- Email: [email protected]
Inverse planning of energy-modulated electron beams in radiotherapy
Medical Dosimetry, Vol. 31, No. 4, pp. 259-268, 2006 Copyright © 2006 American Association of Medical Dosimetrists Printed in the USA. All rights reserved 0958-3947/06/$–see front matter
doi:10.1016/j.meddos.2005.03.008
INVERSE PLANNING OF ENERGY-MODULATED ELECTRON BEAMS IN RADIOTHERAPY JOHN R. GENTRY, M.S., RICHARD STEEVES, M.D., PH.D., and BHUDATT A. PALIWAL, PH.D. Department of Human Oncology, Comprehensive Cancer Care Center, University of Wisconsin-Madison, Madison, WI 53792 (Received 15 November 2004; accepted 14 March 2005)
Abstract—The use of megavoltage electron beams often poses a clinical challenge in that the planning target volume (PTV) is anterior to other radiosensitive structures and has variable depth. To ensure that skin as well as the deepest extent of the PTV receives the prescribed dose entails prescribing to a point beyond the depth of peak dose for a single electron energy. This causes dose inhomogeneities and heightened potential for tissue fibrosis, scarring, and possible soft tissue necrosis. Use of bolus on the skin improves the entrant dose at the cost of decreasing the therapeutic depth that can be treated. Selection of a higher energy to improve dose homogeneity results in increased dose to structures beyond the PTV, as well as enlargement of the volume receiving heightened dose. Measured electron data from a linear accelerator was used as input to create an inverse planning tool employing energy and intensity modulation using bolus (e-IMRT™). Using tools readily available in a radiotherapy department, the applications of energy and intensity modulation on the central axis makes it possible to remove hot spots of 115% or more over the depths clinically encountered. The e-IMRT™ algorithm enables the development of patient-specific dose distributions with user-defined positions of peak dose, range, and reduced dose to points beyond the prescription point. © 2006 American Association of Medical Dosimetrists. Key Words: Electron energy/intensity modulation, Inverse treatment Planning, Breast cancer.
lower density of the lung and therefore extended range of megavoltage electrons in this medium. To offset this problem, effort is often devoted to the construction of a customized bolus that will ensure the prescribed dose is delivered to the skin and that the electrons deliver the prescribed dose to the distal edge of the tumor and expend as much of their energy as possible before crossing the chest wall/lung interface. Such a customized bolus can be created based on a slice-by-slice computed tomography (CT) analysis of the chest wall thickness throughout the treatment portal. Use of the bolus improves the dose distribution, but may create significant dose heterogeneities near the skin surface and in the region of peak dose. While improving the tumor control probability, boost phase therapy with electrons is also associated with higher rates of fibrosis. Fibrosis, breast hardness, and edema within the irradiated breast are sources of discomfort for women undergoing conservative treatment for breast cancer. Campana et al.1 cites the release of free radicals, such as O2 and hydroxyls, as the cause of breast fibrosis. The strong instability of these molecules and their tendency to bind to adjacent structures causes injuries to the connective tissue and vascular networks. In addition to the pain, short- and long-term cosmetic implications are involved. Lymph node aspiration further compromises the local vascular system, contributing to marked or even severe skin reactions and edema.2 Age and weight,2 radiation dose,3,4 dose fractionation,5 and volume effects,6 as well as the overall treatment duration1 of the
INTRODUCTION The depth dose and practical range characteristics of megavoltage electron beams make them a natural choice in the treatment of superficial cancer both as primary therapy, in the case of a skin lesion, or as the boost in the treatment of breast cancer to the region overlying the excision in breast-conserving treatment. In the latter group, the residual chest wall and, in some patients, the internal mammary nodes as well must be treated with electrons to minimize dose to lung and heart. The necessity of delivering the prescribed dose to the skin and at the depth of the chest wall/lung interface while simultaneously minimizing the dose to the lung is a treatment framework in which competing goals exist and trade-offs are required. Treatment, in the boost phase of conservative breast therapy, often consists of a single electron energy prescribed to the chest wall/lung interface. Bolus is prescribed to deliver the treatment dose to a depth no greater than the chest wall or the depth indicated by clips left at the time of surgery to denote the depth of the tumor bed. Because the depth of prescription is typically deeper than the depth of maximum dose, in these cases, dose heterogeneities are created at the depth of peak dose of 110% or more. Care must be exercised in the choice of electron beam energy when treating the chest because of the Reprint requests to: John R. Gentry, M.S., Caromont Cancer Care Center, Gaston Memorial Hospital, 2525 Court Drive, Gastonia, NC 28054. E-mail: [email protected]. 259
260
Medical Dosimetry
radiotherapy contribute to the severity of these radiationrelated injuries. Recent studies at Massachusetts General Hospital further establish the dose volume effect in these patients, where 4 of 48 women experienced moderate to severe thickening of the skin.7 Vrieling et al.8 point specifically at the boost as an important factor in creation of breast fibrosis. They concluded that to achieve a good cosmesis, it is necessary to excise the tumor with a limited margin, to avoid postoperative complications, to assess the need for a boost in the individual patient, and to give the radiation dose as homogeneously as possible. Buchholz9 lends additional support for this line of thought and suggests that the level of dose homogeneity in the lower quadrant of the breast resulting from the photon treatment should be taken into account when planning the application of the boost electron treatment, rather than be ignored. A reasonable way to decrease dose inhomogeneity in the breast boost and reduce the dose to critical structures beyond the depth of prescription would be to deliver the dose by 2 or more electron beam energies, such that an optimized part of the prescribed dose is delivered by each energy. Optimized treatment solutions requiring the use of multiple electron beam energies have taken principally 1 of 2 paths. Some (Ma et al.,10 Karlsson et al.,11 Lee et al.,12 have taken the path of modulating both energy and intensity via a special electron multileaf collimator (MLC) and utilize Monte Carlo-based inverse treatment planning tools. Specifically, Hyödynmaa et al.13 proposed a technique for chest wall irradiation that optimized dose homogeneity via multiple electron beams by varying fluence profiles and energies. Their technique optimizes the dose distribution in 3 dimensions, and is capable of using as many as 5 beams. Subject to the constraints imposed, their computer found the iterative solution in 7 to 17 minutes. Though the tools and methods of this path are promising and represent the future for conformal electron beam therapy, they are currently not widely available. The second path is epitomized by those who have focused on the use of more readily accessible computational and hardware tools. They (Klein,14 Kudchaker et al.,15 Lief et al.,16 Blomquist et al.17) utilize an MLC, traditionally reserved for the shaping of the port for photon beam treatment with the dose intensity pattern generated by an inverse treatment planning system. Delivery of treatment using this method is not available in the clinical mode for most accelerators. It would be advantageous to offer a straightforward treatment solution that results in improved dose distributions, reduced dose to structures beyond the depth of prescription, and improves the ability of existing tools to deliver conformal electron radiotherapy. This paper will present the Trumpet e-IMRT™ Pre-Planning Software, a suite of planning tools for creating composite electron beam depth dose distributions.
Volume 31, Number 4, 2006
METHODS Mathematical algorithm for energy and intensity modulated radiotherapy In general, the dose D at any depth x in tissue subject to N normally incident megavoltage electron beams can be expressed as D共xn兲 ⫽ MU1FDD1共x兲共OAR1共x兲兲共INVSQ1兲 ⫹ MU2FDD2共x兲共OAR2共x兲兲共INVSQ2兲 ⫹ MU3FDD3共x兲共OAR3共x兲兲共INVSQ3兲 ⫹ . . . ⫹ MUN FDDN共x兲共OARN共x兲兲共INVSQN兲 (1) where MU, FDD, x, OAR, and INVSQ are monitor units, fractional depth dose, depth, off-axis ratio, and inverse square factor associated with the total dose at x, respectively. For this discussion, central axis dose with the patient at 100-cm source-to-surface distance (SSD) and field size large enough to ensure electronic equilibrium are considered exclusively. An algorithm has been developed that computes the number of MUs that should be delivered by each of the electron beam energies such that the resulting composite dose distribution meets userdefined constraints expressed as n
兺 MUiFDDi共xj兲共OARi共xj兲兲共INVSQ兲 ⫺ D共xj兲constraint ⫽ 0 i⫽1 j⫽1
(2) Visual evaluation of composite depth dose distributions In the basic application of the inverse planning concept within the Pre-Planning Software, the user specifies the dose to the skin and to a second depth within the patient. The algorithm determines the proper weighting of MUs and bolus to achieve the dose at the skin and at the second point. The set of solutions returned by the first pass of the program is further reduced by the operator through specification of the appropriate range that will be allowed. Then, application of dose homogeneity requirements further reduces the number of treatment solutions leaving the more optimal treatment solutions. A graphical interface was created to show the fractional depth dose characteristics of the composite beam relative to 2 standard beams for comparison. The algorithm’s results were then tested against the dose distribution computed by a Prowess 3000 treatment planning system utilizing the Memorial Sloan Kettering Electron Beam Computation Model for treatment on a Varian 2100 C linear accelerator and Pinnacle3 version 7.0 (ADAC Philips). The dose engine for the electron tool within Pinnacle is the Hogstrom Pencil Beam, which makes use of a combination of measured data and mathematical modeling to characterize each electron beam. The solutions to the constraint equations provide both physical as well as non-physical electron depth dose distributions. A graphical tool was introduced to show
Energy-modulated electron beams ● J. R. GENTRY et al.
261
Fig. 1. Graphical tool shows the location (shaded region) of the possible composite distributions with acceptable solutions to the system of linear equations given a set of dose constraints on the dose at the surface, in this case 1 Gy. The 12-MeV beam is bolused and the resulting dose distribution is shown. The composite dose distribution gives a homogeneous dose from the surface to 1.6-cm depth. This distribution is intermediate between a 6-MeV beam and a 9-MeV beam. The size and shape of the solution space varies with any change in the bolus used or constraints imposed. Given the requirement that a dose of 1 Gy is to be delivered at the skin surface in the figure above, a real composite distribution can be created which has a depth dose of 0.80 at 30 mm, but not 0.1 at 30 mm, because that point lies outside the shaded region.
the achievable solution sets (Figs. 1 and 2). Fig. 1 shows a composite dose distribution created by mixing 9 and 12 MeV electron beams (with bolus) that has depth dose characteristics that are intermediate between 6 MeV and 9 MeV. Several test cases using 9 MeV and 12 MeV
electron beams are presented to show the utility of this technique to deliver a dose distribution over a tumor volume to a depth beyond that of peak dose. The Pre-Planning Software algorithm computes the central axis composite depth dose for the electron beam
Fig. 2. Graphical illustration of the effect of bolus on each beam and the composite beam created by their addition. The shape of the solution space changes with the dose specification and the quantity of bolus each beam receives. A graphical tool reveals quickly the limits of any potential composite dose distribution that is achievable with the electron beam energies and bolus being used.
262
Medical Dosimetry
Volume 31, Number 4, 2006
Fig. 3. A composite dose distribution composed of 9- and 12 MeV electron beams shown beside its component 9- and 12-MeV beams. The dose heterogeneity was 110% of the prescription. The composite beam delivers less dose beyond the prescription point (3 cm) than the 12-MeV beam alone and a smaller practical range. The composite beam gives a dose of 0.61 Gy at 4 cm compared to 0.83 Gy from the 12 MeV beam alone. The monitor units calculated represent the dose which must be delivered to the depth of peak dose for each energy in order to achieve the prescribed dose. Error bars represent 2 standard deviations.
energies, MUs and bolus being used. The algorithm’s calculation of the dose on the central axis was tested against ion chamber in a typical water-equivalent phantom utilizing a 100-cm SSD. Bolus, when necessary, was placed on top of the phantom. Open cones were used exclusively for this testing. RESULTS The depth dose beam characteristics of a composite beam technique were compared to those of 2 intermediate megavoltage electron beams (9 MeV and 12 MeV) for treatment of the internal mammary nodes at a depth of 3 cm. The composite beam creation was guided by the need to create a dose distribution that optimized dose homogeneity while limiting dose to structures beyond the depth of 3 cm. Figure 3 shows the composite beam created as well as the 9- and 12-MeV beams. The composite beam was achieved by adding 0.6-cm bolus to the 12-MeV beam and 0.4 cm of bolus to the 9-MeV beam. The composite beam delivers 0.63 Gy at a depth 1 cm beyond the prescription point as compared to the 0.83 Gy using the 12-MeV beam alone. The composite beam is superior to the 9 MeV in dose homogeneity. The composite beam delivers less dose beyond the prescription point (3 cm) than the 12-MeV beam alone and has a smaller practical range. The composite beam gives a dose of 0.61 Gy at 4 cm compared to 0.83 Gy from the 12-MeV beam alone. The position of peak dose for the
distribution is found at a depth of 1.6 cm as compared to the 1.8-cm peak dose depth for the 9-MeV beam. The 50% fractional depth dose of the composite beam is extended 0.9 cm beyond the 3.3-cm position of the 9-MeV beam but is more shallow than the 12-MeV beam (4.8 cm) by 0.6 cm. The practical range of the composite beam is approximately 54 mm, giving it an effective energy that is close to 11 MeV, but energy as designated by the distance needed for the fractional depth dose to be reduced to 50% of the peak dose (R50) depth is 9.9 MeV. Figure 4 shows how the depth dose of the composite beam compares to the 2 individual beams over the first 10 mm of skin. Thus, a judicious choice of bolus remains an integral part of creating the optimal composite beam with which to treat the patient. Simply combining the beams in achieve the prescribed dose at the surface and at 3 cm yielded a composite beam (Fig. 5) that delivered less dose beyond the prescription depth, but displayed poor dose homogeneity (hot spot of 115%). The fractional depth dose within the first centimeter of a typical patient is shown in Fig. 6. While this distribution meets the dose constraints, this distribution displays less homogeneity than Fig. 3. In addition to improving dose homogeneity and decreasing dose distal to the prescription point, the algorithm facilitates the evaluation of a continuum of composite beams that have user-specified practical ranges
Energy-modulated electron beams ● J. R. GENTRY et al.
263
Fig. 4. The dose distribution of the composite and single energy electron beams (9 and 12 MEV) are shown. As a result of bolus being added to the skin for both the 9- and 12-MeV beams to form the composite beam, the entrant skin fractional depth dose of the composite beam meets the dose prescription. The increase in dose over the first 10 mm is less than 105% of the prescription.
and homogeneities within the patient. It is demonstrated in Fig. 7 that the combination of a 9-MeV beam (0.5-cm bolus, 100-cm SSD, 9 cGy delivered to the depth of peak dose) with a 12-MeV beam (1.3-cm bolus, 100-SSD, 94 cGy delivered to the depth of peak dose) yields a dose distribution that is very similar to 9 MeV, but, has significantly improved dose homogeneity. Figure 8 shows that dose distributions that have practical ranges smaller than the lowest energy component can be created that have
superior homogeneity. In this case, the composite distribution has less than 105% dose heterogeneity over a distance of 1.3 cm, comparable to the depth of peak dose for a 6-MeV electron beam. Lateral Tumor Coverage: Phantom Results The Pre-Planning Software was tasked to compute a dose distribution (Figure 9a, 9b, and 9c) for a 10x10 cm reference field that was more homogeneous than a 9
Fig. 5. A composite beam created with dose specification of 1 Gy at the surface and at depth (30 mm) in the patient. Both 9- and 12-MeV beams are shown for comparison purposes. No bolus is used in this example. Without bolus, the composite beam has dose homogeneity characteristics similar to the 9-MeV beam. This distribution contains a dose heterogeneity of 115%. The practical range is less than that of the 12-MeV beam and the integral dose delivered past the point of the prescription is decreased in comparison to the 12-MeV beam.
264
Medical Dosimetry
Volume 31, Number 4, 2006
Fig. 6. Without adding bolus to either beam, the composite beam’s entrance fractional depth dose was 11% higher than the 12-MeV beam, but the increase in the fractional dose per millimeter was similar to the 9-MeV beam over the first 1-cm depth. Without the use of bolus, the dose heterogeneity is somewhat larger.
MeV beam down to a depth of 2 cm, and provided good coverage by the 95% or higher isodose line, and had a range similar to the 9 MeV beam. In less than a minute, the program found 24 methods for creating the desired dose distribution by varying the number of mu and bolus thicknesses to be used with each energy. The mixed energy solution is shown in the upper row of panes and the single energy 9 MeV beam is shown in the lower row. Coverage by the 95% line (blue) is available at the surface (Fig 9a), and, after 2.5 mm, the 98% isodose line (red) covers a significant fraction of the mixed energy field. Figure 9b shows the lateral extent of the 98% isodose line (red) to a depth of 2 cm as computed by the Pinnacle (Phillips ADAC) planning system. The Pre-Planning Software computed solution brought a useable prescription isodose line to the surface while retaining the range characteristics of the 9 MeV beam. This is in contrast to the single energy 9 MeV beam which only displays the prescence of the 85% isodose (olive) at 0.5 cm. Clinical Application: Patient-Chest Mass A solution for a patient presenting with a chest mass was found to be a prescription using both a mixed
electron beam and a photon beam. For the electron portion of the treatment, the prescription was for 12 MeV with 2.6 cm of bolus combined with 16 MeV and 0.5 cm of bolus. The dose was normalized to the 95% isodose line to provide the best coverage. The patient then received 7 fractions of 2.5 Gy with the mixed electron beam and 7 fractions of 2.5 Gy with photons. This solution is compared to a single energy 16 MeV electron and 6 MV photon treatment in Fig 10. The 3578 cGy treatment dose is shown by the red isodose line in both pictures. The treatment dose delivered by the Pre-Planning technique encompasses the tumor, shown in red, and then falls off more rapidly thereafter, sparing lung and heart to a greater degree than the 6 MV photon/ 16 MeV single electron treatment prescribed to the 90% isodose line. DISCUSSION The ability to create dose distributions possessing improved dose homogeneity while retaining similar practical ranges to that of the standard electron energies on most linear accelerators was demonstrated. The method presented here provides a treatment scheme that
Energy-modulated electron beams ● J. R. GENTRY et al.
Fig. 7. A composite beam created from 9- and 12-MeV electrons, respectively, is shown in comparison with single 9and 12-MeVelectron depth dose distributions. The composite beam is similar to the practical range of the 9-MeV beam while displaying much improved dose homogeneity, extending easily to 2 cm. The dose from the 9-MeV beam varies by over 20% over the first 2 cm, while the composite beam varies by less than 6%. Good agreement between the algorithm’s calculation of the dose and the actual dose distribution was found (data points with error bars). The error bars represent 2 standard deviations.
Fig. 8. The fractional depth dose of a composite beam with practical range less than 9 MeV. The beam energy as determined by the R50 depth is 5.7 MeV. The dose homogeneity is 105% (103–98%) over the first 13 mm of the patient. The 9-MeV beam received a bolus of 1.2 cm and the 12-MeV beam has a bolus of 2 cm. The dose distribution was further specified by setting the dose at 3 cm to 22% of the dose at the skin surface. To achieve the dose specified, 84 cGy from the 9-MeV beam and 20 cGy from the 12-MeV beam was required. Setup of 100-cm SSD to the bolus is made. Errors bars represent 2 standard deviations in the computation of the dose by the treatment planning system.
265
266
Medical Dosimetry
Volume 31, Number 4, 2006
Fig. 9a, 9b and 9c. Treatment plan verification of the prescence and lateral extent of typical prescription isodose lines near the surface of the phantom for the mixed 6/9 MeV beam (upper row) and the single 9 MeV beam (lower row). The blue area is the 95% isodose and red is 98% isodose.
could achieve tumor control while minimizing toxicity through the use of electron energy modulation and the simultaneous computation of the appropriate bolus thickness (intensity modulation). Working alongside a planning system capable of accurately computing the dose in all 3 dimensions, the Pre-Planning Software tools make it possible to explore a number of treatment options with a minimum of effort. The program can display the composite dose distribution of 50 or more combinations in less than 30 seconds. While present methods of dose fractionation and planning for the target volume includes consideration of organ toxicity and skin dose,18,19 the Trumpet e-IMRT™ method includes the ability to specially tailor the range and homogeneity to the patient being treated while enabling the dose goals to be met. Because the Pre-Planning Software algorithm makes use of data that has been acquired previously or can be acquired rapidly in the case of an individualized electron cutout, its clinical implementation is straightforward.
Fig 10. Clinical case shows that mixed energy electron treatment (12 and 16 MeV) on the right creates a dose distribution that falls off more rapidly than the single energy 16 MeV beam (left). The peach colored line shows the prescription isodose curve in each case. The mixed energy electron beam when combined with the photon beam spared underlying heart better.
Energy-modulated electron beams ● J. R. GENTRY et al.
Verification of the composite dose distribution is performed with traditional methods of ion chamber, film and point dose detectors, as well as the treatment planning system. The incorporation of computerized planning aids that are external to the central treatment planning computer is not without precedent. The creation of optimal dose strategies involving electrons is a case in point. Fields et al.20 at M. D. Anderson developed a computer program called EMIX for the combining of electron and photon beams utilizing central axis beam data. Combining a mixture of 2D inverse planning techniques and manual methods, they achieved good dose homogeneity over a range of user-specified depths guided by patient CT data. The Pre-Planning Software algorithm makes possible both energy-modulated as well as intensitymodified treatment. Galbraith and Rawlinson,21 prior to the present era, presented a conceptual and mathematical framework for treating with differing amounts of bolus in different segments of the same treatment session with a single energy only (intensity modulation). Their partial bolus technique did not make a significant impact on clinical practice, due perhaps to the lack of incorporation of their methods into straightforward software planning tools. Intensity modulation, in the sense of providing preferential blocking for portions of the field, can also be supported by the program, but requires the fabrication of 2 blocks. Energy modulation of electron beams adds significantly to Galbraith’s idea for the treatment of superficial sites. By employing both energy and intensity modulation within the same treatment, the physician is able to deliver a uniform dose over a larger volume than would be possible through the use of a single energy alone. The work to date has required specialized equipment (MLC or electron MLC) and specialized computing platforms that compute the dose in 3 dimensions and have only recently been released. The algorithm presented here is less ambitious giving a 1-dimensional central axis depth dose distribution only. Its proper application is for any electron field treatment in which the depth dose of the electron beam is known. The strength is that it gives added value to the tools that are already resident in a department. The Trumpet e-IMRT™ program is a powerful script that the user dose not have to write. It puts in one easily accessible place all the possible solutions on the central axis for composite electron beam creation using both intensity (variable block aperture and/or bolus thickness) and energy modulation. The Pre-Planning Software replaces the necessity of creating thousands of trials on a powerful treatment planning system to optimize the dose distribution and improves results for patients. Whether already-scanned data from commissioning or electron depth dose from a patient-specific cutout gathered from film is used, the treatment planner can, within minutes, visually create, evaluate, and choose optimal beam energies and intensity-modulated solutions. The MUs to be
267
used are then entered into the more powerful 3D planning system for a more complete dosimetric display and analysis. Composite electron beam treatment may be considered favorably over single energy treatment at many sites including scalp,22 brain,16 oral cavity,23 and in the breast boost, whether intraoperatively,24 or after traditional external beam treatment with photons. In the case of the electron boost prescribed after external beam, the algorithm can provide an optimal linear combination for reducing the overall dose to the breast in comparison to single energy treatment. Kantorowitz25 found that despite the similarity in the prescription of the dose in the breast boost by number of facilities (usually the 80% isodose line), a widely varying dose was delivered due to institutional dose specification practices. His year 2000 study of institutions in the United State and Europe, as discussed in published retrospective reports, textbooks, and multi-center trials, found normalized doses between 66 and 76.11 Gy were delivered. This is a dose range that has prompted concerns by he and others about late effects.26 –29 While the algorithm facilitates the planning of treatments that have 1-mm slices of tissue equivalent bolus, in practice, 3–5-thicknesses of tissue equivalent material are the smallest increment that are readily available in most departments without special ordering. To achieve intermediate bolus thicknesses, other techniques or materials may be utilized. One technique, use of 2 or more electron beam energies, can replace the use of bolus in a single energy electron beam over limited distances. For instance, 6-MeV electrons can be bolused from their 80% entrant depth dose to 88% via the application of an unbolused 9-MeV electron beam to the treatment volume. This corresponds to 3 mm of depth. As the requirement for the entrant dose of the composite beam is increased (the 9 MeV beam becomes more heavily weighted), an increase in the practical range is also seen. Conversely, a composite beam can be created that has a smaller entrant depth dose but reduced range using the Pre-Planning Software tool. Thermoplastic material (Thermo-Shield, MED-TEC) on the skin as a compensator could further expand the usefulness of the technique. Paliwal et al.30 have demonstrated the compensation characteristics of this novel plastic that is very malleable at temperatures of 108 –132°F and stable at room temperature. They report the electron transmission percentage at the depth of maximum dose for a range of electron energies and thickness of the thermo-shield material only. By completely characterizing the depth dose distribution at additional points in tissue for several standard thicknesses of this material, it would be possible to achieve a greater number of intermediate bolus thicknesses, improving dose conformality. Compared to the scope of change that has been evident in the application of external photon beam treatment, which includes 3D conformal radiotherapy, inverse treatment planning, and tomotherapy, treatment
268
Medical Dosimetry
with electrons has remained comparatively stable with the exception of the computer intensive techniques already noted. The technique employed by Fields20 or Galbraith21 with electrons did not become developed features in the planning tools that developed later. The tool presented here, because it rapidly computes the number of MUs and applicable bolus thickness subject to dose and range constraints for specific patient variables on the central axis, can augment and improve the electron therapy planning now being done. CONCLUSIONS When the prescription of a single megavoltage electron beam is written to a point that lies beyond the depth of peak dose for single electron energy, large dose heterogeneities may be created within the patient. Numerical as well as graphical tools were created to quickly and accurately display the 1-dimensional depth dose distribution resulting from a linear combination of megavoltage electron beams output by a linear accelerator. The tools make use of depth dose data and treatment tools that departments already possess. The e-IMRT™ algorithm expands the physician’s ability to tailor the dose to each patient in a wide range of treatment situations, enabling a reduction of the radiation dose, improved dose homogeneity, and decreased dose to structures that lie beyond the point of prescription.
Acknowledgments—The authors thank Dr. Bruce Thomadsen, Ph.D; Dr. Jack Fowler, Ph.D.; Dr. James Welsh, M.D., Ph.D.; Dr. Patrick Fernandes, M.D.; and Peter Mahler, M.D., Ph.D. for their timely comments; and Mary Pat Gordon (Health Science Librarian, Freeport Memorial Hospital Library) for her assistance with the manuscript. The work was supported in part by Standard Imaging, Inc of Middleton, Wisconsin.
REFERENCES 1. Campana, F.; Zervoudis, S.; Perdereau, B.; et al. Topical superoxide dismutase reduces post-irradiation breast cancer fibrosis. J. Cell Mol. Med. 8:109 –16; 2004. 2. Porock, D.; Kristjanson, L.; Nikoletti, S.; et al. Predicting the severity of radiation skin reactions in women with breast cancer. Oncol. Nurs. Forum 25:1019 –29; 1998. 3. Calle, R.; Pilleron, J.P.; Schlienger, P.; et al. Conservative management of operable breast cancer: Ten years experience at the Foundation Curie. Cancer 42:2045–53; 1978. 4. Habibollahi, F.; Mayles, H.M.; Mayles, W.P.; et al. Assessment of skin dose and its relation to cosmesis in the conservative treatment of early breast cancer. Int. J. Radiat. Oncol. Biol. Phys. 14:291– 6; 1998. 5. Deore, S.M.; Sarin, R.; Dinshaw, K.A.; et al. Influence of dose-rate and dose per fraction on clinical outcome of breast cancer treated by external beam irradiation plus iridium-192 implants: Analysis of 289 cases. Int. J. Radiat. Oncol. Biol. Phys. 26:601– 6; 1993. 6. Borger, J.H.; Kemperman, H.; Smitt H.S.; et al. Dose and volume effects on fibrosis after breast conservation therapy. Int. J. Radiat. Oncol. Biol. Phys. 30:1073– 81; 1994. 7. Lawenda, B.D.; Taghian, A.G.; Kachnic, L.A.; et al. Dose-volume analysis of radiotherapy for T1N0 invasive breast cancer treated by local excision and partial breast irradiation by low-dose-rate interstitial. Int. J. Radiat. Oncol. Biol. Phys. 56:671– 80; 2003. 8. Vrieling, C.; Collette, L.; Fourquet, A.; et al. The influence of patient, tumor and treatment factors on the cosmetic results after breast-conserving therapy in the EORTC ‘boost vs. no boost’ trial. EORTC Radiotherapy and Breast Cancer Cooperative Groups. Radiother. Oncol. 55:219 –32; 2000.
Volume 31, Number 4, 2006 9. Buchholz, T.A.; Gurgoze, E.; Bice, W.S.; et al. Dosimetric analysis of the intact breast irradiation in off axis planes. Int. J. Radiat. Oncol. Biol. Phys. 39:261–7; 1997. 10. Ma, C.M.; Pawlicki, T.; Lee, M.C.; et al. Energy-and intensity modulated electron beams for radiotherapy. Phys. Med. Biol. 45: 2293–311; 2000. 11. Karlsson, M.G.; Karlsson, M.; Zackrisson, B. Intensity modulation with electrons: Calculations, measurements and clinical applications. Phys. Med. Biol. 43:1159 – 69; 1998. 12. Lee, M.C.; Deng, J.; Li, J.; et al. Monte Carlo based treatment planning for modulated electron beam radiation therapy. Phys. Med. Biol. 46:2177–99; 2001. 13. Hyödynmaa, S.; Gustaffson, A.; Brahme, A. Optimization of conformal electron beam therapy using energy-and fluence modulated beam. Med. Phys. 23:659 – 66; 1996. 14. Klein, E.E. Modulated electron beams using multisegmented multileaf collimation. Radiat. Oncol. 48:307–11; 1998. 15. Kudchaker, R.J.; Hogstrom, K.R.; Garden, A.S.; et al. Electron conformal radiotherapy using bolus and intensity modulation. Int. J. Radiat. Oncol. Biol. Phys. 53:1023–37; 2002. 16. Lief, E.P.; Lo, Y.; Larsson, A.; et al. Treatment of brain tumors with wedged electron beams. Int. J. Radiat. Oncol. Biol. Phys. 32:169; 1995. 17. Blomquist, M.; Karlsson, M.G.; Zackrisson, B.; et al. Multileaf collimation of electrons— clinical effects on electron energy modulation and mixed beam therapy depending on treatment head design. Phys. Med. Biol. 47:1013–24; 2002. 18. Overgaard, M.; Bentzen, S.M.; Christensen, J.J.; et al. The value of the NSD formula in equation of acute and late radiation complications in normal tissue following 2 and 5 fractions per week in breast cancer patients treated with post-mastectomy irradiation. Radiat. Oncol. 9:1–11; 1987. 19. McMillan, T.; Steel, G. Molecular aspects of radiation biology. In: Basic Clinical Radiobiology. Edward Arnold Publishers; 1983: 211–24. 19. von Essen, C. A spatial model of time-dose-area relationships in radiation therapy. Radiology 81:881–3; 1963. 20. Fields, R.S.; Spanos, W.J.; Tapley, N.; et al. Computer optimization for combining electron and photon beams (abstract). Int. J. Radiat. Oncol. Biol. Phys. 6:1444; 1980. 21. Galbraith, D.M.; Rawlinson, J.A. Partial bolusing to improve the depth doses in the surface region of the low energy electron beams. Int. J. Radiat. Oncol. Biol. Phys. 10:313–7; 1984. 22. Yaparpalvi, R.; Fontenla, D.P.; Beitler, J.J. Improved dose homogeneity in scalp irradiation using a single set-up point and different energy electron beams. Br. J. Radiol. 75:670 –7; 2002. 23. Cederbaum, M.; Zalik, M.; Rosenblatt Bar-Deroma, R. A simple oral cone attachment for the varian linear accelerator. Med. Dosim. 23:47–50; 1998. 24. Veronesi, U.; Gatti, G.; Luini, A.; Intraoperative radiation therapy for breast cancer: Technical notes. Breast J. 9:106 –12; 2003. 25. Kantorowitz, D.A. The impact of dose-specification policies upon nominal radiation dose received by breast tissue in the conservation treatment of breast cancer. Int. J. Radiat. Oncol. Biol. Phys. 47:841– 8; 2000. 26. Borger, J.H.; Keijser, A.H. Conservative breast cancer treatment: Analysis of cosmetic results and the role of concomitant adjuvant chemotherapy. Int. J. Radiat. Oncol. Biol. Phys. 13:173–7; 1987. 27. Taylor, M.E.; Perez, C.A.; Halverson, K.J.; et al. Factors influencing cosmetic results conservation therapy for breast cancer. Int. J. Radiat. Oncol. Biol. Phys. 31:753– 64; 1995. 27. Pezner, R.D.; Lipsett, J.A.; Desai, K.; et al. To boost or not to boost: Decreasing radiation therapy in conservative breast cancer treatment when “inked” tumor resection margins are pathologically free of cancer. Int. J. Radiat. Oncol. Biol. Phys. 14:873–7; 1988. 28. Delouche, G.; Bachelot. F.; Premont, M. Conservation treatment of early stage breast cancer: Long term results and complications. Int. J. Radiat. Oncol. Biol. Phys. 13:29 –34; 1987. 29. Dewar, J.A.; Benhamou, S.; Benhamou, E.; Cosmetic results following lumpectomy axillary dissection and radiotherapy for small breast cancers. Radiat. Oncol. 12:273– 80; 1988. 30. Paliwal, B.; Rommelfanger, S.; Das, R. Attenuation characteristics of a new compensator material: Thermo-Shield for high energy electron and photon beams. Med. Phys. 25:484 –7; 1998.
doi:10.1016/j.meddos.2005.03.008
INVERSE PLANNING OF ENERGY-MODULATED ELECTRON BEAMS IN RADIOTHERAPY JOHN R. GENTRY, M.S., RICHARD STEEVES, M.D., PH.D., and BHUDATT A. PALIWAL, PH.D. Department of Human Oncology, Comprehensive Cancer Care Center, University of Wisconsin-Madison, Madison, WI 53792 (Received 15 November 2004; accepted 14 March 2005)
Abstract—The use of megavoltage electron beams often poses a clinical challenge in that the planning target volume (PTV) is anterior to other radiosensitive structures and has variable depth. To ensure that skin as well as the deepest extent of the PTV receives the prescribed dose entails prescribing to a point beyond the depth of peak dose for a single electron energy. This causes dose inhomogeneities and heightened potential for tissue fibrosis, scarring, and possible soft tissue necrosis. Use of bolus on the skin improves the entrant dose at the cost of decreasing the therapeutic depth that can be treated. Selection of a higher energy to improve dose homogeneity results in increased dose to structures beyond the PTV, as well as enlargement of the volume receiving heightened dose. Measured electron data from a linear accelerator was used as input to create an inverse planning tool employing energy and intensity modulation using bolus (e-IMRT™). Using tools readily available in a radiotherapy department, the applications of energy and intensity modulation on the central axis makes it possible to remove hot spots of 115% or more over the depths clinically encountered. The e-IMRT™ algorithm enables the development of patient-specific dose distributions with user-defined positions of peak dose, range, and reduced dose to points beyond the prescription point. © 2006 American Association of Medical Dosimetrists. Key Words: Electron energy/intensity modulation, Inverse treatment Planning, Breast cancer.
lower density of the lung and therefore extended range of megavoltage electrons in this medium. To offset this problem, effort is often devoted to the construction of a customized bolus that will ensure the prescribed dose is delivered to the skin and that the electrons deliver the prescribed dose to the distal edge of the tumor and expend as much of their energy as possible before crossing the chest wall/lung interface. Such a customized bolus can be created based on a slice-by-slice computed tomography (CT) analysis of the chest wall thickness throughout the treatment portal. Use of the bolus improves the dose distribution, but may create significant dose heterogeneities near the skin surface and in the region of peak dose. While improving the tumor control probability, boost phase therapy with electrons is also associated with higher rates of fibrosis. Fibrosis, breast hardness, and edema within the irradiated breast are sources of discomfort for women undergoing conservative treatment for breast cancer. Campana et al.1 cites the release of free radicals, such as O2 and hydroxyls, as the cause of breast fibrosis. The strong instability of these molecules and their tendency to bind to adjacent structures causes injuries to the connective tissue and vascular networks. In addition to the pain, short- and long-term cosmetic implications are involved. Lymph node aspiration further compromises the local vascular system, contributing to marked or even severe skin reactions and edema.2 Age and weight,2 radiation dose,3,4 dose fractionation,5 and volume effects,6 as well as the overall treatment duration1 of the
INTRODUCTION The depth dose and practical range characteristics of megavoltage electron beams make them a natural choice in the treatment of superficial cancer both as primary therapy, in the case of a skin lesion, or as the boost in the treatment of breast cancer to the region overlying the excision in breast-conserving treatment. In the latter group, the residual chest wall and, in some patients, the internal mammary nodes as well must be treated with electrons to minimize dose to lung and heart. The necessity of delivering the prescribed dose to the skin and at the depth of the chest wall/lung interface while simultaneously minimizing the dose to the lung is a treatment framework in which competing goals exist and trade-offs are required. Treatment, in the boost phase of conservative breast therapy, often consists of a single electron energy prescribed to the chest wall/lung interface. Bolus is prescribed to deliver the treatment dose to a depth no greater than the chest wall or the depth indicated by clips left at the time of surgery to denote the depth of the tumor bed. Because the depth of prescription is typically deeper than the depth of maximum dose, in these cases, dose heterogeneities are created at the depth of peak dose of 110% or more. Care must be exercised in the choice of electron beam energy when treating the chest because of the Reprint requests to: John R. Gentry, M.S., Caromont Cancer Care Center, Gaston Memorial Hospital, 2525 Court Drive, Gastonia, NC 28054. E-mail: [email protected]. 259
260
Medical Dosimetry
radiotherapy contribute to the severity of these radiationrelated injuries. Recent studies at Massachusetts General Hospital further establish the dose volume effect in these patients, where 4 of 48 women experienced moderate to severe thickening of the skin.7 Vrieling et al.8 point specifically at the boost as an important factor in creation of breast fibrosis. They concluded that to achieve a good cosmesis, it is necessary to excise the tumor with a limited margin, to avoid postoperative complications, to assess the need for a boost in the individual patient, and to give the radiation dose as homogeneously as possible. Buchholz9 lends additional support for this line of thought and suggests that the level of dose homogeneity in the lower quadrant of the breast resulting from the photon treatment should be taken into account when planning the application of the boost electron treatment, rather than be ignored. A reasonable way to decrease dose inhomogeneity in the breast boost and reduce the dose to critical structures beyond the depth of prescription would be to deliver the dose by 2 or more electron beam energies, such that an optimized part of the prescribed dose is delivered by each energy. Optimized treatment solutions requiring the use of multiple electron beam energies have taken principally 1 of 2 paths. Some (Ma et al.,10 Karlsson et al.,11 Lee et al.,12 have taken the path of modulating both energy and intensity via a special electron multileaf collimator (MLC) and utilize Monte Carlo-based inverse treatment planning tools. Specifically, Hyödynmaa et al.13 proposed a technique for chest wall irradiation that optimized dose homogeneity via multiple electron beams by varying fluence profiles and energies. Their technique optimizes the dose distribution in 3 dimensions, and is capable of using as many as 5 beams. Subject to the constraints imposed, their computer found the iterative solution in 7 to 17 minutes. Though the tools and methods of this path are promising and represent the future for conformal electron beam therapy, they are currently not widely available. The second path is epitomized by those who have focused on the use of more readily accessible computational and hardware tools. They (Klein,14 Kudchaker et al.,15 Lief et al.,16 Blomquist et al.17) utilize an MLC, traditionally reserved for the shaping of the port for photon beam treatment with the dose intensity pattern generated by an inverse treatment planning system. Delivery of treatment using this method is not available in the clinical mode for most accelerators. It would be advantageous to offer a straightforward treatment solution that results in improved dose distributions, reduced dose to structures beyond the depth of prescription, and improves the ability of existing tools to deliver conformal electron radiotherapy. This paper will present the Trumpet e-IMRT™ Pre-Planning Software, a suite of planning tools for creating composite electron beam depth dose distributions.
Volume 31, Number 4, 2006
METHODS Mathematical algorithm for energy and intensity modulated radiotherapy In general, the dose D at any depth x in tissue subject to N normally incident megavoltage electron beams can be expressed as D共xn兲 ⫽ MU1FDD1共x兲共OAR1共x兲兲共INVSQ1兲 ⫹ MU2FDD2共x兲共OAR2共x兲兲共INVSQ2兲 ⫹ MU3FDD3共x兲共OAR3共x兲兲共INVSQ3兲 ⫹ . . . ⫹ MUN FDDN共x兲共OARN共x兲兲共INVSQN兲 (1) where MU, FDD, x, OAR, and INVSQ are monitor units, fractional depth dose, depth, off-axis ratio, and inverse square factor associated with the total dose at x, respectively. For this discussion, central axis dose with the patient at 100-cm source-to-surface distance (SSD) and field size large enough to ensure electronic equilibrium are considered exclusively. An algorithm has been developed that computes the number of MUs that should be delivered by each of the electron beam energies such that the resulting composite dose distribution meets userdefined constraints expressed as n
兺 MUiFDDi共xj兲共OARi共xj兲兲共INVSQ兲 ⫺ D共xj兲constraint ⫽ 0 i⫽1 j⫽1
(2) Visual evaluation of composite depth dose distributions In the basic application of the inverse planning concept within the Pre-Planning Software, the user specifies the dose to the skin and to a second depth within the patient. The algorithm determines the proper weighting of MUs and bolus to achieve the dose at the skin and at the second point. The set of solutions returned by the first pass of the program is further reduced by the operator through specification of the appropriate range that will be allowed. Then, application of dose homogeneity requirements further reduces the number of treatment solutions leaving the more optimal treatment solutions. A graphical interface was created to show the fractional depth dose characteristics of the composite beam relative to 2 standard beams for comparison. The algorithm’s results were then tested against the dose distribution computed by a Prowess 3000 treatment planning system utilizing the Memorial Sloan Kettering Electron Beam Computation Model for treatment on a Varian 2100 C linear accelerator and Pinnacle3 version 7.0 (ADAC Philips). The dose engine for the electron tool within Pinnacle is the Hogstrom Pencil Beam, which makes use of a combination of measured data and mathematical modeling to characterize each electron beam. The solutions to the constraint equations provide both physical as well as non-physical electron depth dose distributions. A graphical tool was introduced to show
Energy-modulated electron beams ● J. R. GENTRY et al.
261
Fig. 1. Graphical tool shows the location (shaded region) of the possible composite distributions with acceptable solutions to the system of linear equations given a set of dose constraints on the dose at the surface, in this case 1 Gy. The 12-MeV beam is bolused and the resulting dose distribution is shown. The composite dose distribution gives a homogeneous dose from the surface to 1.6-cm depth. This distribution is intermediate between a 6-MeV beam and a 9-MeV beam. The size and shape of the solution space varies with any change in the bolus used or constraints imposed. Given the requirement that a dose of 1 Gy is to be delivered at the skin surface in the figure above, a real composite distribution can be created which has a depth dose of 0.80 at 30 mm, but not 0.1 at 30 mm, because that point lies outside the shaded region.
the achievable solution sets (Figs. 1 and 2). Fig. 1 shows a composite dose distribution created by mixing 9 and 12 MeV electron beams (with bolus) that has depth dose characteristics that are intermediate between 6 MeV and 9 MeV. Several test cases using 9 MeV and 12 MeV
electron beams are presented to show the utility of this technique to deliver a dose distribution over a tumor volume to a depth beyond that of peak dose. The Pre-Planning Software algorithm computes the central axis composite depth dose for the electron beam
Fig. 2. Graphical illustration of the effect of bolus on each beam and the composite beam created by their addition. The shape of the solution space changes with the dose specification and the quantity of bolus each beam receives. A graphical tool reveals quickly the limits of any potential composite dose distribution that is achievable with the electron beam energies and bolus being used.
262
Medical Dosimetry
Volume 31, Number 4, 2006
Fig. 3. A composite dose distribution composed of 9- and 12 MeV electron beams shown beside its component 9- and 12-MeV beams. The dose heterogeneity was 110% of the prescription. The composite beam delivers less dose beyond the prescription point (3 cm) than the 12-MeV beam alone and a smaller practical range. The composite beam gives a dose of 0.61 Gy at 4 cm compared to 0.83 Gy from the 12 MeV beam alone. The monitor units calculated represent the dose which must be delivered to the depth of peak dose for each energy in order to achieve the prescribed dose. Error bars represent 2 standard deviations.
energies, MUs and bolus being used. The algorithm’s calculation of the dose on the central axis was tested against ion chamber in a typical water-equivalent phantom utilizing a 100-cm SSD. Bolus, when necessary, was placed on top of the phantom. Open cones were used exclusively for this testing. RESULTS The depth dose beam characteristics of a composite beam technique were compared to those of 2 intermediate megavoltage electron beams (9 MeV and 12 MeV) for treatment of the internal mammary nodes at a depth of 3 cm. The composite beam creation was guided by the need to create a dose distribution that optimized dose homogeneity while limiting dose to structures beyond the depth of 3 cm. Figure 3 shows the composite beam created as well as the 9- and 12-MeV beams. The composite beam was achieved by adding 0.6-cm bolus to the 12-MeV beam and 0.4 cm of bolus to the 9-MeV beam. The composite beam delivers 0.63 Gy at a depth 1 cm beyond the prescription point as compared to the 0.83 Gy using the 12-MeV beam alone. The composite beam is superior to the 9 MeV in dose homogeneity. The composite beam delivers less dose beyond the prescription point (3 cm) than the 12-MeV beam alone and has a smaller practical range. The composite beam gives a dose of 0.61 Gy at 4 cm compared to 0.83 Gy from the 12-MeV beam alone. The position of peak dose for the
distribution is found at a depth of 1.6 cm as compared to the 1.8-cm peak dose depth for the 9-MeV beam. The 50% fractional depth dose of the composite beam is extended 0.9 cm beyond the 3.3-cm position of the 9-MeV beam but is more shallow than the 12-MeV beam (4.8 cm) by 0.6 cm. The practical range of the composite beam is approximately 54 mm, giving it an effective energy that is close to 11 MeV, but energy as designated by the distance needed for the fractional depth dose to be reduced to 50% of the peak dose (R50) depth is 9.9 MeV. Figure 4 shows how the depth dose of the composite beam compares to the 2 individual beams over the first 10 mm of skin. Thus, a judicious choice of bolus remains an integral part of creating the optimal composite beam with which to treat the patient. Simply combining the beams in achieve the prescribed dose at the surface and at 3 cm yielded a composite beam (Fig. 5) that delivered less dose beyond the prescription depth, but displayed poor dose homogeneity (hot spot of 115%). The fractional depth dose within the first centimeter of a typical patient is shown in Fig. 6. While this distribution meets the dose constraints, this distribution displays less homogeneity than Fig. 3. In addition to improving dose homogeneity and decreasing dose distal to the prescription point, the algorithm facilitates the evaluation of a continuum of composite beams that have user-specified practical ranges
Energy-modulated electron beams ● J. R. GENTRY et al.
263
Fig. 4. The dose distribution of the composite and single energy electron beams (9 and 12 MEV) are shown. As a result of bolus being added to the skin for both the 9- and 12-MeV beams to form the composite beam, the entrant skin fractional depth dose of the composite beam meets the dose prescription. The increase in dose over the first 10 mm is less than 105% of the prescription.
and homogeneities within the patient. It is demonstrated in Fig. 7 that the combination of a 9-MeV beam (0.5-cm bolus, 100-cm SSD, 9 cGy delivered to the depth of peak dose) with a 12-MeV beam (1.3-cm bolus, 100-SSD, 94 cGy delivered to the depth of peak dose) yields a dose distribution that is very similar to 9 MeV, but, has significantly improved dose homogeneity. Figure 8 shows that dose distributions that have practical ranges smaller than the lowest energy component can be created that have
superior homogeneity. In this case, the composite distribution has less than 105% dose heterogeneity over a distance of 1.3 cm, comparable to the depth of peak dose for a 6-MeV electron beam. Lateral Tumor Coverage: Phantom Results The Pre-Planning Software was tasked to compute a dose distribution (Figure 9a, 9b, and 9c) for a 10x10 cm reference field that was more homogeneous than a 9
Fig. 5. A composite beam created with dose specification of 1 Gy at the surface and at depth (30 mm) in the patient. Both 9- and 12-MeV beams are shown for comparison purposes. No bolus is used in this example. Without bolus, the composite beam has dose homogeneity characteristics similar to the 9-MeV beam. This distribution contains a dose heterogeneity of 115%. The practical range is less than that of the 12-MeV beam and the integral dose delivered past the point of the prescription is decreased in comparison to the 12-MeV beam.
264
Medical Dosimetry
Volume 31, Number 4, 2006
Fig. 6. Without adding bolus to either beam, the composite beam’s entrance fractional depth dose was 11% higher than the 12-MeV beam, but the increase in the fractional dose per millimeter was similar to the 9-MeV beam over the first 1-cm depth. Without the use of bolus, the dose heterogeneity is somewhat larger.
MeV beam down to a depth of 2 cm, and provided good coverage by the 95% or higher isodose line, and had a range similar to the 9 MeV beam. In less than a minute, the program found 24 methods for creating the desired dose distribution by varying the number of mu and bolus thicknesses to be used with each energy. The mixed energy solution is shown in the upper row of panes and the single energy 9 MeV beam is shown in the lower row. Coverage by the 95% line (blue) is available at the surface (Fig 9a), and, after 2.5 mm, the 98% isodose line (red) covers a significant fraction of the mixed energy field. Figure 9b shows the lateral extent of the 98% isodose line (red) to a depth of 2 cm as computed by the Pinnacle (Phillips ADAC) planning system. The Pre-Planning Software computed solution brought a useable prescription isodose line to the surface while retaining the range characteristics of the 9 MeV beam. This is in contrast to the single energy 9 MeV beam which only displays the prescence of the 85% isodose (olive) at 0.5 cm. Clinical Application: Patient-Chest Mass A solution for a patient presenting with a chest mass was found to be a prescription using both a mixed
electron beam and a photon beam. For the electron portion of the treatment, the prescription was for 12 MeV with 2.6 cm of bolus combined with 16 MeV and 0.5 cm of bolus. The dose was normalized to the 95% isodose line to provide the best coverage. The patient then received 7 fractions of 2.5 Gy with the mixed electron beam and 7 fractions of 2.5 Gy with photons. This solution is compared to a single energy 16 MeV electron and 6 MV photon treatment in Fig 10. The 3578 cGy treatment dose is shown by the red isodose line in both pictures. The treatment dose delivered by the Pre-Planning technique encompasses the tumor, shown in red, and then falls off more rapidly thereafter, sparing lung and heart to a greater degree than the 6 MV photon/ 16 MeV single electron treatment prescribed to the 90% isodose line. DISCUSSION The ability to create dose distributions possessing improved dose homogeneity while retaining similar practical ranges to that of the standard electron energies on most linear accelerators was demonstrated. The method presented here provides a treatment scheme that
Energy-modulated electron beams ● J. R. GENTRY et al.
Fig. 7. A composite beam created from 9- and 12-MeV electrons, respectively, is shown in comparison with single 9and 12-MeVelectron depth dose distributions. The composite beam is similar to the practical range of the 9-MeV beam while displaying much improved dose homogeneity, extending easily to 2 cm. The dose from the 9-MeV beam varies by over 20% over the first 2 cm, while the composite beam varies by less than 6%. Good agreement between the algorithm’s calculation of the dose and the actual dose distribution was found (data points with error bars). The error bars represent 2 standard deviations.
Fig. 8. The fractional depth dose of a composite beam with practical range less than 9 MeV. The beam energy as determined by the R50 depth is 5.7 MeV. The dose homogeneity is 105% (103–98%) over the first 13 mm of the patient. The 9-MeV beam received a bolus of 1.2 cm and the 12-MeV beam has a bolus of 2 cm. The dose distribution was further specified by setting the dose at 3 cm to 22% of the dose at the skin surface. To achieve the dose specified, 84 cGy from the 9-MeV beam and 20 cGy from the 12-MeV beam was required. Setup of 100-cm SSD to the bolus is made. Errors bars represent 2 standard deviations in the computation of the dose by the treatment planning system.
265
266
Medical Dosimetry
Volume 31, Number 4, 2006
Fig. 9a, 9b and 9c. Treatment plan verification of the prescence and lateral extent of typical prescription isodose lines near the surface of the phantom for the mixed 6/9 MeV beam (upper row) and the single 9 MeV beam (lower row). The blue area is the 95% isodose and red is 98% isodose.
could achieve tumor control while minimizing toxicity through the use of electron energy modulation and the simultaneous computation of the appropriate bolus thickness (intensity modulation). Working alongside a planning system capable of accurately computing the dose in all 3 dimensions, the Pre-Planning Software tools make it possible to explore a number of treatment options with a minimum of effort. The program can display the composite dose distribution of 50 or more combinations in less than 30 seconds. While present methods of dose fractionation and planning for the target volume includes consideration of organ toxicity and skin dose,18,19 the Trumpet e-IMRT™ method includes the ability to specially tailor the range and homogeneity to the patient being treated while enabling the dose goals to be met. Because the Pre-Planning Software algorithm makes use of data that has been acquired previously or can be acquired rapidly in the case of an individualized electron cutout, its clinical implementation is straightforward.
Fig 10. Clinical case shows that mixed energy electron treatment (12 and 16 MeV) on the right creates a dose distribution that falls off more rapidly than the single energy 16 MeV beam (left). The peach colored line shows the prescription isodose curve in each case. The mixed energy electron beam when combined with the photon beam spared underlying heart better.
Energy-modulated electron beams ● J. R. GENTRY et al.
Verification of the composite dose distribution is performed with traditional methods of ion chamber, film and point dose detectors, as well as the treatment planning system. The incorporation of computerized planning aids that are external to the central treatment planning computer is not without precedent. The creation of optimal dose strategies involving electrons is a case in point. Fields et al.20 at M. D. Anderson developed a computer program called EMIX for the combining of electron and photon beams utilizing central axis beam data. Combining a mixture of 2D inverse planning techniques and manual methods, they achieved good dose homogeneity over a range of user-specified depths guided by patient CT data. The Pre-Planning Software algorithm makes possible both energy-modulated as well as intensitymodified treatment. Galbraith and Rawlinson,21 prior to the present era, presented a conceptual and mathematical framework for treating with differing amounts of bolus in different segments of the same treatment session with a single energy only (intensity modulation). Their partial bolus technique did not make a significant impact on clinical practice, due perhaps to the lack of incorporation of their methods into straightforward software planning tools. Intensity modulation, in the sense of providing preferential blocking for portions of the field, can also be supported by the program, but requires the fabrication of 2 blocks. Energy modulation of electron beams adds significantly to Galbraith’s idea for the treatment of superficial sites. By employing both energy and intensity modulation within the same treatment, the physician is able to deliver a uniform dose over a larger volume than would be possible through the use of a single energy alone. The work to date has required specialized equipment (MLC or electron MLC) and specialized computing platforms that compute the dose in 3 dimensions and have only recently been released. The algorithm presented here is less ambitious giving a 1-dimensional central axis depth dose distribution only. Its proper application is for any electron field treatment in which the depth dose of the electron beam is known. The strength is that it gives added value to the tools that are already resident in a department. The Trumpet e-IMRT™ program is a powerful script that the user dose not have to write. It puts in one easily accessible place all the possible solutions on the central axis for composite electron beam creation using both intensity (variable block aperture and/or bolus thickness) and energy modulation. The Pre-Planning Software replaces the necessity of creating thousands of trials on a powerful treatment planning system to optimize the dose distribution and improves results for patients. Whether already-scanned data from commissioning or electron depth dose from a patient-specific cutout gathered from film is used, the treatment planner can, within minutes, visually create, evaluate, and choose optimal beam energies and intensity-modulated solutions. The MUs to be
267
used are then entered into the more powerful 3D planning system for a more complete dosimetric display and analysis. Composite electron beam treatment may be considered favorably over single energy treatment at many sites including scalp,22 brain,16 oral cavity,23 and in the breast boost, whether intraoperatively,24 or after traditional external beam treatment with photons. In the case of the electron boost prescribed after external beam, the algorithm can provide an optimal linear combination for reducing the overall dose to the breast in comparison to single energy treatment. Kantorowitz25 found that despite the similarity in the prescription of the dose in the breast boost by number of facilities (usually the 80% isodose line), a widely varying dose was delivered due to institutional dose specification practices. His year 2000 study of institutions in the United State and Europe, as discussed in published retrospective reports, textbooks, and multi-center trials, found normalized doses between 66 and 76.11 Gy were delivered. This is a dose range that has prompted concerns by he and others about late effects.26 –29 While the algorithm facilitates the planning of treatments that have 1-mm slices of tissue equivalent bolus, in practice, 3–5-thicknesses of tissue equivalent material are the smallest increment that are readily available in most departments without special ordering. To achieve intermediate bolus thicknesses, other techniques or materials may be utilized. One technique, use of 2 or more electron beam energies, can replace the use of bolus in a single energy electron beam over limited distances. For instance, 6-MeV electrons can be bolused from their 80% entrant depth dose to 88% via the application of an unbolused 9-MeV electron beam to the treatment volume. This corresponds to 3 mm of depth. As the requirement for the entrant dose of the composite beam is increased (the 9 MeV beam becomes more heavily weighted), an increase in the practical range is also seen. Conversely, a composite beam can be created that has a smaller entrant depth dose but reduced range using the Pre-Planning Software tool. Thermoplastic material (Thermo-Shield, MED-TEC) on the skin as a compensator could further expand the usefulness of the technique. Paliwal et al.30 have demonstrated the compensation characteristics of this novel plastic that is very malleable at temperatures of 108 –132°F and stable at room temperature. They report the electron transmission percentage at the depth of maximum dose for a range of electron energies and thickness of the thermo-shield material only. By completely characterizing the depth dose distribution at additional points in tissue for several standard thicknesses of this material, it would be possible to achieve a greater number of intermediate bolus thicknesses, improving dose conformality. Compared to the scope of change that has been evident in the application of external photon beam treatment, which includes 3D conformal radiotherapy, inverse treatment planning, and tomotherapy, treatment
268
Medical Dosimetry
with electrons has remained comparatively stable with the exception of the computer intensive techniques already noted. The technique employed by Fields20 or Galbraith21 with electrons did not become developed features in the planning tools that developed later. The tool presented here, because it rapidly computes the number of MUs and applicable bolus thickness subject to dose and range constraints for specific patient variables on the central axis, can augment and improve the electron therapy planning now being done. CONCLUSIONS When the prescription of a single megavoltage electron beam is written to a point that lies beyond the depth of peak dose for single electron energy, large dose heterogeneities may be created within the patient. Numerical as well as graphical tools were created to quickly and accurately display the 1-dimensional depth dose distribution resulting from a linear combination of megavoltage electron beams output by a linear accelerator. The tools make use of depth dose data and treatment tools that departments already possess. The e-IMRT™ algorithm expands the physician’s ability to tailor the dose to each patient in a wide range of treatment situations, enabling a reduction of the radiation dose, improved dose homogeneity, and decreased dose to structures that lie beyond the point of prescription.
Acknowledgments—The authors thank Dr. Bruce Thomadsen, Ph.D; Dr. Jack Fowler, Ph.D.; Dr. James Welsh, M.D., Ph.D.; Dr. Patrick Fernandes, M.D.; and Peter Mahler, M.D., Ph.D. for their timely comments; and Mary Pat Gordon (Health Science Librarian, Freeport Memorial Hospital Library) for her assistance with the manuscript. The work was supported in part by Standard Imaging, Inc of Middleton, Wisconsin.
REFERENCES 1. Campana, F.; Zervoudis, S.; Perdereau, B.; et al. Topical superoxide dismutase reduces post-irradiation breast cancer fibrosis. J. Cell Mol. Med. 8:109 –16; 2004. 2. Porock, D.; Kristjanson, L.; Nikoletti, S.; et al. Predicting the severity of radiation skin reactions in women with breast cancer. Oncol. Nurs. Forum 25:1019 –29; 1998. 3. Calle, R.; Pilleron, J.P.; Schlienger, P.; et al. Conservative management of operable breast cancer: Ten years experience at the Foundation Curie. Cancer 42:2045–53; 1978. 4. Habibollahi, F.; Mayles, H.M.; Mayles, W.P.; et al. Assessment of skin dose and its relation to cosmesis in the conservative treatment of early breast cancer. Int. J. Radiat. Oncol. Biol. Phys. 14:291– 6; 1998. 5. Deore, S.M.; Sarin, R.; Dinshaw, K.A.; et al. Influence of dose-rate and dose per fraction on clinical outcome of breast cancer treated by external beam irradiation plus iridium-192 implants: Analysis of 289 cases. Int. J. Radiat. Oncol. Biol. Phys. 26:601– 6; 1993. 6. Borger, J.H.; Kemperman, H.; Smitt H.S.; et al. Dose and volume effects on fibrosis after breast conservation therapy. Int. J. Radiat. Oncol. Biol. Phys. 30:1073– 81; 1994. 7. Lawenda, B.D.; Taghian, A.G.; Kachnic, L.A.; et al. Dose-volume analysis of radiotherapy for T1N0 invasive breast cancer treated by local excision and partial breast irradiation by low-dose-rate interstitial. Int. J. Radiat. Oncol. Biol. Phys. 56:671– 80; 2003. 8. Vrieling, C.; Collette, L.; Fourquet, A.; et al. The influence of patient, tumor and treatment factors on the cosmetic results after breast-conserving therapy in the EORTC ‘boost vs. no boost’ trial. EORTC Radiotherapy and Breast Cancer Cooperative Groups. Radiother. Oncol. 55:219 –32; 2000.
Volume 31, Number 4, 2006 9. Buchholz, T.A.; Gurgoze, E.; Bice, W.S.; et al. Dosimetric analysis of the intact breast irradiation in off axis planes. Int. J. Radiat. Oncol. Biol. Phys. 39:261–7; 1997. 10. Ma, C.M.; Pawlicki, T.; Lee, M.C.; et al. Energy-and intensity modulated electron beams for radiotherapy. Phys. Med. Biol. 45: 2293–311; 2000. 11. Karlsson, M.G.; Karlsson, M.; Zackrisson, B. Intensity modulation with electrons: Calculations, measurements and clinical applications. Phys. Med. Biol. 43:1159 – 69; 1998. 12. Lee, M.C.; Deng, J.; Li, J.; et al. Monte Carlo based treatment planning for modulated electron beam radiation therapy. Phys. Med. Biol. 46:2177–99; 2001. 13. Hyödynmaa, S.; Gustaffson, A.; Brahme, A. Optimization of conformal electron beam therapy using energy-and fluence modulated beam. Med. Phys. 23:659 – 66; 1996. 14. Klein, E.E. Modulated electron beams using multisegmented multileaf collimation. Radiat. Oncol. 48:307–11; 1998. 15. Kudchaker, R.J.; Hogstrom, K.R.; Garden, A.S.; et al. Electron conformal radiotherapy using bolus and intensity modulation. Int. J. Radiat. Oncol. Biol. Phys. 53:1023–37; 2002. 16. Lief, E.P.; Lo, Y.; Larsson, A.; et al. Treatment of brain tumors with wedged electron beams. Int. J. Radiat. Oncol. Biol. Phys. 32:169; 1995. 17. Blomquist, M.; Karlsson, M.G.; Zackrisson, B.; et al. Multileaf collimation of electrons— clinical effects on electron energy modulation and mixed beam therapy depending on treatment head design. Phys. Med. Biol. 47:1013–24; 2002. 18. Overgaard, M.; Bentzen, S.M.; Christensen, J.J.; et al. The value of the NSD formula in equation of acute and late radiation complications in normal tissue following 2 and 5 fractions per week in breast cancer patients treated with post-mastectomy irradiation. Radiat. Oncol. 9:1–11; 1987. 19. McMillan, T.; Steel, G. Molecular aspects of radiation biology. In: Basic Clinical Radiobiology. Edward Arnold Publishers; 1983: 211–24. 19. von Essen, C. A spatial model of time-dose-area relationships in radiation therapy. Radiology 81:881–3; 1963. 20. Fields, R.S.; Spanos, W.J.; Tapley, N.; et al. Computer optimization for combining electron and photon beams (abstract). Int. J. Radiat. Oncol. Biol. Phys. 6:1444; 1980. 21. Galbraith, D.M.; Rawlinson, J.A. Partial bolusing to improve the depth doses in the surface region of the low energy electron beams. Int. J. Radiat. Oncol. Biol. Phys. 10:313–7; 1984. 22. Yaparpalvi, R.; Fontenla, D.P.; Beitler, J.J. Improved dose homogeneity in scalp irradiation using a single set-up point and different energy electron beams. Br. J. Radiol. 75:670 –7; 2002. 23. Cederbaum, M.; Zalik, M.; Rosenblatt Bar-Deroma, R. A simple oral cone attachment for the varian linear accelerator. Med. Dosim. 23:47–50; 1998. 24. Veronesi, U.; Gatti, G.; Luini, A.; Intraoperative radiation therapy for breast cancer: Technical notes. Breast J. 9:106 –12; 2003. 25. Kantorowitz, D.A. The impact of dose-specification policies upon nominal radiation dose received by breast tissue in the conservation treatment of breast cancer. Int. J. Radiat. Oncol. Biol. Phys. 47:841– 8; 2000. 26. Borger, J.H.; Keijser, A.H. Conservative breast cancer treatment: Analysis of cosmetic results and the role of concomitant adjuvant chemotherapy. Int. J. Radiat. Oncol. Biol. Phys. 13:173–7; 1987. 27. Taylor, M.E.; Perez, C.A.; Halverson, K.J.; et al. Factors influencing cosmetic results conservation therapy for breast cancer. Int. J. Radiat. Oncol. Biol. Phys. 31:753– 64; 1995. 27. Pezner, R.D.; Lipsett, J.A.; Desai, K.; et al. To boost or not to boost: Decreasing radiation therapy in conservative breast cancer treatment when “inked” tumor resection margins are pathologically free of cancer. Int. J. Radiat. Oncol. Biol. Phys. 14:873–7; 1988. 28. Delouche, G.; Bachelot. F.; Premont, M. Conservation treatment of early stage breast cancer: Long term results and complications. Int. J. Radiat. Oncol. Biol. Phys. 13:29 –34; 1987. 29. Dewar, J.A.; Benhamou, S.; Benhamou, E.; Cosmetic results following lumpectomy axillary dissection and radiotherapy for small breast cancers. Radiat. Oncol. 12:273– 80; 1988. 30. Paliwal, B.; Rommelfanger, S.; Das, R. Attenuation characteristics of a new compensator material: Thermo-Shield for high energy electron and photon beams. Med. Phys. 25:484 –7; 1998.