 American Association of Medical Dosimetrists. Key Words: Intra-oral cone, Small-field electron dosimetry, Film dosimetry, Relative output factor.
respectively. In addition, the two smaller tubes are bevelled at the treatment end to angles of 0, 15, 30, and 45 degrees, while the largest tube is bevelled at 15° only. The nominal source-skin (SSD) distance is 114.6 cm for the straight tube. The assembly is held together by gravity and will slide upwards into the applicator if an undue force is applied from below. We consider this to be an important patient safety feature. Dosimetric measurements were done for all tubes and for all electron energies. The nominal electron beam energies of our accelerator are 6, 9, 12, 16, and 20 MeV. Measurements of percentage depth dose (PDD) were done using film inserted tightly into a machined recess in a black polystyrene cassette. The cassette, together with additional scattering material, was placed on its edge parallel to the beam axis. By this technique we should have avoided the classic pitfalls of electron film dosimetry—air pockets and misaligned film. Used in this way, film is a reliable vehicle for relative electron dosimetry except for the region of dose build-up3,4 where the higher average Z of the film causes the electrons to scatter out, and thus reduces the dose. The accelerator gantry was rotated so that the cassette edge was parallel to the bevelled end of the cone, as would tissue in a clinical setup (Fig. 2). Optical density was measured with a densitometer (Victoreen Model 07-424) with a 1 mm aperture and was corrected for film non-linearity by using an analytical fit of the film sensitometric curve. Because of the cassette construction our first measurable point was at a depth of 4 mm inside the phantom. A typical PDD curve was compared to depth dose as measured with a parallel plate ion chamber, with good agreement within the limits of spatial accuracy of the densitometer measurements. The ion chamber read-
INTRODUCTION Lesions in the oral cavity are often treated with two opposed lateral fields that include a significant amount of normal healthy tissue. The radiation tolerance of these tissues limits the tumor dose to around 60 Gy while a higher dose is often desirable. The tumor dose can be boosted by brachytherapy or by small electron fields directed straight upon the lesion— both systems strive to give a full dose to the target volume while sparing surrounding structures. The use of small electron fields has been pioneered by C.C. Wang and P.J. Biggs1,2 at Massachusetts General Hospital in Boston. The fields are delivered through acrylic tubes, the oral cones, directly on the tumor bed in the oral cavity. This work describes a simple (and inexpensive) attachment to a standard electron applicator that we have designed for the oral cone technique. Dosimetric evaluation has been done using film for relative dosimetry and ion chamber for calibrations. METHODS AND MATERIALS A brass plate was machined to fit into the trimmer holder of a standard 10 3 10 cm electron applicator for the Varian Clinac 1800 accelerator (Fig. 1). A central hole with a step in the brass plate accommodates aluminum collars that are attached to the oral cones by three screws. There is a separate collar for every tube diameter. The cones are made of commercial acrylic tubes with a wall thickness of 3 mm and are about 20 cm long. The tubes have an inner diameter of 19, 24, and 34 mm, Reprint requests to: M. Cederbaum, Rambam Medical Center, Northern Israel Oncology Center, Physics Unit, POB 9602, 31096 Haifa, Israel. 47
48
Medical Dosimetry
Volume 23, Number 1, 1998
further constricted at the treatment depth, as is always the case for electron beams. The oral cone output factors were measured relative to the output factor of the open 10 3 10 cm field at SSD 5 100 cm. Typical values of these factors are given in Table 2. We also explored the influence of an air gap between the end of the tube and the surface. For the 19 mm tube the dose decreased by 6.2% in a 6 MeV beam and 4.2% in a 12 MeV beam with a 1 cm air gap. DISCUSSION
Fig. 1. The oral cone assembly mounted in the trimmer holder of the Varian Clinac 1800 electron applicator.
ings were converted to relative dose, taking into account the changes in stopping power with depth. Kodak XOmat V film was used throughout, cut to fit snugly into the cassette. All films were developed in the same processor though not at the same time. Dose measurements were always done along the beam axis, and also for beams with oblique incidence on the phantom. Lateral profiles were obtained by irradiating a film in its envelope at different depths in polystyrene. Absolute calibration was done at the depth of dose maximum with a Markus chamber (PTW, Freiburg, Germany) in a polystyrene phantom. The influence of air gaps was also explored with the Markus chamber.
The main clinical parameter is the depth of penetration, as determined by the chosen treatment isodose from the PDD curves. As described earlier, the curves are not reliable for the build-up region due to the high Z of the film. For larger open fields, the dose maximum tends to be broader as the energy increases. The dependence of PDD on field size is shown in Fig. 5. At 6 MeV there is only a small difference between the PDD of the oral cones and that of an open 6 3 6 cm field. The dose maximum is closer to the surface for the smaller tube as expected3 but the penetration depth is almost the same; however, for a 20 MeV beam the differences in penetration are very large. This is due to the loss of side-scatter that contributes significantly to the dose on the central
RESULTS The PDD curves were measured for every combination of cone diameter, bevel angle and electron energy. Figure 3 shows the depth dose as measured for three different cone diameters (15° bevel) along the beam axis. Every curve is normalized to its dose maximum. The accuracy in depth determination is 2 mm. The influence of the bevel angle on the PDD is shown in Fig. 4 for the 24 mm cone at 6 MeV and 16 MeV. When the field size is smaller than a certain critical value (that depends on energy) there is an appreciable influence of field cross-sectional area on the PDD. Figure 5 shows the PDD for the 19 and 24 mm diameter cones as well as for the open 6 3 6 cm electron applicator for the two extremes of the energy range, 6 and 20 MeV. The lateral profiles can be used to construct isodoses. The treatment isodose, here taken to be 90%, is significantly smaller than the nominal tube diameter; its width is given in Table 1. The treatment isodose is
Fig. 2. The measurement setup.
A simple oral cone attachment ● M. CEDERBAUM et al.
49
Fig. 4. PDD curves for different bevel angles in a 6 MeV and 16 MeV beam.
Fig. 3. PDD curves for the three cones (bevel angle 15°).
axis, occurring when the field dimension is smaller than a certain critical value that depends on energy and on machine design.5 This value is usually approximately the practical range of the electrons. In our case, the practical range of 6 and 20 MeV electrons is 3 and 10 cm, respectively; thus, this effect is very pronounced for the 20 MeV electrons when the field size is only a quarter or a third of the practical range. A disadvantage of our technique with the film inserted in a cassette is that the first measurable point is at 4 mm beyond the surface; thus, providing no information on the surface dose. The film method is not suitable for measurements in the build-up zone; the appropriate method for doing measurements in this zone is by using a plane-parallel ion chamber in a polystyrene or water phantom. Biggs6 has measured the influence of oblique incidence of an electron beam on the PDD, and found a large influence for the lower energies (6 MeV) that decreases
Fig. 5. PDD curves for different field sizes in a 6 MeV and 20 MeV beam.
50
Medical Dosimetry
Table 1. Effective field size of the oral cones Cone (mm)
Width of 90% isodose (mm)
19 24 34
16 20 28
Measured at depth of dose maximum.
in importance as the energy increases. A very similar behavior was found in these measurements as shown in Fig. 4 for a low and a high energy, respectively. The depth of dose maximum decreases as the angle of obliquity increases. The change in output factor is, as expected, quite significant (especially for the smallest field size) and is due to the change in SSD, the restricted field sizes, and contributions from the materials of the oral cone assembly itself. These measurements do not allow for the separation of the three possible contributions to the change in output factor. The large changes caused by the small field sizes and the angle of incidence stress the importance of measuring all parameters for every combination of tube diameter, bevel angle, and beam energy. In a clinical situation there are often air gaps in part of the field between the lesion and the tube end. The gap causes the dose to decrease in an unpredictable way. One method of quantifying this effect is by defining an effective sourceto-skin distance that can be very much shorter than the nominal one, especially for the lower energies.8,9 The effective SSD is strongly dependent on the particular accelerator used and the design and size of the collimators, as well as on energy. This concept is not readily applicable to the particular geometry of a long and narrow tube that collimates the beam; hence, the effective
Table 2. Typical calibration factors for the oral cones Cone (mm)
Bevel angle (°)
Energy 6 MeV
Energy 20 MeV
19
0 15 30 45 0 15 30 45 15
0.36 0.32 0.32 0.38 0.45 0.46 0.46 0.48 0.60
0.71 0.68 0.67 0.64 0.76 0.76 0.76 0.76 0.74
24
34
The calibration factors are relative to the output of a 10 3 10 cm open field at SSD 5 100 cm.
Volume 23, Number 1, 1998
SSD measured for small fields shaped with conventional trimmers cannot be applied to the oral cones. In clinical practice, one should avoid air gaps where possible. The effective field size (Table 1) is about 20% smaller than the physical diameter of the cone; this should be taken into account while choosing the most appropriate cone for a given clinical situation. Tumors in the oral cavity often protrude the oral mucosa; therefore, the full dose should be given to the surface. This is achieved by placing a piece of bolus material on the lesion. The thickness of the bolus should correspond to the depth of dose maximum as determined by the PDD curves for the oral cones actually in use. In treatment, the cone is simply put directly on the target. Positioning is facilitated by turning the treatment couch 90° and adjusting the accelerator gantry. For lesions close to the mouth the positioning is done by sight. For lesions further back in the oral cavity we use a mirror and a penlight to look down the tube and adjust its position. CONCLUSION A simple oral cone attachment to an existing electron applicator has been designed and built. It can be manufactured locally at low cost. Dosimetric evaluation has been performed using a film technique for relative dosimetry and ion chamber for output measurements. The cone assembly is now in routine clinical use. REFERENCES 1. Biggs, P.J.; Wang, C.C. An intra-oral cone for an 18 MeV linear accelerator. Int. J. Radiat. Oncol. Biol. Phys. 8:1251–1256; 1982. 2. Wang, C.C.; Biggs, P.J. Technical and radiotherapeutic considerations of intra-oral cone electron beam radiation for head and neck cancer. Semin. Radiat. Oncol. 2:171–179; 1992. 3. Niroomand-Rad, A.; Gillin, M.T.; Kline, R.W.; Grimm, D.F. Film dosimetry of small electron beams for routine radiotherapy planning. Med. Phys. 13:416 – 421; 1986. 4. Shiu, A.S.; Otte, V.A.; Hogstrom, K.R. Measurement of dose distributions using film in therapeutic electron beams. Med. Phys. 16:911–915; 1989. 5. Redpath, A.T.; Williams, J.R.; Thwaites, D.I. Treatment planning for external beam therapy. In: Williams, J.R.; Thwaites, D.I., editors. Radiotherapy physics in practice. Oxford University Press; 1993:135–185. 6. Biggs, P.J. The effect of beam angulation on central axis percent depth dose for 4 –29 MeV electrons. Phys. Med. Biol. 29:1089 – 1096; 1984. 7. Horton, J.L. Relative dosimetry. In: Williams, J.R.; Thwaites, D.I., editors. Radiotherapy physics in practice. Oxford University Press; 1993:53–75. 8. Hakin, G.; Faermann, S.; Krutman, Y.; Kushilevski, A. Dosimetry of small and irregularly shaped electron beams for the Varian Clinac 18 linear accelerator. In: Measurement assurance in dosimetry. International Symposium Proceedings, May 1993. International Atomic Energy Agency, Vienna; 1994.
PII S0958-3947(97)00119-2
A SIMPLE ORAL CONE ATTACHMENT FOR THE VARIAN LINEAR ACCELERATOR M. CEDERBAUM, M. ZALIK, E. ROSENBLATT, R. BAR-DEROMA, and A. KUTEN Northern Israel Oncology Center, Rambam Medical Center, 31096 Haifa, Israel Abstract—Lesions in the oral cavity are often treated with two opposed lateral fields. These include a significant amount of normal healthy tissue whose radiation tolerance is dose-limiting. The tumor dose can be boosted to tumorcidal levels by brachytherapy or by small electron fields directed straight on the lesion. We have developed a simple attachment to the standard electron applicator of the Varian Clinac 1800 that allows irradiation of small electron fields through acrylic tubes—the oral cones. These tubes have been evaluated in terms of depth dose and field profiles for 6, 9, 12, 16, and 20 MeV electrons using film for relative dosimetry. At these small field sizes there are significant changes in output factors, in the depth dose as well as in the effective size of the field, and a thorough dosimetric evaluation is imperative prior to treatment. The attachment can be manufactured locally at low cost. For reasons of patient safety the assembly is collapsible. In clinical practice the cone is directed directly on the tumor. For deep-seated lesions we use a penlight and a mirror for positioning. © 1998 American Association of Medical Dosimetrists. Key Words: Intra-oral cone, Small-field electron dosimetry, Film dosimetry, Relative output factor.
respectively. In addition, the two smaller tubes are bevelled at the treatment end to angles of 0, 15, 30, and 45 degrees, while the largest tube is bevelled at 15° only. The nominal source-skin (SSD) distance is 114.6 cm for the straight tube. The assembly is held together by gravity and will slide upwards into the applicator if an undue force is applied from below. We consider this to be an important patient safety feature. Dosimetric measurements were done for all tubes and for all electron energies. The nominal electron beam energies of our accelerator are 6, 9, 12, 16, and 20 MeV. Measurements of percentage depth dose (PDD) were done using film inserted tightly into a machined recess in a black polystyrene cassette. The cassette, together with additional scattering material, was placed on its edge parallel to the beam axis. By this technique we should have avoided the classic pitfalls of electron film dosimetry—air pockets and misaligned film. Used in this way, film is a reliable vehicle for relative electron dosimetry except for the region of dose build-up3,4 where the higher average Z of the film causes the electrons to scatter out, and thus reduces the dose. The accelerator gantry was rotated so that the cassette edge was parallel to the bevelled end of the cone, as would tissue in a clinical setup (Fig. 2). Optical density was measured with a densitometer (Victoreen Model 07-424) with a 1 mm aperture and was corrected for film non-linearity by using an analytical fit of the film sensitometric curve. Because of the cassette construction our first measurable point was at a depth of 4 mm inside the phantom. A typical PDD curve was compared to depth dose as measured with a parallel plate ion chamber, with good agreement within the limits of spatial accuracy of the densitometer measurements. The ion chamber read-
INTRODUCTION Lesions in the oral cavity are often treated with two opposed lateral fields that include a significant amount of normal healthy tissue. The radiation tolerance of these tissues limits the tumor dose to around 60 Gy while a higher dose is often desirable. The tumor dose can be boosted by brachytherapy or by small electron fields directed straight upon the lesion— both systems strive to give a full dose to the target volume while sparing surrounding structures. The use of small electron fields has been pioneered by C.C. Wang and P.J. Biggs1,2 at Massachusetts General Hospital in Boston. The fields are delivered through acrylic tubes, the oral cones, directly on the tumor bed in the oral cavity. This work describes a simple (and inexpensive) attachment to a standard electron applicator that we have designed for the oral cone technique. Dosimetric evaluation has been done using film for relative dosimetry and ion chamber for calibrations. METHODS AND MATERIALS A brass plate was machined to fit into the trimmer holder of a standard 10 3 10 cm electron applicator for the Varian Clinac 1800 accelerator (Fig. 1). A central hole with a step in the brass plate accommodates aluminum collars that are attached to the oral cones by three screws. There is a separate collar for every tube diameter. The cones are made of commercial acrylic tubes with a wall thickness of 3 mm and are about 20 cm long. The tubes have an inner diameter of 19, 24, and 34 mm, Reprint requests to: M. Cederbaum, Rambam Medical Center, Northern Israel Oncology Center, Physics Unit, POB 9602, 31096 Haifa, Israel. 47
48
Medical Dosimetry
Volume 23, Number 1, 1998
further constricted at the treatment depth, as is always the case for electron beams. The oral cone output factors were measured relative to the output factor of the open 10 3 10 cm field at SSD 5 100 cm. Typical values of these factors are given in Table 2. We also explored the influence of an air gap between the end of the tube and the surface. For the 19 mm tube the dose decreased by 6.2% in a 6 MeV beam and 4.2% in a 12 MeV beam with a 1 cm air gap. DISCUSSION
Fig. 1. The oral cone assembly mounted in the trimmer holder of the Varian Clinac 1800 electron applicator.
ings were converted to relative dose, taking into account the changes in stopping power with depth. Kodak XOmat V film was used throughout, cut to fit snugly into the cassette. All films were developed in the same processor though not at the same time. Dose measurements were always done along the beam axis, and also for beams with oblique incidence on the phantom. Lateral profiles were obtained by irradiating a film in its envelope at different depths in polystyrene. Absolute calibration was done at the depth of dose maximum with a Markus chamber (PTW, Freiburg, Germany) in a polystyrene phantom. The influence of air gaps was also explored with the Markus chamber.
The main clinical parameter is the depth of penetration, as determined by the chosen treatment isodose from the PDD curves. As described earlier, the curves are not reliable for the build-up region due to the high Z of the film. For larger open fields, the dose maximum tends to be broader as the energy increases. The dependence of PDD on field size is shown in Fig. 5. At 6 MeV there is only a small difference between the PDD of the oral cones and that of an open 6 3 6 cm field. The dose maximum is closer to the surface for the smaller tube as expected3 but the penetration depth is almost the same; however, for a 20 MeV beam the differences in penetration are very large. This is due to the loss of side-scatter that contributes significantly to the dose on the central
RESULTS The PDD curves were measured for every combination of cone diameter, bevel angle and electron energy. Figure 3 shows the depth dose as measured for three different cone diameters (15° bevel) along the beam axis. Every curve is normalized to its dose maximum. The accuracy in depth determination is 2 mm. The influence of the bevel angle on the PDD is shown in Fig. 4 for the 24 mm cone at 6 MeV and 16 MeV. When the field size is smaller than a certain critical value (that depends on energy) there is an appreciable influence of field cross-sectional area on the PDD. Figure 5 shows the PDD for the 19 and 24 mm diameter cones as well as for the open 6 3 6 cm electron applicator for the two extremes of the energy range, 6 and 20 MeV. The lateral profiles can be used to construct isodoses. The treatment isodose, here taken to be 90%, is significantly smaller than the nominal tube diameter; its width is given in Table 1. The treatment isodose is
Fig. 2. The measurement setup.
A simple oral cone attachment ● M. CEDERBAUM et al.
49
Fig. 4. PDD curves for different bevel angles in a 6 MeV and 16 MeV beam.
Fig. 3. PDD curves for the three cones (bevel angle 15°).
axis, occurring when the field dimension is smaller than a certain critical value that depends on energy and on machine design.5 This value is usually approximately the practical range of the electrons. In our case, the practical range of 6 and 20 MeV electrons is 3 and 10 cm, respectively; thus, this effect is very pronounced for the 20 MeV electrons when the field size is only a quarter or a third of the practical range. A disadvantage of our technique with the film inserted in a cassette is that the first measurable point is at 4 mm beyond the surface; thus, providing no information on the surface dose. The film method is not suitable for measurements in the build-up zone; the appropriate method for doing measurements in this zone is by using a plane-parallel ion chamber in a polystyrene or water phantom. Biggs6 has measured the influence of oblique incidence of an electron beam on the PDD, and found a large influence for the lower energies (6 MeV) that decreases
Fig. 5. PDD curves for different field sizes in a 6 MeV and 20 MeV beam.
50
Medical Dosimetry
Table 1. Effective field size of the oral cones Cone (mm)
Width of 90% isodose (mm)
19 24 34
16 20 28
Measured at depth of dose maximum.
in importance as the energy increases. A very similar behavior was found in these measurements as shown in Fig. 4 for a low and a high energy, respectively. The depth of dose maximum decreases as the angle of obliquity increases. The change in output factor is, as expected, quite significant (especially for the smallest field size) and is due to the change in SSD, the restricted field sizes, and contributions from the materials of the oral cone assembly itself. These measurements do not allow for the separation of the three possible contributions to the change in output factor. The large changes caused by the small field sizes and the angle of incidence stress the importance of measuring all parameters for every combination of tube diameter, bevel angle, and beam energy. In a clinical situation there are often air gaps in part of the field between the lesion and the tube end. The gap causes the dose to decrease in an unpredictable way. One method of quantifying this effect is by defining an effective sourceto-skin distance that can be very much shorter than the nominal one, especially for the lower energies.8,9 The effective SSD is strongly dependent on the particular accelerator used and the design and size of the collimators, as well as on energy. This concept is not readily applicable to the particular geometry of a long and narrow tube that collimates the beam; hence, the effective
Table 2. Typical calibration factors for the oral cones Cone (mm)
Bevel angle (°)
Energy 6 MeV
Energy 20 MeV
19
0 15 30 45 0 15 30 45 15
0.36 0.32 0.32 0.38 0.45 0.46 0.46 0.48 0.60
0.71 0.68 0.67 0.64 0.76 0.76 0.76 0.76 0.74
24
34
The calibration factors are relative to the output of a 10 3 10 cm open field at SSD 5 100 cm.
Volume 23, Number 1, 1998
SSD measured for small fields shaped with conventional trimmers cannot be applied to the oral cones. In clinical practice, one should avoid air gaps where possible. The effective field size (Table 1) is about 20% smaller than the physical diameter of the cone; this should be taken into account while choosing the most appropriate cone for a given clinical situation. Tumors in the oral cavity often protrude the oral mucosa; therefore, the full dose should be given to the surface. This is achieved by placing a piece of bolus material on the lesion. The thickness of the bolus should correspond to the depth of dose maximum as determined by the PDD curves for the oral cones actually in use. In treatment, the cone is simply put directly on the target. Positioning is facilitated by turning the treatment couch 90° and adjusting the accelerator gantry. For lesions close to the mouth the positioning is done by sight. For lesions further back in the oral cavity we use a mirror and a penlight to look down the tube and adjust its position. CONCLUSION A simple oral cone attachment to an existing electron applicator has been designed and built. It can be manufactured locally at low cost. Dosimetric evaluation has been performed using a film technique for relative dosimetry and ion chamber for output measurements. The cone assembly is now in routine clinical use. REFERENCES 1. Biggs, P.J.; Wang, C.C. An intra-oral cone for an 18 MeV linear accelerator. Int. J. Radiat. Oncol. Biol. Phys. 8:1251–1256; 1982. 2. Wang, C.C.; Biggs, P.J. Technical and radiotherapeutic considerations of intra-oral cone electron beam radiation for head and neck cancer. Semin. Radiat. Oncol. 2:171–179; 1992. 3. Niroomand-Rad, A.; Gillin, M.T.; Kline, R.W.; Grimm, D.F. Film dosimetry of small electron beams for routine radiotherapy planning. Med. Phys. 13:416 – 421; 1986. 4. Shiu, A.S.; Otte, V.A.; Hogstrom, K.R. Measurement of dose distributions using film in therapeutic electron beams. Med. Phys. 16:911–915; 1989. 5. Redpath, A.T.; Williams, J.R.; Thwaites, D.I. Treatment planning for external beam therapy. In: Williams, J.R.; Thwaites, D.I., editors. Radiotherapy physics in practice. Oxford University Press; 1993:135–185. 6. Biggs, P.J. The effect of beam angulation on central axis percent depth dose for 4 –29 MeV electrons. Phys. Med. Biol. 29:1089 – 1096; 1984. 7. Horton, J.L. Relative dosimetry. In: Williams, J.R.; Thwaites, D.I., editors. Radiotherapy physics in practice. Oxford University Press; 1993:53–75. 8. Hakin, G.; Faermann, S.; Krutman, Y.; Kushilevski, A. Dosimetry of small and irregularly shaped electron beams for the Varian Clinac 18 linear accelerator. In: Measurement assurance in dosimetry. International Symposium Proceedings, May 1993. International Atomic Energy Agency, Vienna; 1994.