- Email: [email protected]
Dosimetric comparison of three different treatment techniques in extensive scalp lesion irradiation
Radiotherapy and Oncology 91 (2009) 255–260
Contents lists available at ScienceDirect
Radiotherapy and Oncology journal homepage: www.thegreenjournal.com
Scalp irradiation
Dosimetric comparison of three different treatment techniques in extensive scalp lesion irradiation Jadwiga B. Wojcicka, Donette E. Lasher *, Sandra S. McAfee, Gregory A. Fortier Department of Radiation Oncology, York Cancer Center, York, PA, USA
a r t i c l e
i n f o
Article history: Received 19 May 2008 Received in revised form 29 August 2008 Accepted 7 September 2008 Available online 25 October 2008 Keywords: Scalp irradiation High dose rate brachytherapy Intensity modulated radiotherapy Mold-based brachytherapy
a b s t r a c t Background and purpose: This study compared lateral photon/electron plan (3DCRT), intensity modulated radiation therapy (IMRT) plan, and high dose rate (HDR) brachytherapy plan for total scalp irradiation. Materials and methods: The techniques were planned on a patient with squamous cell carcinoma of the scalp for a prescribed dose of 60 Gy. Conformity indexes and dose volume histograms were used for the comparison. Results: Clinical target volume coverage factors for 3DCRT, IMRT, and HDR were 0.976, 0.998, and 0.967, and Conformation Numbers were 0.532, 0.713, and 0.761, respectively. The dose gradient across the target was 59–136%, 91–129%, and 58–242% for 3DCRT, IMRT, and HDR techniques, respectively. The 3DCRT and IMRT techniques produced low optical structure doses. 3DCRT produced hotspots in the brain, while IMRT produced brain sparing. HDR produced the highest integral doses to the brain and optical structures. Conclusions: IMRT provided the best target dose homogeneity and coverage, and delivered clinically acceptable doses to normal structures. HDR produced the most conformal plan, but the total dose delivered is limited by doses to the brain and eyes. HDR is a clinically feasible alternative for less extensive lesions, lower prescription doses, and for patients who cannot lie on the treatment table. Ó 2008 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology 91 (2009) 255–260
Total scalp irradiation may be used to treat such conditions as angiosarcoma, lymphoma, melanoma, mycosis fungoides, and squamous cell carcinoma. The goal of total scalp irradiation is to deliver a homogeneous dose to the scalp while sparing the brain and optical structures. Planning and delivering the radiation treatment is technically challenging due to the complex shape and superficial nature of the target. Megavoltage electron beams are the traditional choice for total scalp irradiation due to their high surface dose and rapid dose falloff. The characteristics of electron beams require normal incidence to the treatment surface to achieve a homogeneous dose distribution. However, the shape of the scalp does not permit normal incidence by a single electron field. Many techniques have been described in the literature to provide a homogenous dose distribution across the scalp. For example, Able et al. [1] and Mellenberg and Schoeppel [10] used several matching fields and shifted the gap during the treatment course. Sagar and Pujara [14] employed a similar technique with overlapping fields. The technique in Walker et al. [16] involved overlapping normal and tangential electron
* Corresponding author. Address: Department of Radiation Oncology, Suite 94, York Cancer Center, 25 Monument Road, York, PA 17403. E-mail addresses: [email protected] (J.B. Wojcicka), [email protected] (D.E. Lasher), [email protected] (S.S. McAfee), [email protected] (G.A. Fortier). 0167-8140/$ - see front matter Ó 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.radonc.2008.09.022
fields to improve the homogeneity at the junction. Yaparpalvi et al. [17] utilized concentric energy modulated beams. The foregoing electron techniques have problems associated with them, including dose heterogeneity across the target volume, construction of multiple field apertures, and laborious treatment setup with many field junctions. Megavoltage photon beam approaches have been utilized to overcome the difficulties with electron beams. Akazawa [2] described a combined photon–electron technique with shifting junctions where lateral electron fields were matched with lateral photon fields, which covered the outer rind of scalp around the top of the head and spared the brain. An improvement to this technique was presented in Tung et al. [15], where the photon and electron fields were overlapped by 3–4 mm to improve dose uniformity, again shifting the junction midway through the treatment. Kinard et al. [7] introduced a photon shell technique with four 90° photon arcs around the head with central blocks for sensitive structures. The ability of intensity modulated radiation therapy (IMRT) to produce concave dose distributions makes it potentially well-suited for the complicated target shape in total scalp irradiation for an extensive lesion of the total scalp. Locke et al. [9] performed a treatment planning study by comparing the lateral photon–electron technique with serial tomotherapy planned for delivery on the Peacock system (Best Nomos Radiation Oncology, Sewickley,
256
Dosimetric comparison of extensive scalp irradiation techniques
PA). The study concluded that although serial tomotherapy provided a more homogeneous dose distribution throughout the target volume, the doses to the eyes and brain were much higher. A later planning study done by Orton et al. [12] compared helical tomotherapy (Tomotherapy, Inc., Madison, WI) with the lateral photon–electron technique. This study produced tomotherapy plans with a more uniform dose to the scalp and lower doses to the brain and eyes than the more conventional technique. Bedford et al. [3] compared linear accelerator (LINAC) based IMRT with five fixed gantry angles to static electron and arcing electron techniques. The results indicated that IMRT provided superior target coverage of the treatment volume. The doses to the brain and eyes were higher with IMRT, but clinically acceptable. Most recently, Chan et al. [4] combined static electron fields with photon IMRT, improving target homogeneity compared to the electron-only approach and reducing normal tissue doses compared to photon IMRT alone. The photon techniques, including IMRT, require a conformal bolus for adequate dose buildup. It is technically challenging to construct adequate bolus and immobilization, and at the same time provide reasonable comfort for the patient. Mold-based high dose rate (HDR) brachytherapy is a potential alternative to external beam techniques. Imai et al. [6] described treating patients with angiosarcoma and malignant lymphoma of the scalp with HDR. The HDR catheters were fixed to the inner surface of a helmetshaped mold made of plaster casting tape. The patients sat comfortably for treatment. Nakamura et al. [11] presented a similar case report for an angiosarcoma patient. They created the surface mold from thermoplastic mask material. Both studies reported good local control and minimal side effects, but did not provide specific data regarding target volume coverage and doses to critical structures. Recently, Liebmann et al. [8] conducted a case study of a patient with chronic lymphatic leukemia lesions on the scalp with helmet mold-based surface brachytherapy. Dose information for the scalp, brain, and eyes was provided, and the authors reported good local control. We performed a treatment planning study by comparing the dosimetry of the lateral photon–electron, LINAC-based segmental photon IMRT, and helmet mold-based HDR brachytherapy techniques. Target volume coverage was evaluated using conformity indexes [5], and doses to critical structures were analyzed with dose volume histograms and isodose distributions. A simple helmetshaped bolus and a surface mold were constructed for this study.
scalp nodules and with a right neck mass. The patient was treated with wide local excision of the scalp nodules and a right neck dissection. The margins of the scalp lesions were positive for metastatic squamous cell carcinoma with infiltration to soft tissues, skeletal muscle, and fascia. Both sides of the patient’s neck as well as the entire scalp were treated sequentially to avoid excessive morbidity. The scalp treatment was initiated after recovery from the neck radiation and the entire scalp was treated to 60 Gy using an IMRT technique, but we also performed treatment planning with a lateral 3D-CRT photon– electron technique and a HDR mold technique. Photon IMRT During simulation, the physician defined the clinical target volume (CTV) from the skin surface to the depth of the cranium over the extent of the scalp shown in Fig. 1. The planning target volume (PTV) was delineated as the CTV plus a 0.5 cm margin. A 1 cm thick helmet-shaped bolus (Fig. 2) was constructed from solid thermoplastic material (WFR/Aquaplast Corporation, Wickoff, NJ). The IMRT plan was generated with nine equally spaced coplanar 6 MV photon beams with a total of 174 step-and-shoot segments. Planning was conducted on the Pinnacle 7.6c treatment planning system (Philips Medical Systems, Madison, WI). The dose grid was 3 mm 3 mm 3 mm. The prescription dose was 60 Gy in 30 fractions to the isodose line covering the PTV. For daily treatments, a thin layer of petroleum jelly was spread on the patient’s head to fill air gaps between the skin and the bolus.
Methods and materials Clinical background A 73-year-old male with a prior history of squamous cell carcinoma of the scalp presented to the department with recurrent
Fig. 2. Patient treatment position for the IMRT and photon–electron treatment techniques. The setup consists of the 1 cm thermoplastic mold, customized headrest, and thermoplastic mask.
Fig. 1. Treatment area for total scalp irradiation.
257
J.B. Wojcicka et al. / Radiotherapy and Oncology 91 (2009) 255–260
Fig. 3. Brachytherapy mold for the HDR technique. Catheters were attached to the helmet at 1 cm intervals. The catheters extended at least 1 cm beyond the treatment volume in all directions.
The treatment was delivered on a Clinac iX linear accelerator (Varian Medical Systems, Palo Alto, CA). The On-Board-ImagerÒ was used for daily position verification with orthogonal images. The treatments required 30 min of machine time, 15 min for bolus placement, setup, and imaging, and 15 min to execute the treatment. The IMRT treatment was verified with an ionization chamber measurement (0.3 cm3 PTW 31003) and an EDR2 film (Carestream Health, Inc., Rochester, NY) in a 30 cm 30 cm 18 cm Plastic WaterÒ phantom (Computerized Imaging Reference Systems, Inc., Norfolk, VA). The composite dose distribution was analyzed with RIT Software (Radiological Imaging Technology, Colorado Springs, CO) and found to be acceptable within 3 mm and 3%. The ion chamber measurement was within 3% of the planned values. In addition, the planned dose was verified with thermoluminescent dosimeters (TLDs) placed on the patient’s scalp. 3D-CRT photon–electron treatment plan
The catheters were reconstructed in the Plato 14.2.6 (Nucletron, Veenedaal, Netherlands) brachytherapy planning system utilizing the axial and multi-planar reconstruction views and a dwell position spacing of 5 mm. The external beam PTV, brain, and optical structures were reproduced on the new CT scans. The Nucletron MicroSelectron Remote Afterloader restricts the treatment length along each catheter to 23 cm. The majority of the PTV was covered with a margin by the catheters, but the long mid-sagittal area required that the seven central catheters have alternating offsets from the catheter tip. The isodose distribution was optimized graphically so that at least 95% of the PTV was covered by the prescription dose of 60 Gy in 30 fractions. The treatment plan was verified by placing TLD chips on the inner surface of the mold before inserting the phantom. Flexible bolus material was placed over the catheters during delivery to provide backscatter, because the treatment planning system assumes an infinite unit density medium in the dose calculation. Based on [13], four centimeters of bolus was chosen to reduce the dose uncertainty to within 3% of full scatter conditions without adding excessive weight to the helmet. The treatment was delivered on the MicroSelectron unit with an Ir192 source. The plan had a total of 1028 dwell positions and the total reference air kerma of 1.835 cGy m2. Dose conformity Feuvret et al. [5] presented a variety of dose conformity indexes useful in evaluating treatment plan dose distributions. The Lesion Coverage Factor (CVF) was defined as
Lesion Coverage Factor ¼
TVRI TV
ð1Þ
where TVRI was the volume of the target covered by the reference isodose line and TV was the target volume. The Healthy Tissue Conformity Index (HTCI) was defined as
TVRI V RI
The 3D-CRT plan was performed on the same CT data set, treatment planning system, linear accelerator, and PTV as the IMRT plan. The treatment plan included lateral photon and electron beams as described in [6]. The inner field was a mixed energy electron beam, evenly weighted with 6 MeV and 9 MeV. The outer semi-circular field was a 6 MV photon beam matched on the skin to the electron beam. The prescribed dose was 60 Gy in 30 fractions to the isodose line that covered the PTV. The treatment plan included a 1 cm shift in the junction after 15 treatments.
Healthy Tissue Conformity Index ¼
Helmet mold-based HDR treatment plan
Results
29 catheters were attached to the outside of the bolus helmet previously constructed for the IMRT treatment at 1 cm spacing (Fig. 3). The catheters were placed to extend at least 1 cm beyond the treatment area. Informed consent was obtained for the patient to return for this treatment planning CT study after the majority of his scalp had healed from the IMRT. The CT was performed with 3 mm slice spacing and in the prone position to avoid damage to the catheters. The patient had scalp tenderness that prevented him from placing the helmet firmly into position for the treatment planning CT scan. As a result, there were undesirable air cavities between the helmet and the patient’s scalp. Therefore, two treatment plans were performed, one with the patient’s data set and the other with the helmet filled with phantom material (Super Stuff bolus material, VanArsdale Innovative Products, Inc., Pensacola, FL), eliminating the majority of air gaps. The two HDR plans were compared to determine how the greater distance between radiation source and target volume affected the dosimetric parameters.
Isodose plots for the three treatment plans are displayed in Fig. 4, and dose volume histograms (DVH) are compared in Figs. 5 and 6. The IMRT plan provided the best coverage of the PTV and the most homogenous dose within the target. The HDR plan did not spare the optical structures as well as the 3D-CRT and IMRT techniques. The dose to the optical structures was slightly higher in the IMRT plan than the 3D-CRT plan, but it was within a clinically acceptable level. For example, the maximum dose to the optical structures was 4.3% of the prescription dose delivered to the optic chiasm, or 2.6 Gy over the treatment course of 60 Gy. The DVH for the brain revealed that the maximum dose to the brain was lowest with the HDR plan, but the mean dose to the brain was much higher than in the other two techniques. The DVHs of the 3D-CRT and IMRT plans crossed at 40% of the prescription dose, so that the IMRT plan irradiated a larger volume to a lower dose, and the 3D-CRT plan irradiated larger volumes to higher doses. Only the 3D-CRT technique produced hotspots greater than 100% of the prescription dose in the brain. The Lesion Coverage
ð2Þ
where VRI was the volume of the reference isodose. The Conformation Number (CN)
Conformation Number ¼ CVF HTCI ¼
TVRI TVRI TV V RI
ð3Þ
was a single index incorporating both the amount of target coverage and the extent to which normal tissues are spared.
258
Dosimetric comparison of extensive scalp irradiation techniques
Fig. 4. 3D-CRT lateral photon–electron, IMRT, and HDR isodose distributions in the sagittal plane. The isodose lines are as follows: white = 110%, red = 100%, blue = 95%, aqua = 80%, yellow = 50%, green = 20%.
Fig. 5. Target and brain dose volume Histogram comparison for the three techniques: IMRT, helmet mold-based HDR, and lateral photon–electron 3D-CRT.
Fig. 6. Optical structure dose volume histogram comparison for the three techniques: IMRT, helmet mold-based HDR, and lateral photon–electron 3D-CRT.
259
J.B. Wojcicka et al. / Radiotherapy and Oncology 91 (2009) 255–260
Table 1 Treatment plan evaluation via the Lesion Coverage Factor, Healthy Tissue Conformity Index, and Conformation Number. The volumes indicated were determined from the dose volume histograms.
TVRIa b
TV VRIc Lesion Coverage Factor (TVRI/TV) Healthy Tissue Conformity Index (TVRI/VRI) Conformation Number (LCF HTCI)
Photon–electron plan (cm3)
IMRT plan (cm3)
HDR patient plan (cm3)
HDR phantom plan (cm3)
418.4 428.8 767.8 0.976 0.545 0.532
428.1 428.8 598.9 0.998 0.715 0.713
414.0 428.0 526.0 0.967 0.787 0.761
409.7 428.0 462.0 0.957 0.887 0.848
For an idealized plan with perfect conformance between the target volume and reference isodose line, LCF = HTCI = CN = 1.0. a TVRI = volume of the target covered by the reference isodose line. b TV = target volume. c VRI = volume of the reference isodose.
Factor, Healthy Tissue Conformity Index, and Conformation Number are presented for each plan in Table 1. The measured TLD results from the IMRT and HDR phantom plans were within 5% of the planned values. Discussion Extensive lesions of the scalp present relatively rarely in the radiotherapy clinic. The cases encountered vary in the extent, shape, and concavity of the treatment area and the proximity to critical structures. This variability and low sample size make it difficult to select one approach for all patients, and each technique has advantages and disadvantages. In this work, a detailed comparison of 3D-CRT, IMRT, and HDR treatment planning and delivery techniques was done for a case involving a large lesion of the total scalp. The photon–electron technique has been the most commonly used one. Matching the photon and electron beams and shifting the match-line during the treatment course increase the technical difficulty in executing this plan. This technique provided low doses to the optical structures. However, it produced the least conformal plan, had a dose gradient of 59–136% in the target, and included a hotspot in the brain. The IMRT technique offered the best target coverage and the most homogenous dose distribution (91–129%) within the target. This technique has also been evaluated by Locke et al. [9] and Bedford et al. [3]. Although Locke et al. [9] found the dose to the brain and eyes unacceptably high with total scalp IMRT, Bedford et al. [3] found the doses clinically acceptable for irradiation of an extensive lesion of the partial scalp. Our DVH results were similar to those of Bedford et al. [3], despite the fact that our PTV was more extensive and exceptionally convex. Among the three techniques, the IMRT plan produced the least amount of brain tissue irradiated above 60% of the prescription dose. The optical structure doses were slightly higher than the 3D-CRT plan, but clinically acceptable at less than 3 Gy. IMRT plans involving non-coplanar beams were evaluated in an effort to reduce the dose to the optical structures. The optical structure DVHs from these plans were only slightly improved from the original plan, falling in between the 3D-CRT and coplanar IMRT plan DVHs. The coplanar plan was used to treat the patient because it could be delivered more efficiently on the treatment machine, without a clinically significant increase in critical structure dose. The IMRT plan had higher conformity indexes than the photon– electron plan and did not involve complicated field matching for daily treatments or junction shifts. However, the sharp dose gradient increased the effect of setup uncertainty on target coverage [3], so employment of daily setup verification with orthogonal radiographs was needed, increasing the amount of time the patient was on the table. The HDR plans were the most conformal of the three techniques, but produced a 58–242% dose gradient within the target.
This result was not surprising due to the dose fall-off characteristics of Ir192. The brain and optical structures received the highest doses with the HDR technique. Our treatment plan results were consistent with those of Liebmann et al. [8]. This consequence of brachytherapy was hard to avoid, because the radiation was emitted isotropically from each dwell position and summed in the surrounding space. In addition, the geometry of the treatment made it impossible to shield the brain or optical structures from the radiation source. Our intention to deliver 60 Gy to an extensive lesion of the entire scalp precluded the use of HDR for this particular patient. However, this technique could be clinically appropriate for treatments requiring a lower prescription dose, as a boost, or for less extensive lesions (as in [6,8,11]). Despite the total dose limitations, the HDR technique has several advantages. Treatment delivery is simple, setup is highly reproducible, and the treatment is well tolerated. It is an alternative for patients who cannot lie on the treatment table, because the patient sits comfortably for treatment. It is feasible to cover the whole scalp despite the limitations of the HDR afterloader unit. Attaching the catheters to the mold is time-consuming for the radiotherapy staff, but the patient only requires two short simulation procedures. The differences between the patient and phantom HDR plans demonstrated the importance of fitting the helmet to avoid air gaps between the mold and skin. The treatment planning system did not account for dose differences due to medium heterogeneities, which increased the uncertainty in the actual dose delivered to the scalp. The dosimetric discrepancy reported in the two plans was mainly influenced by the variation in source to target difference. The increased distance improved the homogeneity within the target at the expense of increased brain dose. The HDR plan with the air gaps was less conformal than the one without the gaps. Delivery of both the IMRT and HDR plans was verified within clinically acceptable levels of accuracy with TLD. Conclusions LINAC-based segmental IMRT, helmet mold-based HDR, and lateral photon–electron 3D-CRT are all clinically feasible options for irradiation of the total scalp. The photon–electron technique has the longest history and spares the optical structures, but involves troublesome field matching and dose heterogeneity. IMRT produces a homogeneous and conformal dose distribution with only a moderate increase in dose to the optical structures. The IMRT is sensitive to setup uncertainty and can involve long treatment times in an uncomfortable position. The HDR technique is the most conformal and simple to deliver, but is limited in total prescription dose due to the brain and optical structure tolerances and target dose inhomogeneity.
260
Dosimetric comparison of extensive scalp irradiation techniques
References [1] Able CM, Mills MD, McNeese MD, et al. Evaluation of a total scalp electron irradiation technique. Int J Radiat Oncol Biol Phys 1991;21:1063–72. [2] Akazawa C. Treatment of the scalp using photon and electron beams. Med Dosim 1989;14:129–31. [3] Bedford JL, Childs PJ, Hansen VN, et al. Treatment of extensive scalp lesions with segmental intensity-modulated photon therapy. Int J Radiat Oncol Biol Phys 2005;62:1549–58. [4] Chan F, Song Y, Burman C, Chui C, Schupak K. The treatment of extensive scalp lesions combining electrons with intensity-modulated photons. Eng Med Biol Soc 2006;1:152–5. [5] Feuvret L, Noel G, Mazeron J, et al. Conformity index: a review. Int J Radiat Oncol Biol Phys 2006;64:333–42. [6] Imai M, Nishimura T, Nozue M, et al. The 192Ir surface-mold technique for a whole scalp irradiation. J Jpn Soc Ther Radiol Oncol 1999;11:27–31. [7] Kinard JD, Zwicker RD, Schmidt-Ullrich RK, et al. Short communication: total craniofacial photon shell technique for radiotherapy of extensive angiosarcomas of the head. Br J Radiol 1996;69:351–5. [8] Liebmann A, Pohlmann S, Heinicke F, et al. Helmet mold-based surface brachytherapy for homogenous scalp treatment: a case report. Strahlenther Onkol 2007;183:211–4.
[9] Locke J, Low DA, Grigireit T, et al. Potential of tomotherapy for total scalp treatment. Int J Radiat Oncol Biol Phys 2002;52:553–9. [10] Mellenberg D, Schoeppel S. Total scalp irradiation of mycosis fungoides: the 4 4 technique. Int J Radiat Oncol Biol Phys 1993;27:953–8. [11] Nakamura R, Harada S, Obara T, et al. Iridium-192 brachytherapy for hemorrhagic angiosarcoma of the scalp: a case report. Jpn J Clin Oncol 2003;33:198–201. [12] Orton N, Jaradat H, Welsh J, et al. Total scalp irradiation using helical tomotherapy. Med Dosim 2005;30:162–8. [13] Raina S, Avadhani JS, Moonseong O, et al. Quantifying IOHDR brachytherapy underdosage resulting from an incomplete scatter environment. Int J Radiat Oncol Biol Phys 2005;61:1582–6. [14] Sagar S, Pujara C. Radical Treatment of angiosarcoma of the scalp using megavoltage electron beam therapy. Br J Radiol 1992;65:421–4. [15] Tung S, Shiu A, Starkschall G, et al. Dosimetric evaluation of total scalp irradiation using a lateral electron–photon technique. Int J Radiat Oncol Biol Phys 1993;27:153–60. [16] Walker C, Wadd N, Lucraft H. Novel solutions to the problems encountered in electron radiation to the surface of the head. Br J Radiol 1999;2:787–91. [17] Yaparpalvi R, Fontenla D, Beitler J. Improved dose homogeneity in scalp irradiation using a single set-up point and different energy electron beams. Br J Radiol 2002;75:670–7.
Contents lists available at ScienceDirect
Radiotherapy and Oncology journal homepage: www.thegreenjournal.com
Scalp irradiation
Dosimetric comparison of three different treatment techniques in extensive scalp lesion irradiation Jadwiga B. Wojcicka, Donette E. Lasher *, Sandra S. McAfee, Gregory A. Fortier Department of Radiation Oncology, York Cancer Center, York, PA, USA
a r t i c l e
i n f o
Article history: Received 19 May 2008 Received in revised form 29 August 2008 Accepted 7 September 2008 Available online 25 October 2008 Keywords: Scalp irradiation High dose rate brachytherapy Intensity modulated radiotherapy Mold-based brachytherapy
a b s t r a c t Background and purpose: This study compared lateral photon/electron plan (3DCRT), intensity modulated radiation therapy (IMRT) plan, and high dose rate (HDR) brachytherapy plan for total scalp irradiation. Materials and methods: The techniques were planned on a patient with squamous cell carcinoma of the scalp for a prescribed dose of 60 Gy. Conformity indexes and dose volume histograms were used for the comparison. Results: Clinical target volume coverage factors for 3DCRT, IMRT, and HDR were 0.976, 0.998, and 0.967, and Conformation Numbers were 0.532, 0.713, and 0.761, respectively. The dose gradient across the target was 59–136%, 91–129%, and 58–242% for 3DCRT, IMRT, and HDR techniques, respectively. The 3DCRT and IMRT techniques produced low optical structure doses. 3DCRT produced hotspots in the brain, while IMRT produced brain sparing. HDR produced the highest integral doses to the brain and optical structures. Conclusions: IMRT provided the best target dose homogeneity and coverage, and delivered clinically acceptable doses to normal structures. HDR produced the most conformal plan, but the total dose delivered is limited by doses to the brain and eyes. HDR is a clinically feasible alternative for less extensive lesions, lower prescription doses, and for patients who cannot lie on the treatment table. Ó 2008 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology 91 (2009) 255–260
Total scalp irradiation may be used to treat such conditions as angiosarcoma, lymphoma, melanoma, mycosis fungoides, and squamous cell carcinoma. The goal of total scalp irradiation is to deliver a homogeneous dose to the scalp while sparing the brain and optical structures. Planning and delivering the radiation treatment is technically challenging due to the complex shape and superficial nature of the target. Megavoltage electron beams are the traditional choice for total scalp irradiation due to their high surface dose and rapid dose falloff. The characteristics of electron beams require normal incidence to the treatment surface to achieve a homogeneous dose distribution. However, the shape of the scalp does not permit normal incidence by a single electron field. Many techniques have been described in the literature to provide a homogenous dose distribution across the scalp. For example, Able et al. [1] and Mellenberg and Schoeppel [10] used several matching fields and shifted the gap during the treatment course. Sagar and Pujara [14] employed a similar technique with overlapping fields. The technique in Walker et al. [16] involved overlapping normal and tangential electron
* Corresponding author. Address: Department of Radiation Oncology, Suite 94, York Cancer Center, 25 Monument Road, York, PA 17403. E-mail addresses: [email protected] (J.B. Wojcicka), [email protected] (D.E. Lasher), [email protected] (S.S. McAfee), [email protected] (G.A. Fortier). 0167-8140/$ - see front matter Ó 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.radonc.2008.09.022
fields to improve the homogeneity at the junction. Yaparpalvi et al. [17] utilized concentric energy modulated beams. The foregoing electron techniques have problems associated with them, including dose heterogeneity across the target volume, construction of multiple field apertures, and laborious treatment setup with many field junctions. Megavoltage photon beam approaches have been utilized to overcome the difficulties with electron beams. Akazawa [2] described a combined photon–electron technique with shifting junctions where lateral electron fields were matched with lateral photon fields, which covered the outer rind of scalp around the top of the head and spared the brain. An improvement to this technique was presented in Tung et al. [15], where the photon and electron fields were overlapped by 3–4 mm to improve dose uniformity, again shifting the junction midway through the treatment. Kinard et al. [7] introduced a photon shell technique with four 90° photon arcs around the head with central blocks for sensitive structures. The ability of intensity modulated radiation therapy (IMRT) to produce concave dose distributions makes it potentially well-suited for the complicated target shape in total scalp irradiation for an extensive lesion of the total scalp. Locke et al. [9] performed a treatment planning study by comparing the lateral photon–electron technique with serial tomotherapy planned for delivery on the Peacock system (Best Nomos Radiation Oncology, Sewickley,
256
Dosimetric comparison of extensive scalp irradiation techniques
PA). The study concluded that although serial tomotherapy provided a more homogeneous dose distribution throughout the target volume, the doses to the eyes and brain were much higher. A later planning study done by Orton et al. [12] compared helical tomotherapy (Tomotherapy, Inc., Madison, WI) with the lateral photon–electron technique. This study produced tomotherapy plans with a more uniform dose to the scalp and lower doses to the brain and eyes than the more conventional technique. Bedford et al. [3] compared linear accelerator (LINAC) based IMRT with five fixed gantry angles to static electron and arcing electron techniques. The results indicated that IMRT provided superior target coverage of the treatment volume. The doses to the brain and eyes were higher with IMRT, but clinically acceptable. Most recently, Chan et al. [4] combined static electron fields with photon IMRT, improving target homogeneity compared to the electron-only approach and reducing normal tissue doses compared to photon IMRT alone. The photon techniques, including IMRT, require a conformal bolus for adequate dose buildup. It is technically challenging to construct adequate bolus and immobilization, and at the same time provide reasonable comfort for the patient. Mold-based high dose rate (HDR) brachytherapy is a potential alternative to external beam techniques. Imai et al. [6] described treating patients with angiosarcoma and malignant lymphoma of the scalp with HDR. The HDR catheters were fixed to the inner surface of a helmetshaped mold made of plaster casting tape. The patients sat comfortably for treatment. Nakamura et al. [11] presented a similar case report for an angiosarcoma patient. They created the surface mold from thermoplastic mask material. Both studies reported good local control and minimal side effects, but did not provide specific data regarding target volume coverage and doses to critical structures. Recently, Liebmann et al. [8] conducted a case study of a patient with chronic lymphatic leukemia lesions on the scalp with helmet mold-based surface brachytherapy. Dose information for the scalp, brain, and eyes was provided, and the authors reported good local control. We performed a treatment planning study by comparing the dosimetry of the lateral photon–electron, LINAC-based segmental photon IMRT, and helmet mold-based HDR brachytherapy techniques. Target volume coverage was evaluated using conformity indexes [5], and doses to critical structures were analyzed with dose volume histograms and isodose distributions. A simple helmetshaped bolus and a surface mold were constructed for this study.
scalp nodules and with a right neck mass. The patient was treated with wide local excision of the scalp nodules and a right neck dissection. The margins of the scalp lesions were positive for metastatic squamous cell carcinoma with infiltration to soft tissues, skeletal muscle, and fascia. Both sides of the patient’s neck as well as the entire scalp were treated sequentially to avoid excessive morbidity. The scalp treatment was initiated after recovery from the neck radiation and the entire scalp was treated to 60 Gy using an IMRT technique, but we also performed treatment planning with a lateral 3D-CRT photon– electron technique and a HDR mold technique. Photon IMRT During simulation, the physician defined the clinical target volume (CTV) from the skin surface to the depth of the cranium over the extent of the scalp shown in Fig. 1. The planning target volume (PTV) was delineated as the CTV plus a 0.5 cm margin. A 1 cm thick helmet-shaped bolus (Fig. 2) was constructed from solid thermoplastic material (WFR/Aquaplast Corporation, Wickoff, NJ). The IMRT plan was generated with nine equally spaced coplanar 6 MV photon beams with a total of 174 step-and-shoot segments. Planning was conducted on the Pinnacle 7.6c treatment planning system (Philips Medical Systems, Madison, WI). The dose grid was 3 mm 3 mm 3 mm. The prescription dose was 60 Gy in 30 fractions to the isodose line covering the PTV. For daily treatments, a thin layer of petroleum jelly was spread on the patient’s head to fill air gaps between the skin and the bolus.
Methods and materials Clinical background A 73-year-old male with a prior history of squamous cell carcinoma of the scalp presented to the department with recurrent
Fig. 2. Patient treatment position for the IMRT and photon–electron treatment techniques. The setup consists of the 1 cm thermoplastic mold, customized headrest, and thermoplastic mask.
Fig. 1. Treatment area for total scalp irradiation.
257
J.B. Wojcicka et al. / Radiotherapy and Oncology 91 (2009) 255–260
Fig. 3. Brachytherapy mold for the HDR technique. Catheters were attached to the helmet at 1 cm intervals. The catheters extended at least 1 cm beyond the treatment volume in all directions.
The treatment was delivered on a Clinac iX linear accelerator (Varian Medical Systems, Palo Alto, CA). The On-Board-ImagerÒ was used for daily position verification with orthogonal images. The treatments required 30 min of machine time, 15 min for bolus placement, setup, and imaging, and 15 min to execute the treatment. The IMRT treatment was verified with an ionization chamber measurement (0.3 cm3 PTW 31003) and an EDR2 film (Carestream Health, Inc., Rochester, NY) in a 30 cm 30 cm 18 cm Plastic WaterÒ phantom (Computerized Imaging Reference Systems, Inc., Norfolk, VA). The composite dose distribution was analyzed with RIT Software (Radiological Imaging Technology, Colorado Springs, CO) and found to be acceptable within 3 mm and 3%. The ion chamber measurement was within 3% of the planned values. In addition, the planned dose was verified with thermoluminescent dosimeters (TLDs) placed on the patient’s scalp. 3D-CRT photon–electron treatment plan
The catheters were reconstructed in the Plato 14.2.6 (Nucletron, Veenedaal, Netherlands) brachytherapy planning system utilizing the axial and multi-planar reconstruction views and a dwell position spacing of 5 mm. The external beam PTV, brain, and optical structures were reproduced on the new CT scans. The Nucletron MicroSelectron Remote Afterloader restricts the treatment length along each catheter to 23 cm. The majority of the PTV was covered with a margin by the catheters, but the long mid-sagittal area required that the seven central catheters have alternating offsets from the catheter tip. The isodose distribution was optimized graphically so that at least 95% of the PTV was covered by the prescription dose of 60 Gy in 30 fractions. The treatment plan was verified by placing TLD chips on the inner surface of the mold before inserting the phantom. Flexible bolus material was placed over the catheters during delivery to provide backscatter, because the treatment planning system assumes an infinite unit density medium in the dose calculation. Based on [13], four centimeters of bolus was chosen to reduce the dose uncertainty to within 3% of full scatter conditions without adding excessive weight to the helmet. The treatment was delivered on the MicroSelectron unit with an Ir192 source. The plan had a total of 1028 dwell positions and the total reference air kerma of 1.835 cGy m2. Dose conformity Feuvret et al. [5] presented a variety of dose conformity indexes useful in evaluating treatment plan dose distributions. The Lesion Coverage Factor (CVF) was defined as
Lesion Coverage Factor ¼
TVRI TV
ð1Þ
where TVRI was the volume of the target covered by the reference isodose line and TV was the target volume. The Healthy Tissue Conformity Index (HTCI) was defined as
TVRI V RI
The 3D-CRT plan was performed on the same CT data set, treatment planning system, linear accelerator, and PTV as the IMRT plan. The treatment plan included lateral photon and electron beams as described in [6]. The inner field was a mixed energy electron beam, evenly weighted with 6 MeV and 9 MeV. The outer semi-circular field was a 6 MV photon beam matched on the skin to the electron beam. The prescribed dose was 60 Gy in 30 fractions to the isodose line that covered the PTV. The treatment plan included a 1 cm shift in the junction after 15 treatments.
Healthy Tissue Conformity Index ¼
Helmet mold-based HDR treatment plan
Results
29 catheters were attached to the outside of the bolus helmet previously constructed for the IMRT treatment at 1 cm spacing (Fig. 3). The catheters were placed to extend at least 1 cm beyond the treatment area. Informed consent was obtained for the patient to return for this treatment planning CT study after the majority of his scalp had healed from the IMRT. The CT was performed with 3 mm slice spacing and in the prone position to avoid damage to the catheters. The patient had scalp tenderness that prevented him from placing the helmet firmly into position for the treatment planning CT scan. As a result, there were undesirable air cavities between the helmet and the patient’s scalp. Therefore, two treatment plans were performed, one with the patient’s data set and the other with the helmet filled with phantom material (Super Stuff bolus material, VanArsdale Innovative Products, Inc., Pensacola, FL), eliminating the majority of air gaps. The two HDR plans were compared to determine how the greater distance between radiation source and target volume affected the dosimetric parameters.
Isodose plots for the three treatment plans are displayed in Fig. 4, and dose volume histograms (DVH) are compared in Figs. 5 and 6. The IMRT plan provided the best coverage of the PTV and the most homogenous dose within the target. The HDR plan did not spare the optical structures as well as the 3D-CRT and IMRT techniques. The dose to the optical structures was slightly higher in the IMRT plan than the 3D-CRT plan, but it was within a clinically acceptable level. For example, the maximum dose to the optical structures was 4.3% of the prescription dose delivered to the optic chiasm, or 2.6 Gy over the treatment course of 60 Gy. The DVH for the brain revealed that the maximum dose to the brain was lowest with the HDR plan, but the mean dose to the brain was much higher than in the other two techniques. The DVHs of the 3D-CRT and IMRT plans crossed at 40% of the prescription dose, so that the IMRT plan irradiated a larger volume to a lower dose, and the 3D-CRT plan irradiated larger volumes to higher doses. Only the 3D-CRT technique produced hotspots greater than 100% of the prescription dose in the brain. The Lesion Coverage
ð2Þ
where VRI was the volume of the reference isodose. The Conformation Number (CN)
Conformation Number ¼ CVF HTCI ¼
TVRI TVRI TV V RI
ð3Þ
was a single index incorporating both the amount of target coverage and the extent to which normal tissues are spared.
258
Dosimetric comparison of extensive scalp irradiation techniques
Fig. 4. 3D-CRT lateral photon–electron, IMRT, and HDR isodose distributions in the sagittal plane. The isodose lines are as follows: white = 110%, red = 100%, blue = 95%, aqua = 80%, yellow = 50%, green = 20%.
Fig. 5. Target and brain dose volume Histogram comparison for the three techniques: IMRT, helmet mold-based HDR, and lateral photon–electron 3D-CRT.
Fig. 6. Optical structure dose volume histogram comparison for the three techniques: IMRT, helmet mold-based HDR, and lateral photon–electron 3D-CRT.
259
J.B. Wojcicka et al. / Radiotherapy and Oncology 91 (2009) 255–260
Table 1 Treatment plan evaluation via the Lesion Coverage Factor, Healthy Tissue Conformity Index, and Conformation Number. The volumes indicated were determined from the dose volume histograms.
TVRIa b
TV VRIc Lesion Coverage Factor (TVRI/TV) Healthy Tissue Conformity Index (TVRI/VRI) Conformation Number (LCF HTCI)
Photon–electron plan (cm3)
IMRT plan (cm3)
HDR patient plan (cm3)
HDR phantom plan (cm3)
418.4 428.8 767.8 0.976 0.545 0.532
428.1 428.8 598.9 0.998 0.715 0.713
414.0 428.0 526.0 0.967 0.787 0.761
409.7 428.0 462.0 0.957 0.887 0.848
For an idealized plan with perfect conformance between the target volume and reference isodose line, LCF = HTCI = CN = 1.0. a TVRI = volume of the target covered by the reference isodose line. b TV = target volume. c VRI = volume of the reference isodose.
Factor, Healthy Tissue Conformity Index, and Conformation Number are presented for each plan in Table 1. The measured TLD results from the IMRT and HDR phantom plans were within 5% of the planned values. Discussion Extensive lesions of the scalp present relatively rarely in the radiotherapy clinic. The cases encountered vary in the extent, shape, and concavity of the treatment area and the proximity to critical structures. This variability and low sample size make it difficult to select one approach for all patients, and each technique has advantages and disadvantages. In this work, a detailed comparison of 3D-CRT, IMRT, and HDR treatment planning and delivery techniques was done for a case involving a large lesion of the total scalp. The photon–electron technique has been the most commonly used one. Matching the photon and electron beams and shifting the match-line during the treatment course increase the technical difficulty in executing this plan. This technique provided low doses to the optical structures. However, it produced the least conformal plan, had a dose gradient of 59–136% in the target, and included a hotspot in the brain. The IMRT technique offered the best target coverage and the most homogenous dose distribution (91–129%) within the target. This technique has also been evaluated by Locke et al. [9] and Bedford et al. [3]. Although Locke et al. [9] found the dose to the brain and eyes unacceptably high with total scalp IMRT, Bedford et al. [3] found the doses clinically acceptable for irradiation of an extensive lesion of the partial scalp. Our DVH results were similar to those of Bedford et al. [3], despite the fact that our PTV was more extensive and exceptionally convex. Among the three techniques, the IMRT plan produced the least amount of brain tissue irradiated above 60% of the prescription dose. The optical structure doses were slightly higher than the 3D-CRT plan, but clinically acceptable at less than 3 Gy. IMRT plans involving non-coplanar beams were evaluated in an effort to reduce the dose to the optical structures. The optical structure DVHs from these plans were only slightly improved from the original plan, falling in between the 3D-CRT and coplanar IMRT plan DVHs. The coplanar plan was used to treat the patient because it could be delivered more efficiently on the treatment machine, without a clinically significant increase in critical structure dose. The IMRT plan had higher conformity indexes than the photon– electron plan and did not involve complicated field matching for daily treatments or junction shifts. However, the sharp dose gradient increased the effect of setup uncertainty on target coverage [3], so employment of daily setup verification with orthogonal radiographs was needed, increasing the amount of time the patient was on the table. The HDR plans were the most conformal of the three techniques, but produced a 58–242% dose gradient within the target.
This result was not surprising due to the dose fall-off characteristics of Ir192. The brain and optical structures received the highest doses with the HDR technique. Our treatment plan results were consistent with those of Liebmann et al. [8]. This consequence of brachytherapy was hard to avoid, because the radiation was emitted isotropically from each dwell position and summed in the surrounding space. In addition, the geometry of the treatment made it impossible to shield the brain or optical structures from the radiation source. Our intention to deliver 60 Gy to an extensive lesion of the entire scalp precluded the use of HDR for this particular patient. However, this technique could be clinically appropriate for treatments requiring a lower prescription dose, as a boost, or for less extensive lesions (as in [6,8,11]). Despite the total dose limitations, the HDR technique has several advantages. Treatment delivery is simple, setup is highly reproducible, and the treatment is well tolerated. It is an alternative for patients who cannot lie on the treatment table, because the patient sits comfortably for treatment. It is feasible to cover the whole scalp despite the limitations of the HDR afterloader unit. Attaching the catheters to the mold is time-consuming for the radiotherapy staff, but the patient only requires two short simulation procedures. The differences between the patient and phantom HDR plans demonstrated the importance of fitting the helmet to avoid air gaps between the mold and skin. The treatment planning system did not account for dose differences due to medium heterogeneities, which increased the uncertainty in the actual dose delivered to the scalp. The dosimetric discrepancy reported in the two plans was mainly influenced by the variation in source to target difference. The increased distance improved the homogeneity within the target at the expense of increased brain dose. The HDR plan with the air gaps was less conformal than the one without the gaps. Delivery of both the IMRT and HDR plans was verified within clinically acceptable levels of accuracy with TLD. Conclusions LINAC-based segmental IMRT, helmet mold-based HDR, and lateral photon–electron 3D-CRT are all clinically feasible options for irradiation of the total scalp. The photon–electron technique has the longest history and spares the optical structures, but involves troublesome field matching and dose heterogeneity. IMRT produces a homogeneous and conformal dose distribution with only a moderate increase in dose to the optical structures. The IMRT is sensitive to setup uncertainty and can involve long treatment times in an uncomfortable position. The HDR technique is the most conformal and simple to deliver, but is limited in total prescription dose due to the brain and optical structure tolerances and target dose inhomogeneity.
260
Dosimetric comparison of extensive scalp irradiation techniques
References [1] Able CM, Mills MD, McNeese MD, et al. Evaluation of a total scalp electron irradiation technique. Int J Radiat Oncol Biol Phys 1991;21:1063–72. [2] Akazawa C. Treatment of the scalp using photon and electron beams. Med Dosim 1989;14:129–31. [3] Bedford JL, Childs PJ, Hansen VN, et al. Treatment of extensive scalp lesions with segmental intensity-modulated photon therapy. Int J Radiat Oncol Biol Phys 2005;62:1549–58. [4] Chan F, Song Y, Burman C, Chui C, Schupak K. The treatment of extensive scalp lesions combining electrons with intensity-modulated photons. Eng Med Biol Soc 2006;1:152–5. [5] Feuvret L, Noel G, Mazeron J, et al. Conformity index: a review. Int J Radiat Oncol Biol Phys 2006;64:333–42. [6] Imai M, Nishimura T, Nozue M, et al. The 192Ir surface-mold technique for a whole scalp irradiation. J Jpn Soc Ther Radiol Oncol 1999;11:27–31. [7] Kinard JD, Zwicker RD, Schmidt-Ullrich RK, et al. Short communication: total craniofacial photon shell technique for radiotherapy of extensive angiosarcomas of the head. Br J Radiol 1996;69:351–5. [8] Liebmann A, Pohlmann S, Heinicke F, et al. Helmet mold-based surface brachytherapy for homogenous scalp treatment: a case report. Strahlenther Onkol 2007;183:211–4.
[9] Locke J, Low DA, Grigireit T, et al. Potential of tomotherapy for total scalp treatment. Int J Radiat Oncol Biol Phys 2002;52:553–9. [10] Mellenberg D, Schoeppel S. Total scalp irradiation of mycosis fungoides: the 4 4 technique. Int J Radiat Oncol Biol Phys 1993;27:953–8. [11] Nakamura R, Harada S, Obara T, et al. Iridium-192 brachytherapy for hemorrhagic angiosarcoma of the scalp: a case report. Jpn J Clin Oncol 2003;33:198–201. [12] Orton N, Jaradat H, Welsh J, et al. Total scalp irradiation using helical tomotherapy. Med Dosim 2005;30:162–8. [13] Raina S, Avadhani JS, Moonseong O, et al. Quantifying IOHDR brachytherapy underdosage resulting from an incomplete scatter environment. Int J Radiat Oncol Biol Phys 2005;61:1582–6. [14] Sagar S, Pujara C. Radical Treatment of angiosarcoma of the scalp using megavoltage electron beam therapy. Br J Radiol 1992;65:421–4. [15] Tung S, Shiu A, Starkschall G, et al. Dosimetric evaluation of total scalp irradiation using a lateral electron–photon technique. Int J Radiat Oncol Biol Phys 1993;27:153–60. [16] Walker C, Wadd N, Lucraft H. Novel solutions to the problems encountered in electron radiation to the surface of the head. Br J Radiol 1999;2:787–91. [17] Yaparpalvi R, Fontenla D, Beitler J. Improved dose homogeneity in scalp irradiation using a single set-up point and different energy electron beams. Br J Radiol 2002;75:670–7.