- Email: [email protected]
The physics of intraoperative radiotherapy using high energy electron beams
1102
THE PHYSICS OF I~OPE~~E BEAMS
Nuclear Instruments and Methods in Physics Research.BlO/ll (1985) 1102-1106 North-Holland, Amsterdam
~DIO~FY
USING HIGH ENERGY ELECI-IZON
Peter J. BIGGS Department of Radiafion Medicine, Mawachusetts General Hospiral: Hatvatd Medical S&a&, Bosfon, Massachusetrs 021Id, USA
Conventional radiotherapy involves the treatment of malignant tumors by ionizing radiation applied externally to the body. However, injury to internal organs and normal tissue often limits the tumor dose delivered to less than cancericidal values. This problem is circumvented in intraoperative radiotherapy, where the tumor is surgicaby exposed, normal tissue and organs displaced, and the diseased area treated directly with radiation. Electron beams are used for these procedures because of their finite range compared with X-rays, to spare normal tissue below the tumor.
1.
Introduction
Most major advances in the treatment of cancer with radiation have been based on the differences which can be obtained in the distribution of the radiation dose between the tumor and adjacent normal tissues. The probabi~ty of achieving locat control of a localized tumor mass generally increases as larger radiation doses are delivered to the tumor. However, in many clinical situations the dose that can safely be delivered to the tumor is restricted by the limited radiation tolerance of su~oundi~ normal tissues. The rationale for intraoperative radiation therapy (TORT) is to expose the tumorbearing structure and apply the radiation directly to the tumor, thus avoiding irradiation of normal structures. It is the purpose of this paper to describe the physical characteristics of the electron beams used to deliver radiation doses iuteroperativeiy to the tumor-bearing areas. Such characteristics concern the % depth-dose distributions, effect of field angulation, surface dose and bremsstra~ung background.
linear accelerator. The purpose of the adjustable collimator was to eliminate the need for a large number of fixed apertures. A thin (approximately 0.25 mm) transparent mylar sheet was fastened directly over the jaw aperture to prevent the possibility of any stray material from the machine falling directly into the patient. The field in the patient is defined by a transparent lucite tube that slides into an aluminum jacket, which in turn is rigidly attached to the secondary collimator. To align the treatment field with the electron beam, the lucite tube has to be “docked” with the aluminum jacket. This is achieved by rotating the gantry and moving the treatment couch until the head of the tube is aligned with and almost touching the metal jacket. The
ionization Chamber
-
PRIMARY
2. Methods and materials
COLLIMATORS
SECONDARY COLLIMATORS
2. I. Appara fw and inttaopeta~~ve procedure
The linear accelerator used for this procedure has a wide range of electron energies available, namely 6, 9, 12, 15, 18, 23 and 29 MeV, corresponding to a range of 1.7 to 6.1 cm in water at the 90% depth dose. The equipment is shown schematically in fig. 1. In addition to the fixed collimators and photon jaws (primary collimators) of the machine a set of adjustable jaws (secondary collimators) was used to define the electron field size at the surface. The adjustable jaws are attached to a plate that is bolted to the front face of the 0168-583X/85/$03.30 @ Elsevier Science Publishers B.V. (North-Holland Physics Publishing Division)
Trrotmmt
I
Tuba
I
Nottoscolo Fig. 1. GeneraI ~IIi~tion layout of the Iinear accelerator for intraoperative radiotherapy.
P.J. Siggs / The physics of intraoperative
couch is then moved in very short steps so as to ease the tube up into the jacket until it has reached the desired depth. The majority of the tubes are circular in cross section, ranging in inner diameter from 4 cm to 9 cm in steps of 1 cm. Many of these are bevelied. Wrapped around the top of each tube there is a 5 cm long, Delrin cylinder which slides into the metal jacket. This cylinder is of variable wall thickness, so that all the cones may slide into the same metal jacket. The gap between the cyhnder and the jacket was made to be a ~mpro~se between an easy fit and minimal lateral displacement of the tube at its lower end. Typically, the gap is between 0.05 mm and 0.075 mm and the lateral displacement is + (l-2) mm. 2.2. Beam
0.8
CONE SIZE
9Mev 0.7
s 3
0.5 -
8
0.4 -
h 2i=
0.3 -
1
~~
An extensive study was made of the relative dose delivered through each tube, the per cent depth dose, beam profiles, surface dose, and the per cent bremsstrahlung background to determine the optimum setting of the photon collimators. Relaiioe dose ratio: The variation of the relative dose for tube sizes 4,6 and 9 cm, utilizing 9 MeV electrons is shown in fig. 2 as a function of the photon collimator setting of the accelerator. As the collimator is opened for a setting just greater than the field size, the relative dose ratio rises rapidly at first and then flattens off to an asymptotic value. Clearly, the eolhmator cannot be set too narrowly because (a) the relative dose ratio becomes too small, so that the number of monitor units required to deliver a given dose becomes very large, and
.+-m
0.6i
l 5x 5cmz b *5xl5cm* a 35x35cm2
30 20
characterization
1103
radiotheqv
l
9cm
i0 t 0’
I
L
I
1
I
t
2
3
4
5
DEPTH IN POLYSTYRENE
I
fgmcm-PI
Fig. 3. % ionization curves for primary photon jaw settings of (0)5X5cm2,(a)15X15cm2 and(0)3fX35cm2 fora4cm tube at 9 MeV. (b) the uncertaintly in the relative dose ratio, and hence also the delivered dose, increases.
Percent depth dose: A comparison of the % depth-dose curves for three different collimator settings at 9 MeV for a 4 cm cone is shown in fig. 3. The curves are displayed only to the 50% dose level to demonstrate the effect of varying the primary collimator setting on the depth of the 90% dose. It can be seen that at 9 MeV there is a shift in the depth of the 90% ionization level by approximately 2 mm when changing from a 5 X 5 cm* setting to a 15 x 15 cm2 setting. It was found that for increasing energy and tube size, the effect of the
/
6cm
4cm
0.2 -
I
0
,
5
I
10
I
15
I
20
I
25
I
30
1
35
PRIMARY CDLLIMATOR SETTIM? fern J
Fig. 2. Graph of the relative dose ratio versus photon jaw setting for a 9 MeV electron beam. (8) 9 cm tube, (A) 6 cm cone and (Of 4 cm tube.
PRIMdRY
COLLMCd?W
SEfTMS
la/
Fig. 4. %
br~tr~~ng background as a junction of the photon jaw setting for a 4 cm tube.
XIV. RADIATION THERAPY/MEDICAL RESEARCH
1104
P.J. Biggs / Thephysics of ~~tr~~per~t~ve r~j~therapy
photon jaw setting on the % depth-dose curves diminished. Brcmssfrahlung: There is a dramatic effect of the primary collimator setting on the degree of bremsstrahlung background. Fig. 4 shows the per cent bremsstrahlung as a percentage of the ionization at d,,, for three energies 9, 12 and 15 MeV. for a 4 cm tube. At 9 MeV, the background is 35% for a primary collimator setting of 5 X 5 cm’. As the setting is increased to 20 X 20 cm*, this value is reduced to about 15%. The corresponding values for I5 and 23 MeV are 24% reducing to 14% and 15% reducing to 9%, respectively. Thus a primary collimator setting close to the field size gives an unacceptably large bremsstrahlung component to the dose curve, particularly at low energies. For larger tube sizes, the per cent bremsstrahlung decreases; generally it is about 10% or less. Summary: It can be seen from the foregoing results that the choice of primary collimator setting is not un~biguous. On the one hand, a setting just greater than the chosen field size gives the best depth-dose distribution and flattest beam profile (not shown), while on the other it provides relative dose ratios which are undesirably low and sensitive to collimator setting and results in a large bremsstrahlung tail. If the collimator setting is made too large, the beam profile becomes too rounded, ~nt~but~ng to a non-unifo~ dose distribution. Moreover, beyond a certain collimator size, the relative dose ratio approaches a constant value, resulting in no further increase in the effective electron dose rate. These considerations have led to the choice of a fixed photon jaw setting of 15 x 15 cm*. This is a compromise between the advantages and disadvantages outlined above; furthermore, it is larger than the maximum circular field size that could be feasibly used in IORT. 2.3. Surfacedose In our institution, the doses for intraoperative therapy are notably specified at the 90% isodose level, so it is desirable to have the surface dose equal to or greater than 90%, since the tumor is generally assumed to involve the surface. It is well known that the surface dose for electron beams increases with energy [1,2f; this has been verified by our observations. 2.4. Angtdar cones When treating lesions deep within the pelvis, it is often necessary to set the tube over the treatment area at a sharp angle with respect to the normal to the tissue surface. In these cases one has to be concerned about the decrease in penetrability of the beam. Fig. 5 shows the per cent ionization for a 7 cm tube at 9 MeV, inclined at various angles to the surface. The angle
-*-
Norma)
_-a__ .. . . 0
.
Incidence
20’
Incidmce
JO*
lncidrncs
45’
Incidence
F& d“‘..,
i P b
0 DEPTH
1
2 IN
3
POLYSTYRENE,
4 d
5 f am m-2
)
Fig. 5. % ionization curves for a 7 cm tube and 9 MeV electrons taken at normal incidence (O), 20’ incidence (X), 30” incidence (Cl)and 45O incidence (n) with respect to the normat to the surface.
shown is defined with respect to the normal to the surface. The depth plotted is along the central axis of the inclined beam. Between 0” and 20’ there is little change in the curve, but at 30” the depth of the 90% ionization point drops from 2.3 cm to 2.1 cm, while at 45” it is reduced to 1.7 cm. The effect was found to be greatest at the lowest energies and decreased with increasing energy 131. 2.5. Therm~~~mi~eseen~dosimetty (TLI>) An important aspect of the intraoperative procedure is an in oiuo verification of the dose delivered. For this purpose, lithium fluoride TLD extruded ribbons were used to measured the dose at the surface. Three such extruded ribbons, each measuring approximately 3 X 3 x 1 mm3, were placed close together in a short length of plastic tubing (< 3 cm long), the ends hermetically sealed with a hot iron and the assembly gas sterilized. Check measurements were made to ensure that there was no systematic error in this method of dose verification; the dose at the surface was compared to the dose measured with the TLD package when placed at the surface of a polystryrene phantom. The results agreed to within f 3%.
P.J.
I
ae
l
Eggs
6cm
/ The
l
physics
9cm c
7oL
6
10
14
ELECTRON
22
16 ENERGY
26
30
( MeV)
Fig. 6. Graph of surface dose versus electron energy for all circular tubes. (X) 4 cm, (A) 5 cm, (*) 6 cm, (0) 7 cm, (0) 8 cm and (+) 9 cm.
3.
Results
Fig. 6 shows the surface dose for all field sizes as a function of electron energy. As the electron energy was raised, the surface dose increased. For energies greater than 12 MeV, the surface dose was about 90%, while at 12, 9 and 6 MeV it was approximately 88%, 86% and 82% respectively. A comparison between the calculated dose, using this dosimetry system, and the in vivo measured dose, using the encapsulated TLDs, is shown in fig. 7. The average value of the in vivo measured dose was in reasonable agreement with the calculated dose. The standard deviation of the data was about f 6%, larger than the f 3% indicated earlier. This implied that the uncertainty in the dose calculation was greater than supposed, probably because of the clinical environment.
4. Discussion A compromise was selected in the setting for the primary collimator because of the varied effect it had on many of the beam characteristics. It was observed that the bremsstrahlung component of the per cent depth-dose curves increased as the field size set by the primary collimator was decreased. This effect has been
21 Cases
NO. OF CASES
Mean Value = 1.01 t
[CALCULATED
8
DOSE/TLD
S.D. = 0.06
READING]
Fig. 7. Comparison between the calculated dose and the dose as measured by TLD.
of inrraoperutioe
radiotherapy
1105
observed before [4] and the explanation is related to the thickness of the scattering foils. For a field size of 4 cm, the election beam is collimated by the photon jaws to an area subtending an angle of approximately 2” relative to the scattering foil. Thus, a large fraction of electrons scattered from the fiied collimators that would otherwise contribute to the dose is eliminated and the electron dose per monitor unit is strongly reduced. Thus, the electron dose per monitor unit, and hence the relative dose ratio, will increase as the primary collimator is widened. This explains the photon jaw setting dependence shown in fig. 2. The fact that the relative dose ratios approach a constant value beyond a certain photon jaw setting probably indicates that the fixed collimators (see fig. 1) are completely exposed at that point. This is borne out by the fact that the curves for the smallest tubes saturate before the larger ones do. For the same reason, the relative dose ratio will increase with secondary collimator setting, since they permit more scattered electrons to contribute to the dose. The same rationale also explains the energy dependence of the relative dose ratio. As the electron energy is raised, the electrons scattered by the lead foil emerge more in the forward direction, resulting in a greater contribution to the dose. For example, the root mean square scattering angles for 9 and 23 MeV electrons for a 0.6 mm lead foil are 31.5’ and 12.3” respectively. However, since the bremsstrahlung tail produced in the target is extremely forward peaked, it does not depend upon field size. Thus, as the field size is reduced to a very small area, the percentage of bremsstrahlung increases for a given energy. The scattering system in this linear accelerator is of the single foil type with the foil thicknesses chosen by the manufacturer to achieve as flat a dose distribution as possible for a 30 X 30 cm’ field size. By using a double foil scattering system, optimized to obtain a flat dose distribution of a 30 X 10 cm2 field, for example, the thickness of the foils could be greatly reduced [5]. This in turn will increase the effective electron dose rate and reduce the amount of bremsstrahlung background. ’ It is well known that high energy electron beams flattened by scattering foils have a much shallower depth for the 90% isodose level than those beams of the same initial energy that use a scanning magnet to achieve the same field flatness [6). Therefore, any attempt to reduce the thickness of the scattering foils will also help improve the quality of the electron beam by increasing the depth of the 90% isodose level for a fixed electron energy. If treatment to a particular depth is required, a lower energy beam can be used, thereby sparing underlying normal tissue. An example of the importance of this is the treatment for unresectable carcinoma of the head of the pancreas as shown in fig. 8. In this case the spinal cord or one of the kidneys will probably be in the radiation field. Although the density of the vertebral XIV. RADIATION THERAPY/MEDICAL RESEARCH
1106
P. J. Biggs / 7?te physics of intraoperative
INTRAOPERATIVE ELECTRON BEAM IRRADIATION Curchomu, Head of Pancreas 7cm LUCITE CONE 23MeV
radiotherapy
tude of which has implications for the amount of room shielding. The thicker the foil, the greater the per cent bremsstrahlung in the tail, so a greater amount of shielding is required.
5. Conclusions
The intraoperative program has achieved a high degree of success to date. The radiotherapists have expressed satisfaction with the way in which this trial has proceeded, and the surgeons and anesthesiologists have considerable confidence in the general surgical safety of the procedure. As clinical data is accumulated, the efficacy of IORT as a curative procedure will be assessed. Fig. 8. Dose distribution for treatment of carcinoma of head of the pancreas using a 23 MeV electron beam.
References
bodies will reduce the dose to most of the cord to a relatively low value, this could still be quite high at the level of the intervertebral disk. Thus, an improvement in the quality of the electron beam by reducing foil thickness or by employing the scanning beam itself will reduce the dose to these normal structures. An additional consequence of the thickness of the scattering foil lies in the bremsstrahlung tail, the magni-
[1] J.W. Boag, Br. J. Radio]. 45 (1972) 229. [2] S.C. Lilhcrap and M. Rosenbloom, Br. J. Radiol. 45 (1972) 229. (31 P.J. Biggs, Phys. Med. Biol. 9 (1984) 1089. [4] P.J. Biggs and C.C. Ling, Med. Phys. 6 (1979) 291. [5] H.E. Johns and J.A. Rawlinson, High Energy Photons and Electrons, eds., S. Kramer, N. Suntharalingum and G.F. Zinninger (Wiley, New York, 1976) p. 5. (61 P. Almond. Clinical Applications of the Electron Beam, ed., N.V. Tapley (Wiley, New York, 1976) p. 7.
THE PHYSICS OF I~OPE~~E BEAMS
Nuclear Instruments and Methods in Physics Research.BlO/ll (1985) 1102-1106 North-Holland, Amsterdam
~DIO~FY
USING HIGH ENERGY ELECI-IZON
Peter J. BIGGS Department of Radiafion Medicine, Mawachusetts General Hospiral: Hatvatd Medical S&a&, Bosfon, Massachusetrs 021Id, USA
Conventional radiotherapy involves the treatment of malignant tumors by ionizing radiation applied externally to the body. However, injury to internal organs and normal tissue often limits the tumor dose delivered to less than cancericidal values. This problem is circumvented in intraoperative radiotherapy, where the tumor is surgicaby exposed, normal tissue and organs displaced, and the diseased area treated directly with radiation. Electron beams are used for these procedures because of their finite range compared with X-rays, to spare normal tissue below the tumor.
1.
Introduction
Most major advances in the treatment of cancer with radiation have been based on the differences which can be obtained in the distribution of the radiation dose between the tumor and adjacent normal tissues. The probabi~ty of achieving locat control of a localized tumor mass generally increases as larger radiation doses are delivered to the tumor. However, in many clinical situations the dose that can safely be delivered to the tumor is restricted by the limited radiation tolerance of su~oundi~ normal tissues. The rationale for intraoperative radiation therapy (TORT) is to expose the tumorbearing structure and apply the radiation directly to the tumor, thus avoiding irradiation of normal structures. It is the purpose of this paper to describe the physical characteristics of the electron beams used to deliver radiation doses iuteroperativeiy to the tumor-bearing areas. Such characteristics concern the % depth-dose distributions, effect of field angulation, surface dose and bremsstra~ung background.
linear accelerator. The purpose of the adjustable collimator was to eliminate the need for a large number of fixed apertures. A thin (approximately 0.25 mm) transparent mylar sheet was fastened directly over the jaw aperture to prevent the possibility of any stray material from the machine falling directly into the patient. The field in the patient is defined by a transparent lucite tube that slides into an aluminum jacket, which in turn is rigidly attached to the secondary collimator. To align the treatment field with the electron beam, the lucite tube has to be “docked” with the aluminum jacket. This is achieved by rotating the gantry and moving the treatment couch until the head of the tube is aligned with and almost touching the metal jacket. The
ionization Chamber
-
PRIMARY
2. Methods and materials
COLLIMATORS
SECONDARY COLLIMATORS
2. I. Appara fw and inttaopeta~~ve procedure
The linear accelerator used for this procedure has a wide range of electron energies available, namely 6, 9, 12, 15, 18, 23 and 29 MeV, corresponding to a range of 1.7 to 6.1 cm in water at the 90% depth dose. The equipment is shown schematically in fig. 1. In addition to the fixed collimators and photon jaws (primary collimators) of the machine a set of adjustable jaws (secondary collimators) was used to define the electron field size at the surface. The adjustable jaws are attached to a plate that is bolted to the front face of the 0168-583X/85/$03.30 @ Elsevier Science Publishers B.V. (North-Holland Physics Publishing Division)
Trrotmmt
I
Tuba
I
Nottoscolo Fig. 1. GeneraI ~IIi~tion layout of the Iinear accelerator for intraoperative radiotherapy.
P.J. Siggs / The physics of intraoperative
couch is then moved in very short steps so as to ease the tube up into the jacket until it has reached the desired depth. The majority of the tubes are circular in cross section, ranging in inner diameter from 4 cm to 9 cm in steps of 1 cm. Many of these are bevelied. Wrapped around the top of each tube there is a 5 cm long, Delrin cylinder which slides into the metal jacket. This cylinder is of variable wall thickness, so that all the cones may slide into the same metal jacket. The gap between the cyhnder and the jacket was made to be a ~mpro~se between an easy fit and minimal lateral displacement of the tube at its lower end. Typically, the gap is between 0.05 mm and 0.075 mm and the lateral displacement is + (l-2) mm. 2.2. Beam
0.8
CONE SIZE
9Mev 0.7
s 3
0.5 -
8
0.4 -
h 2i=
0.3 -
1
~~
An extensive study was made of the relative dose delivered through each tube, the per cent depth dose, beam profiles, surface dose, and the per cent bremsstrahlung background to determine the optimum setting of the photon collimators. Relaiioe dose ratio: The variation of the relative dose for tube sizes 4,6 and 9 cm, utilizing 9 MeV electrons is shown in fig. 2 as a function of the photon collimator setting of the accelerator. As the collimator is opened for a setting just greater than the field size, the relative dose ratio rises rapidly at first and then flattens off to an asymptotic value. Clearly, the eolhmator cannot be set too narrowly because (a) the relative dose ratio becomes too small, so that the number of monitor units required to deliver a given dose becomes very large, and
.+-m
0.6i
l 5x 5cmz b *5xl5cm* a 35x35cm2
30 20
characterization
1103
radiotheqv
l
9cm
i0 t 0’
I
L
I
1
I
t
2
3
4
5
DEPTH IN POLYSTYRENE
I
fgmcm-PI
Fig. 3. % ionization curves for primary photon jaw settings of (0)5X5cm2,(a)15X15cm2 and(0)3fX35cm2 fora4cm tube at 9 MeV. (b) the uncertaintly in the relative dose ratio, and hence also the delivered dose, increases.
Percent depth dose: A comparison of the % depth-dose curves for three different collimator settings at 9 MeV for a 4 cm cone is shown in fig. 3. The curves are displayed only to the 50% dose level to demonstrate the effect of varying the primary collimator setting on the depth of the 90% dose. It can be seen that at 9 MeV there is a shift in the depth of the 90% ionization level by approximately 2 mm when changing from a 5 X 5 cm* setting to a 15 x 15 cm2 setting. It was found that for increasing energy and tube size, the effect of the
/
6cm
4cm
0.2 -
I
0
,
5
I
10
I
15
I
20
I
25
I
30
1
35
PRIMARY CDLLIMATOR SETTIM? fern J
Fig. 2. Graph of the relative dose ratio versus photon jaw setting for a 9 MeV electron beam. (8) 9 cm tube, (A) 6 cm cone and (Of 4 cm tube.
PRIMdRY
COLLMCd?W
SEfTMS
la/
Fig. 4. %
br~tr~~ng background as a junction of the photon jaw setting for a 4 cm tube.
XIV. RADIATION THERAPY/MEDICAL RESEARCH
1104
P.J. Biggs / Thephysics of ~~tr~~per~t~ve r~j~therapy
photon jaw setting on the % depth-dose curves diminished. Brcmssfrahlung: There is a dramatic effect of the primary collimator setting on the degree of bremsstrahlung background. Fig. 4 shows the per cent bremsstrahlung as a percentage of the ionization at d,,, for three energies 9, 12 and 15 MeV. for a 4 cm tube. At 9 MeV, the background is 35% for a primary collimator setting of 5 X 5 cm’. As the setting is increased to 20 X 20 cm*, this value is reduced to about 15%. The corresponding values for I5 and 23 MeV are 24% reducing to 14% and 15% reducing to 9%, respectively. Thus a primary collimator setting close to the field size gives an unacceptably large bremsstrahlung component to the dose curve, particularly at low energies. For larger tube sizes, the per cent bremsstrahlung decreases; generally it is about 10% or less. Summary: It can be seen from the foregoing results that the choice of primary collimator setting is not un~biguous. On the one hand, a setting just greater than the chosen field size gives the best depth-dose distribution and flattest beam profile (not shown), while on the other it provides relative dose ratios which are undesirably low and sensitive to collimator setting and results in a large bremsstrahlung tail. If the collimator setting is made too large, the beam profile becomes too rounded, ~nt~but~ng to a non-unifo~ dose distribution. Moreover, beyond a certain collimator size, the relative dose ratio approaches a constant value, resulting in no further increase in the effective electron dose rate. These considerations have led to the choice of a fixed photon jaw setting of 15 x 15 cm*. This is a compromise between the advantages and disadvantages outlined above; furthermore, it is larger than the maximum circular field size that could be feasibly used in IORT. 2.3. Surfacedose In our institution, the doses for intraoperative therapy are notably specified at the 90% isodose level, so it is desirable to have the surface dose equal to or greater than 90%, since the tumor is generally assumed to involve the surface. It is well known that the surface dose for electron beams increases with energy [1,2f; this has been verified by our observations. 2.4. Angtdar cones When treating lesions deep within the pelvis, it is often necessary to set the tube over the treatment area at a sharp angle with respect to the normal to the tissue surface. In these cases one has to be concerned about the decrease in penetrability of the beam. Fig. 5 shows the per cent ionization for a 7 cm tube at 9 MeV, inclined at various angles to the surface. The angle
-*-
Norma)
_-a__ .. . . 0
.
Incidence
20’
Incidmce
JO*
lncidrncs
45’
Incidence
F& d“‘..,
i P b
0 DEPTH
1
2 IN
3
POLYSTYRENE,
4 d
5 f am m-2
)
Fig. 5. % ionization curves for a 7 cm tube and 9 MeV electrons taken at normal incidence (O), 20’ incidence (X), 30” incidence (Cl)and 45O incidence (n) with respect to the normat to the surface.
shown is defined with respect to the normal to the surface. The depth plotted is along the central axis of the inclined beam. Between 0” and 20’ there is little change in the curve, but at 30” the depth of the 90% ionization point drops from 2.3 cm to 2.1 cm, while at 45” it is reduced to 1.7 cm. The effect was found to be greatest at the lowest energies and decreased with increasing energy 131. 2.5. Therm~~~mi~eseen~dosimetty (TLI>) An important aspect of the intraoperative procedure is an in oiuo verification of the dose delivered. For this purpose, lithium fluoride TLD extruded ribbons were used to measured the dose at the surface. Three such extruded ribbons, each measuring approximately 3 X 3 x 1 mm3, were placed close together in a short length of plastic tubing (< 3 cm long), the ends hermetically sealed with a hot iron and the assembly gas sterilized. Check measurements were made to ensure that there was no systematic error in this method of dose verification; the dose at the surface was compared to the dose measured with the TLD package when placed at the surface of a polystryrene phantom. The results agreed to within f 3%.
P.J.
I
ae
l
Eggs
6cm
/ The
l
physics
9cm c
7oL
6
10
14
ELECTRON
22
16 ENERGY
26
30
( MeV)
Fig. 6. Graph of surface dose versus electron energy for all circular tubes. (X) 4 cm, (A) 5 cm, (*) 6 cm, (0) 7 cm, (0) 8 cm and (+) 9 cm.
3.
Results
Fig. 6 shows the surface dose for all field sizes as a function of electron energy. As the electron energy was raised, the surface dose increased. For energies greater than 12 MeV, the surface dose was about 90%, while at 12, 9 and 6 MeV it was approximately 88%, 86% and 82% respectively. A comparison between the calculated dose, using this dosimetry system, and the in vivo measured dose, using the encapsulated TLDs, is shown in fig. 7. The average value of the in vivo measured dose was in reasonable agreement with the calculated dose. The standard deviation of the data was about f 6%, larger than the f 3% indicated earlier. This implied that the uncertainty in the dose calculation was greater than supposed, probably because of the clinical environment.
4. Discussion A compromise was selected in the setting for the primary collimator because of the varied effect it had on many of the beam characteristics. It was observed that the bremsstrahlung component of the per cent depth-dose curves increased as the field size set by the primary collimator was decreased. This effect has been
21 Cases
NO. OF CASES
Mean Value = 1.01 t
[CALCULATED
8
DOSE/TLD
S.D. = 0.06
READING]
Fig. 7. Comparison between the calculated dose and the dose as measured by TLD.
of inrraoperutioe
radiotherapy
1105
observed before [4] and the explanation is related to the thickness of the scattering foils. For a field size of 4 cm, the election beam is collimated by the photon jaws to an area subtending an angle of approximately 2” relative to the scattering foil. Thus, a large fraction of electrons scattered from the fiied collimators that would otherwise contribute to the dose is eliminated and the electron dose per monitor unit is strongly reduced. Thus, the electron dose per monitor unit, and hence the relative dose ratio, will increase as the primary collimator is widened. This explains the photon jaw setting dependence shown in fig. 2. The fact that the relative dose ratios approach a constant value beyond a certain photon jaw setting probably indicates that the fixed collimators (see fig. 1) are completely exposed at that point. This is borne out by the fact that the curves for the smallest tubes saturate before the larger ones do. For the same reason, the relative dose ratio will increase with secondary collimator setting, since they permit more scattered electrons to contribute to the dose. The same rationale also explains the energy dependence of the relative dose ratio. As the electron energy is raised, the electrons scattered by the lead foil emerge more in the forward direction, resulting in a greater contribution to the dose. For example, the root mean square scattering angles for 9 and 23 MeV electrons for a 0.6 mm lead foil are 31.5’ and 12.3” respectively. However, since the bremsstrahlung tail produced in the target is extremely forward peaked, it does not depend upon field size. Thus, as the field size is reduced to a very small area, the percentage of bremsstrahlung increases for a given energy. The scattering system in this linear accelerator is of the single foil type with the foil thicknesses chosen by the manufacturer to achieve as flat a dose distribution as possible for a 30 X 30 cm’ field size. By using a double foil scattering system, optimized to obtain a flat dose distribution of a 30 X 10 cm2 field, for example, the thickness of the foils could be greatly reduced [5]. This in turn will increase the effective electron dose rate and reduce the amount of bremsstrahlung background. ’ It is well known that high energy electron beams flattened by scattering foils have a much shallower depth for the 90% isodose level than those beams of the same initial energy that use a scanning magnet to achieve the same field flatness [6). Therefore, any attempt to reduce the thickness of the scattering foils will also help improve the quality of the electron beam by increasing the depth of the 90% isodose level for a fixed electron energy. If treatment to a particular depth is required, a lower energy beam can be used, thereby sparing underlying normal tissue. An example of the importance of this is the treatment for unresectable carcinoma of the head of the pancreas as shown in fig. 8. In this case the spinal cord or one of the kidneys will probably be in the radiation field. Although the density of the vertebral XIV. RADIATION THERAPY/MEDICAL RESEARCH
1106
P. J. Biggs / 7?te physics of intraoperative
INTRAOPERATIVE ELECTRON BEAM IRRADIATION Curchomu, Head of Pancreas 7cm LUCITE CONE 23MeV
radiotherapy
tude of which has implications for the amount of room shielding. The thicker the foil, the greater the per cent bremsstrahlung in the tail, so a greater amount of shielding is required.
5. Conclusions
The intraoperative program has achieved a high degree of success to date. The radiotherapists have expressed satisfaction with the way in which this trial has proceeded, and the surgeons and anesthesiologists have considerable confidence in the general surgical safety of the procedure. As clinical data is accumulated, the efficacy of IORT as a curative procedure will be assessed. Fig. 8. Dose distribution for treatment of carcinoma of head of the pancreas using a 23 MeV electron beam.
References
bodies will reduce the dose to most of the cord to a relatively low value, this could still be quite high at the level of the intervertebral disk. Thus, an improvement in the quality of the electron beam by reducing foil thickness or by employing the scanning beam itself will reduce the dose to these normal structures. An additional consequence of the thickness of the scattering foil lies in the bremsstrahlung tail, the magni-
[1] J.W. Boag, Br. J. Radio]. 45 (1972) 229. [2] S.C. Lilhcrap and M. Rosenbloom, Br. J. Radiol. 45 (1972) 229. (31 P.J. Biggs, Phys. Med. Biol. 9 (1984) 1089. [4] P.J. Biggs and C.C. Ling, Med. Phys. 6 (1979) 291. [5] H.E. Johns and J.A. Rawlinson, High Energy Photons and Electrons, eds., S. Kramer, N. Suntharalingum and G.F. Zinninger (Wiley, New York, 1976) p. 5. (61 P. Almond. Clinical Applications of the Electron Beam, ed., N.V. Tapley (Wiley, New York, 1976) p. 7.