- Email: [email protected]
Dosimetric evaluation of the INTRABEAM system for breast intraoperative radiotherapy: A single-institution experience
ARTICLE IN PRESS
JID: MDO
[mUS5Gb;October 9, 2019;11:9]
Medical Dosimetry xxx (xxxx) xxx
Medical Dosimetry journal homepage: www.meddos.org
Dosimetric evaluation of the INTRABEAM system for breast intraoperative radiotherapy: A single-institution experience Mubin Y. Shaikh, M.S. ∗,∗∗, Adrian Nalichowski, Ph.D. †, Michael C. Joiner, M.A., Ph.D. ‡, Jay Burmeister, Ph.D. † ∗
Department of Radiation Oncology, Rochester Regional, Rochester, NY 14621, USA Department of Oncology, Wayne State University School of Medicine, Gershenson Radiation Oncology Center, Barbara Ann Karmanos Cancer Institute, Detroit, MI ‡ Department of Oncology, Wayne State University, Gershenson Radiation Oncology Center, Detroit, MI 48201-2013, USA †
a r t i c l e
i n f o
Article history: Received 20 June 2019 Revised 22 August 2019 Accepted 11 September 2019 Available online xxx Keywords: INTRABEAM IORT Dose-volume histograms Breast IORT Intraoperative dosimetry
a b s t r a c t Breast intraoperative radiotherapy (IORT) with the INTRABEAM system uses a 50 kV x-ray source to deliver a single fraction of radiation therapy to the lumpectomy cavity during breast-conserving surgery. We seek to perform a dosimetric analysis of the lumpectomy cavity for rigid spherical applicators. Water phantom measurements were acquired to validate the vendor-provided x-ray calibration. The planning target volume (PTV) was defined as a 10 mm expansion beyond the spherical applicator, a dose-volume histogram (DVH) was generated and dose-volume parameters [Dmin , D1mm , V90 , V80 , V50 , HI] were reported. Additionally, the therapeutic treatment depth using the 90 and 80% isodose level was computed [R90 , R80 ]. When the percent depth dose (PDD) is normalized to the surface of the applicator, smaller applicators have a steeper PDD. For a prescription dose of 20 Gy to the surface of the applicator, the range of dose-volume parameters for the PTV was: 3.15 to 6.84 Gy for Dmin , 16.2 to 17.6 Gy for D1mm , 2.6 to 6.9% for V90 , 5.5 to 15.1% for V80 , and 21.1 to 55.6% for V50 . For applicators 15 to 50 mm in diameter, the reported values were: 6.35 to 2.9 for HI, 0.53 to 0.85 mm for R90 , and 1.18 to 1.85 mm for R80 . Smaller applicators have reduced PTV coverage but elevated HI because the attenuation of the beam proximal to the source is more pronounced. Additionally, the presence of the aluminum filter for small applicators (≤30 mm) increases PTV coverage but reduces the dose rate on the applicator surface. The delivery of IORT is performed in the OR without the use of image-based planning. To overcome this limitation, we have generated sample DVH’s and report dosimetric parameters to offer clinicians a unique dosimetric perspective. © 2019 American Association of Medical Dosimetrists. Published by Elsevier Inc. All rights reserved.
Introduction Breast cancer is the second most common type of nonskin cancer worldwide and it is estimated that 1 in 8 women, and one in 1 in 833 men, will be affected during their lifetime.1 Self-awareness campaigns and extensive screening have led to early diagnosis where early-stage cancers can be treated with surgical excision and intraoperative radiation therapy (IORT) to the lumpectomy cavity using electronic brachytherapy sources such as the INTRABEAM (Carl Zeiss Meditec AG, Oberkochen, Germany) and Xoft Axxent eBx (iCAD Inc., Nashua, NH). For the INTRABEAM system, radiation is delivered via a stationary radiation source at the center of a rigid spherical applicator, while the Axxent system
∗∗ Reprint requests to Mubin Shaikh, MS, DABR, Department of Radiation Oncology, Rochester Regional, 1425 Portland Ave, Rochester, NY 14621, USA. E-mail address: [email protected] (M.Y. Shaikh).
features a miniaturized mobile x-ray source that is stepped inside a saline inflated balloon to deliver the prescription dose to the balloon surface. This study exclusively considers the INTRABEAM system. The advantages of these low-energy systems are reduced shielding requirements, patient convenience, and increased relative biological effectiveness .2-4 In contrast, the drawbacks of IORT include lack of pathological confirmation of tumor-free margins. If cancer is discovered in the margin, then the delivered IORT becomes a boost to the tumor bed and the patient requires additional radiation therapy. Also, the lack of image-based planning limits our clinical ability to document useful information such as the dose to lumpectomy cavity and its adjacent structures (i.e., skin, chest wall, lung, and heart). Previous studies have characterized the change in dose with depth from the spherical applicators using film measurements,5 , 6 Monte Carlo simulations,7 and in vitro cell line experiments.4
https://doi.org/10.1016/j.meddos.2019.09.002 0958-3947/© 2019 American Association of Medical Dosimetrists. Published by Elsevier Inc. All rights reserved.
JID: MDO 2
ARTICLE IN PRESS
[mUS5Gb;October 9, 2019;11:9]
M.Y. Shaikh, A. Nalichowski and M.C. Joiner et al. / Medical Dosimetry xxx (xxxx) xxx
For breast brachytherapy with MammoSite and Contura, the planning target volume (PTV) is a 10 mm spherical expansion around the treatment balloon and the prescription point is at the distal edge of the PTV which is 10 mm from the balloon surface. In contrast, the TARGeted Intraoperative radioTherapy (TARGIT) protocol does not define a PTV for IORT treatments but it offers two single fraction prescription methods: a surface prescription of 20 Gy to the surface of the applicator or 6 Gy at 10 mm depth.8 When a prescription at depth is chosen, we observe a significant overdosage for smaller applicators, therefore, peer reviewed TARGIT publications endorse prescription at the applicator surface.8-10 Additionally, there are country-specific rules where the prescription must include the highest dose delivered. Consequently, surface prescriptions are the only methods used in those countries (i.e., Germany).10 Therefore, in this study, we have adopted the surface prescription methodology in order to stay consistent with the INTRABEAM user community. A significant attraction of electron beam radiotherapy is the shape of the electron depth dose curve which offers an increased surface dose combined with rapid dose fall-off beyond the target. This distinct clinical difference makes electron beam therapy favorable over conventional megavoltage x-ray photons for superficial and shallow tumors. In contrast, low energy photons (i.e., 50 kV) share a depth dose profile similar to electrons, therefore, we can apply a metric common to electron radiotherapy to IORT treatments. The therapeutic treatment depth in electron radiotherapy is defined by either the 90% isodose level (i.e., R90 ), or the 80% isodose level (i.e., R80 ), and the appropriate electron energy is chosen to cover the distal edge of the tumor.11 In breast IORT, the tumor is surgically removed so distal tumor edge is not applicable, but the distance between the boundary of the lesion and the edge of the excised specimen (i.e., margin width) is an important metric. Numerous investigators have shown that postoperative radiation therapy markedly improved the outcome of patients with small margin widths (< 1mm).12 , 13 These studies indicate the importance of delivering adequate dose to the proximal 1 mm of tissue. Therefore, we report the dose being delivered to the proximal 1 mm [D1mm ] and calculate therapeutic treatment depths [R90, R80 ] as a function of applicator size and design. For example, the smaller spherical applicators (≤ 30 mm) feature an aluminum filter to harden the spectrum. The present delivery workflow of IORT is similar to the initial MammoSite experience in which image-based planning was not required and dose-volume histograms (DVH) were not computed because the treatment involved a single dwell position with treatment time being chosen from a look-up table. The objective of the current study is to perform a dosimetric analysis of the lumpectomy cavity using dose-volume parameters [V90 , V80 , V50 , Dmin , D1mm, and HI] and to evaluate the change in treatment depth [R80, R90 ] because of spherical applicator size and filter design. While percent depth dose has been previously published for the range of spherical applicators, this study computes the DVH to offer an alternative dosimetric perspective, and is the first to report treatment depth values for the applicators.
Fig. 1. An internal schematic of the INTRABEAM x-ray source courtesy of Carl Zeiss Meditec AG©.
Methods and Materials X-ray source The INTRABEAM x-ray generator accelerates a beam of electrons down a 100 mm drift tube towards a thin hemispherical gold target situated inside a hemispherical beryllium window.7 The steering coils, at the base of the source, oscillate the beam around the tube axis to create an approximately isotropic distribution of bremsstrahlung radiation.7 Figure 1 illustrates the internal components of the x-ray device used to produce 50 kV x-rays. Investigators have validated the depth dose characteristics of the x-ray source and have proposed guidelines for commissioning and quality assurance.2 , 3
Fig. 2. Shielded water phantom setup to verify the depth dose characteristics of the x-ray source courtesy of Carl Zeiss Meditec AG©.
JID: MDO
ARTICLE IN PRESS M.Y. Shaikh, A. Nalichowski and M.C. Joiner et al. / Medical Dosimetry xxx (xxxx) xxx
[mUS5Gb;October 9, 2019;11:9] 3
Fig. 3. A 3-dimensional representation of 2 isodose surfaces. The center sphere represents the spherical applicator which receives the prescription dose, and the outer sphere represents the PTV. Images courtesy of Carl Zeiss Meditec AG©.
Spherical applicators The INTRABEAM system has a finite range of spherical applicators that range from 15 to 50 mm in increments of 5 mm. These applicators are constructed of a water-equivalent plastic known as polyetherimide (ULTEM) with a hollow internal cavity. An aluminum filter is embedded between the applicator body and x-ray source (XRS) probe for the smaller spherical applicators (≤ 30 mm) to intentionally harden the beam by removing very low energy photons from the treatment spectrum.2 For larger applicators (> 30 mm), beam hardening is sufficiently performed by the applicator body and no aluminum attenuator is required.14 We performed a phantom study which considered a comprehensive range of clinically available applicators for the INTRABEAM system. INTRABEAM calibration verification The INTRABEAM x-ray source is calibrated by the manufacturer annually during servicing, and the vendor provides measured output data that will be used in this study to calculate dosimetric coverage. Therefore, we independently validated the manufacturer’s calibration by mounting the x-ray source on a precise 3-dimensional translational stage and making measurements using a fixed PTW model 34013 (Physikalisch Technische Werkstaetten, Freiburg, Germany) parallelplate ionization chamber. The aforementioned chamber has a width of 0.9 mm in
the direction of the dose fall-off, cavity radius of 1.45 mm and a sensitive volume of 0.005 cm3 , and is the only choice for verifying the output characteristics of the xray source when using the Zeiss water phantom.15 Figure 2 shows the water phantom setup used to validate the vendor-provided depth dose from 5 to 30 mm in intervals of 5 mm. The vendor provided calibration dataset extends to a depth of 45 mm, but the water phantom setup is limited to a depth of 30 mm. Dosimetry Figure 3 represents the spherical applicator positioned in the breast tissue and the diagram on the left represents the position of the applicator relative to the PTV. A prescription of 20 Gy is delivered to the surface of the applicator, and due to the steep dose gradients, the PTV receives a nonhomogenous dose. For the PTV, DVH metrics [V90 , V80 , V50 , D1mm , Dmin, and HI] were used to quantify the dose distribution and coverage. Vx is the volume of target tissue receiving x percent of the prescription dose, D1mm is absorbed dose in Gray to the tissue at 1 mm depth, Dmin is the minimum dose to the PTV at 10 mm depth, and HI is the ratio of the maximum and minimum dose in the PTV. The maximum PTV dose occurs at the surface of the implanted device, while the minimum dose occurs at a depth of 10 mm. The product of the bare source output and the applicator transfer function is used to generate a relative percent depth dose profile. Using the applicator specific radial percent depth dose, the therapeutic treatment depth in millimeters is calculated
Fig. 4. Vendor calibrated x-ray beam dose rate in water for a 50 kV x-ray source operated at 40 μA and validated with ionization chamber measurements in a specialized water phantom. The calibrated x-ray beam data was modeled by a power fit trend line using Microsoft Excel.
JID: MDO 4
ARTICLE IN PRESS
[mUS5Gb;October 9, 2019;11:9]
M.Y. Shaikh, A. Nalichowski and M.C. Joiner et al. / Medical Dosimetry xxx (xxxx) xxx
Fig. 5. (a) Relative percent depth dose from the surface of 15 to 50 mm in diameter spherical applicators with 5 mm intervals. (b) Comparison of relative percent depth dose from the surface of 30, 35, and 40 mm diameter spherical applicators.
using the 80 and 90% isodose line [R80, R90 ]. Furthermore, we fit the data to a power series using the trend line functionality of Microsoft Excel (2017, Redmond, Washington). The power series function format is given by Eq. (1). Where A and B are model parameters calculated to fit the dataset, x will represent depth in millimeters, and y represents the output dose rate in Gy/min. y = AeBx
(1)
Results and discussion Figure 4 shows the vendor-provided calibrated dose rate (Gy/min) in water for the bare x-ray source (i.e., no spherical applicator), ionization chamber measurements and a trendline of best fit. For the 5 mm depth measurement point, the difference between the calibrated dose rate and measured dose rate is less
than 5%, and for deeper depths, the difference was less than 3%. At the 5 mm depth, small positional devaitions have an amplified impact and deviations exceeding 3% are expected.5 The intention behind the measurements was to validate the vendor-provided calibration before performing the dosimetry study. Additionally, the dataset was fit to a power series function and the line of best fit was expressed mathematically on the figure (R2 = 0.9994). We found the change in dose rate with depth to be more rapid than the inverse square because of beam attenuation. Given the rapid change in dose with depth, a logarithmic scale is better suited to highlight the subtle differences between measurements and the factory calibration. Figure 5a shows the percent depth dose in water normalized to the applicator surface on a logarithmic scale, as differences between applicators are easily obscured on a linear scale. For smaller spherical applicators, the percent depth dose changes rapidly with depth because the impact of beam attention is more pronounced closer to the x-ray source. When evaluating the percent depth dose for a
ARTICLE IN PRESS
JID: MDO
[mUS5Gb;October 9, 2019;11:9]
M.Y. Shaikh, A. Nalichowski and M.C. Joiner et al. / Medical Dosimetry xxx (xxxx) xxx Table 2 Dosimetric metrics for the spherical INTRABEAM applicators
Table 1 A summary of the applicator size and dose rate Applicator diameter (mm) 15 20 25 30 35 40 45 50
5
TARGIT: Dose rate on applicator surface (Gy/min) 2.755 1.764 1.163 0.775 1.023 0.751 0.546 0.399
full spectrum of applicator sizes, the impact of the aluminum filter for the smaller applicators (≤ 30 mm) needs to be considered. Figure 5b compares the normalized percent depth dose of the 30, 35, and 40 mm diameter spherical applicators. The 30 mm spherical applicator is unique in that it features an aluminum filter not present in the 35 and 40 mm diameter spherical applicator. The aluminum attenuator increases penetration depth and offsets some of the effects of beam attention, such that the 30 and 40 mm spherical applicators have a similar normalized percent depth dose. Table 1 presents the dose rate at the surface of the applicator. The full range of spherical applicators can be separated into 2 groups: applicators less than 30 mm which feature an aluminum filter, and applicators greater than 30 mm without an aluminum filter. While analyzing each applicator group separately we observed that the dose rate on the applicator surface decreases with increasing applicator diameter. However, when we compare all applicators as a single group we observe that the surface dose rates of the 35 mm applicators are higher than the 30 mm spherical applicator. This effect can be explained by the presence of the aluminum filter, which increases penetration depth [R80, R90 ] but reduces the dose rate at the surface of the applicator. Figure 6 shows the DVH for the PTV according to spherical applicator diameter, and Table 2 summarizes the dose-volume metrics (V90 , V80 , V50 , HI, Dmin and D1mm ). For a prescription of 20 Gy to the surface of the applicator and applicators between 15 to 50 mm, the Dmin to the PTV ranges between 3.15 to 6.84 Gy, and D1mm ranges between 16.2 to 17.6 Gy. Smaller applicators have reduced Dmin and D1mm doses. The presence of the aluminum filter increases the Dmin and D1mm dose and can be observed by comparing the adjacent 30 and 35 mm diameter spherical applicators. The range of dose-volume parameters across all applicator sizes was: 2.6 to 6.9% for V90 , 5.5 to 15.1% for V80 , 21.1 to 55.6% for V50 . Additionally, for the applicators 15 to 50 mm HI ranged from 6.35 to 2.9, with smaller applicators exhibiting increased HI values. Armoogum et al.16 conducted a functional comparison
Applicator diameter (mm)
V90 (%)
V80 (%)
V50 (%)
HI
Dmin (Gy)
D1mm (Gy)
15 20 25 30 35 40 45 50
2.6 3.6 4.2 5.0 4.7 5.2 6.8 6.9
5.5 7.4 8.9 10.7 10.2 11.5 14.0 15.1
21.1 27.0 33.1 41.0 37.2 43.5 49.4 55.6
6.35 5.03 4.32 3.59 4.03 3.61 3.22 2.90
3.15 3.97 4.66 5.54 4.97 5.55 6.20 6.84
16.2 16.6 16.9 17.3 16.9 17.2 17.6 17.6
of four INTRABEAM x-ray sources and found similar output characteristics. Therefore, the dosimetric comparison in this study is relevant to all INTRABEAM users. The concept of therapeutic treatment depth comes from electron beam radiotherapy and has not been previously calculated for IORT treatments with a 50 kV X-ray source. Figure 7 presents the R90 , R80 and D1mm for the full range of spherical applicators. For applicators between 15 and 50 mm, R90 values ranged from 0.53 to 0.85 mm and R80 values ranged from 1.18 to 1.85 mm. When we separate the applicators into 2 groups: applicators less than 30 mm which feature an aluminum filter, and applicators greater than 30 mm without an aluminum filter, we observed that values for R90 , R80 and D1mm increase with increasing applicator size. When comparing adjacent applicator sizes of 30 and 35 mm from different groups, we observe that the presence of an aluminum filter increases R90 , R80, and D1mm . If the aluminum filter was removed we would expect a linearly decreasing trend in the aforementioned values. Numerous studies have shown that patients with a small margin width (< 1mm) benefit from postoperative radiation therapy,12 , 13 indicating the relevance of dose to the proximal tissue. This study is the first to compute the dose R90 , R80 , and D1mm as a function of applicator size and design. The delivery of IORT is unique in that it is performed within the OR without the use of CT-based planning, which limits the ability to produce a DVH for the lumpectomy cavity. The presented dosimetric dataset provides the framework for clinicians to estimate coverage for homogeneous water-equivalent breast tissue, and to our knowledge, a similar dosimetric comparison has not been reported for the INTRABEAM system. A limitation of our study is that we have modeled breast tissue as being water-equivalent, however, for low-energy photon sources, breast tissue is not perfectly water-equivalent. Ebert and Carruthers reported a reduction in dose of 4% for 40 mm depth when comparing INTRABEAM dose calculations in water to breast tissue.17 Additionally, the introduction of heterogeneities such as the bone and lung would modify the dose-volume coverage and is beyond the scope of this paper.
Fig. 6. DVH for the PTV according to applicator size.
ARTICLE IN PRESS
JID: MDO 6
[mUS5Gb;October 9, 2019;11:9]
M.Y. Shaikh, A. Nalichowski and M.C. Joiner et al. / Medical Dosimetry xxx (xxxx) xxx
Fig. 7. For each applicator size we report the therapeutic treatment depth in millimeters as defined by the 80 and 90% [R80, R90] isodose line and the dose to 1 mm depth [D1mm ] in gray.
Conclusions Accurate knowledge of the absorbed dose to water is crucial for achieving optimal treatment outcomes and managing normal tissue toxicities. To overcome the lack of CT-imaging data in IORT treatments, we have generated sample DVH’s to document dose to the PTV and report a range of dosimetric parameters (V90 , V80 , V50 , D1mm , and Dmin ) for the full range of spherical breast applicators (15 to 50 mm). The PTV receives a more homogeneous dose with increasing spherical applicator size because the impact of beam attenuation with depth is less pronounced for larger applicators. The presence of the aluminum filter in the smaller applicator increases minimum PTV dose but reduces the surface dose rate. Declaration of Competing Interest The authors declare no conflicts of interest. Acknowledgments The authors thank Theresa Kwiatkowski and Robin Scott for their assistance with technical editing and comments that greatly improved the manuscript. References 1. Siegel, R.L.; Miller, K.D.; Jemal, A. Cancer statistics, 2019. CA. Cancer. J. Clin. 69:7– 34; 2019. doi:10.3322/caac.21551. 2. Eaton, D.J. Quality assurance and independent dosimetry for an intraoperative x-ray device. Med. Phys. 39:6908–20; 2012. doi:10.1118/1.4761865. 3. Beatty, J.; Biggs, P.J.; Gall, K.; et al. A new miniature x-ray device for interstitial radiosurgery: Dosimetry. Med. Phys. 23:53–62; 1996. doi:10.1118/1.597791. 4. Liu, Q.; Schneider, F.; Ma, L.; et al. Relative Biologic Effectiveness (RBE) of 50 kV X-rays measured in a phantom for intraoperative tumor-bed irradiation. Int. J. Radiat. Oncol. Biol. Phys. 85:1127–33; 2013. doi:10.1016/j.ijrobp.2012.08.005.
5. Watson, P.G.F.; Bekerat, H.; Papaconstadopoulos, P.; et al. An investigation into the INTRABEAM miniature x-ray source dosimetry using ionization chamber and radiochromic film measurements. Med. Phys. 45:4274–86; 2018. doi:10. 1002/mp.13059. 6. Sethi, A.; Emami, B.; Small, W.; et al. Intraoperative radiotherapy with INTRABEAM: Technical and dosimetric considerations. Front. Ancol. 8:74; 2018. doi:10.3389/fonc.2018.0 0 074. 7. Yanch, J.C.; Harte, K.J. Monte Carlo simulation of a miniature, radiosurgery xray tube using the ITS 3.0 coupled electron-photon transport code. Med. Phys. 23:1551–8; 1996. doi:10.1118/1.597885. 8. Vaidya, J.S.; Joseph, D.J.; Tobias, J.S.; et al. Targeted intraoperative radiotherapy versus whole breast radiotherapy for breast cancer (TARGIT-A trial): An international, prospective, randomised, non-inferiority phase 3 trial. Lancet (London, England) 376:91–102; 2010. doi:10.1016/S0140- 6736(10)60837- 9. 9. Vaidya, J.S.; Baum, M.; Tobias, J.S.; et al. The novel technique of delivering targeted intraoperative radiotherapy (Targit) for early breast cancer. Eur. J. Surg. Oncol. 28:447–54; 2002. doi:10.1053/ejso.2002.1275. 10. Vaidya, J.S.; Wenz, F.; Bulsara, M.; et al. An international randomised controlled trial to compare TARGeted Intraoperative radioTherapy (TARGIT) with conventional postoperative radiotherapy after breast-conserving surgery for women with early-stage breast cancer (the TARGIT-A trial). Health. Technol. Assess. 20:13; 2016. https://njl-admin.nihr.ac.uk/document/download/2003454. 11. Ibbott, G.S Radiation Dosimetry: Electron beams with energies between 1 and 50 MeV (ICRU Report No. 35). Med. Phys. 12:813; 1985. doi:10.1118/1.595780. 12. Silverstein, M.J.; Lagios, M.D.; Groshen, S.; et al. The influence of margin width on local control of ductal carcinoma in situ of the breast. N. Engl. J. Med. 340:1455–61; 1999. doi:10.1056/NEJM199905133401902. 13. Dunne, C.; Burke, J.P.; Morrow, M.; et al. Effect of margin status on local recurrence after breast conservation and radiation therapy for ductal carcinoma in situ. J. Clin. Oncol. 27:1615–20; 2009. doi:10.120 0/JCO.20 08.17.5182. 14. Keshtgar, M.R.S.; Vaidya, J.S.; Tobias, J.S.; et al. Targeted intraoperative radiotherapy for breast cancer in patients in whom external beam radiation is not possible. Int. J. Radiat. Oncol. Biol. Phys. 80:31–8; 2011. doi:10.1016/j.ijrobp.2010. 01.045. 15. Watson, P.G.F.; Popovic, M.; Seuntjens, J. Determination of absorbed dose to water from a miniature kilovoltage x-ray source using a parallel-plate ionization chamber. Phys. Med. Biol. 63:15016; 2017. doi:10.1088/1361-6560/aa9560. 16. Armoogum, K.S.; Parry, J.M.; Souliman, S.K.; et al. Functional intercomparison of intraoperative radiotherapy equipment – Photon Radiosurgery System. Radiat. Oncol. 2:11; 2007. doi:10.1186/1748-717X-2-11. 17. Ebert, M.A.; Carruthers, B. Dosimetric characteristics of a low-kV intra-operative x-ray source: implications for use in a clinical trial for treatment of low-risk breast cancer. Med. Phys. 30:2424–31; 2003. doi:10.1118/1.1595611.
JID: MDO
[mUS5Gb;October 9, 2019;11:9]
Medical Dosimetry xxx (xxxx) xxx
Medical Dosimetry journal homepage: www.meddos.org
Dosimetric evaluation of the INTRABEAM system for breast intraoperative radiotherapy: A single-institution experience Mubin Y. Shaikh, M.S. ∗,∗∗, Adrian Nalichowski, Ph.D. †, Michael C. Joiner, M.A., Ph.D. ‡, Jay Burmeister, Ph.D. † ∗
Department of Radiation Oncology, Rochester Regional, Rochester, NY 14621, USA Department of Oncology, Wayne State University School of Medicine, Gershenson Radiation Oncology Center, Barbara Ann Karmanos Cancer Institute, Detroit, MI ‡ Department of Oncology, Wayne State University, Gershenson Radiation Oncology Center, Detroit, MI 48201-2013, USA †
a r t i c l e
i n f o
Article history: Received 20 June 2019 Revised 22 August 2019 Accepted 11 September 2019 Available online xxx Keywords: INTRABEAM IORT Dose-volume histograms Breast IORT Intraoperative dosimetry
a b s t r a c t Breast intraoperative radiotherapy (IORT) with the INTRABEAM system uses a 50 kV x-ray source to deliver a single fraction of radiation therapy to the lumpectomy cavity during breast-conserving surgery. We seek to perform a dosimetric analysis of the lumpectomy cavity for rigid spherical applicators. Water phantom measurements were acquired to validate the vendor-provided x-ray calibration. The planning target volume (PTV) was defined as a 10 mm expansion beyond the spherical applicator, a dose-volume histogram (DVH) was generated and dose-volume parameters [Dmin , D1mm , V90 , V80 , V50 , HI] were reported. Additionally, the therapeutic treatment depth using the 90 and 80% isodose level was computed [R90 , R80 ]. When the percent depth dose (PDD) is normalized to the surface of the applicator, smaller applicators have a steeper PDD. For a prescription dose of 20 Gy to the surface of the applicator, the range of dose-volume parameters for the PTV was: 3.15 to 6.84 Gy for Dmin , 16.2 to 17.6 Gy for D1mm , 2.6 to 6.9% for V90 , 5.5 to 15.1% for V80 , and 21.1 to 55.6% for V50 . For applicators 15 to 50 mm in diameter, the reported values were: 6.35 to 2.9 for HI, 0.53 to 0.85 mm for R90 , and 1.18 to 1.85 mm for R80 . Smaller applicators have reduced PTV coverage but elevated HI because the attenuation of the beam proximal to the source is more pronounced. Additionally, the presence of the aluminum filter for small applicators (≤30 mm) increases PTV coverage but reduces the dose rate on the applicator surface. The delivery of IORT is performed in the OR without the use of image-based planning. To overcome this limitation, we have generated sample DVH’s and report dosimetric parameters to offer clinicians a unique dosimetric perspective. © 2019 American Association of Medical Dosimetrists. Published by Elsevier Inc. All rights reserved.
Introduction Breast cancer is the second most common type of nonskin cancer worldwide and it is estimated that 1 in 8 women, and one in 1 in 833 men, will be affected during their lifetime.1 Self-awareness campaigns and extensive screening have led to early diagnosis where early-stage cancers can be treated with surgical excision and intraoperative radiation therapy (IORT) to the lumpectomy cavity using electronic brachytherapy sources such as the INTRABEAM (Carl Zeiss Meditec AG, Oberkochen, Germany) and Xoft Axxent eBx (iCAD Inc., Nashua, NH). For the INTRABEAM system, radiation is delivered via a stationary radiation source at the center of a rigid spherical applicator, while the Axxent system
∗∗ Reprint requests to Mubin Shaikh, MS, DABR, Department of Radiation Oncology, Rochester Regional, 1425 Portland Ave, Rochester, NY 14621, USA. E-mail address: [email protected] (M.Y. Shaikh).
features a miniaturized mobile x-ray source that is stepped inside a saline inflated balloon to deliver the prescription dose to the balloon surface. This study exclusively considers the INTRABEAM system. The advantages of these low-energy systems are reduced shielding requirements, patient convenience, and increased relative biological effectiveness .2-4 In contrast, the drawbacks of IORT include lack of pathological confirmation of tumor-free margins. If cancer is discovered in the margin, then the delivered IORT becomes a boost to the tumor bed and the patient requires additional radiation therapy. Also, the lack of image-based planning limits our clinical ability to document useful information such as the dose to lumpectomy cavity and its adjacent structures (i.e., skin, chest wall, lung, and heart). Previous studies have characterized the change in dose with depth from the spherical applicators using film measurements,5 , 6 Monte Carlo simulations,7 and in vitro cell line experiments.4
https://doi.org/10.1016/j.meddos.2019.09.002 0958-3947/© 2019 American Association of Medical Dosimetrists. Published by Elsevier Inc. All rights reserved.
JID: MDO 2
ARTICLE IN PRESS
[mUS5Gb;October 9, 2019;11:9]
M.Y. Shaikh, A. Nalichowski and M.C. Joiner et al. / Medical Dosimetry xxx (xxxx) xxx
For breast brachytherapy with MammoSite and Contura, the planning target volume (PTV) is a 10 mm spherical expansion around the treatment balloon and the prescription point is at the distal edge of the PTV which is 10 mm from the balloon surface. In contrast, the TARGeted Intraoperative radioTherapy (TARGIT) protocol does not define a PTV for IORT treatments but it offers two single fraction prescription methods: a surface prescription of 20 Gy to the surface of the applicator or 6 Gy at 10 mm depth.8 When a prescription at depth is chosen, we observe a significant overdosage for smaller applicators, therefore, peer reviewed TARGIT publications endorse prescription at the applicator surface.8-10 Additionally, there are country-specific rules where the prescription must include the highest dose delivered. Consequently, surface prescriptions are the only methods used in those countries (i.e., Germany).10 Therefore, in this study, we have adopted the surface prescription methodology in order to stay consistent with the INTRABEAM user community. A significant attraction of electron beam radiotherapy is the shape of the electron depth dose curve which offers an increased surface dose combined with rapid dose fall-off beyond the target. This distinct clinical difference makes electron beam therapy favorable over conventional megavoltage x-ray photons for superficial and shallow tumors. In contrast, low energy photons (i.e., 50 kV) share a depth dose profile similar to electrons, therefore, we can apply a metric common to electron radiotherapy to IORT treatments. The therapeutic treatment depth in electron radiotherapy is defined by either the 90% isodose level (i.e., R90 ), or the 80% isodose level (i.e., R80 ), and the appropriate electron energy is chosen to cover the distal edge of the tumor.11 In breast IORT, the tumor is surgically removed so distal tumor edge is not applicable, but the distance between the boundary of the lesion and the edge of the excised specimen (i.e., margin width) is an important metric. Numerous investigators have shown that postoperative radiation therapy markedly improved the outcome of patients with small margin widths (< 1mm).12 , 13 These studies indicate the importance of delivering adequate dose to the proximal 1 mm of tissue. Therefore, we report the dose being delivered to the proximal 1 mm [D1mm ] and calculate therapeutic treatment depths [R90, R80 ] as a function of applicator size and design. For example, the smaller spherical applicators (≤ 30 mm) feature an aluminum filter to harden the spectrum. The present delivery workflow of IORT is similar to the initial MammoSite experience in which image-based planning was not required and dose-volume histograms (DVH) were not computed because the treatment involved a single dwell position with treatment time being chosen from a look-up table. The objective of the current study is to perform a dosimetric analysis of the lumpectomy cavity using dose-volume parameters [V90 , V80 , V50 , Dmin , D1mm, and HI] and to evaluate the change in treatment depth [R80, R90 ] because of spherical applicator size and filter design. While percent depth dose has been previously published for the range of spherical applicators, this study computes the DVH to offer an alternative dosimetric perspective, and is the first to report treatment depth values for the applicators.
Fig. 1. An internal schematic of the INTRABEAM x-ray source courtesy of Carl Zeiss Meditec AG©.
Methods and Materials X-ray source The INTRABEAM x-ray generator accelerates a beam of electrons down a 100 mm drift tube towards a thin hemispherical gold target situated inside a hemispherical beryllium window.7 The steering coils, at the base of the source, oscillate the beam around the tube axis to create an approximately isotropic distribution of bremsstrahlung radiation.7 Figure 1 illustrates the internal components of the x-ray device used to produce 50 kV x-rays. Investigators have validated the depth dose characteristics of the x-ray source and have proposed guidelines for commissioning and quality assurance.2 , 3
Fig. 2. Shielded water phantom setup to verify the depth dose characteristics of the x-ray source courtesy of Carl Zeiss Meditec AG©.
JID: MDO
ARTICLE IN PRESS M.Y. Shaikh, A. Nalichowski and M.C. Joiner et al. / Medical Dosimetry xxx (xxxx) xxx
[mUS5Gb;October 9, 2019;11:9] 3
Fig. 3. A 3-dimensional representation of 2 isodose surfaces. The center sphere represents the spherical applicator which receives the prescription dose, and the outer sphere represents the PTV. Images courtesy of Carl Zeiss Meditec AG©.
Spherical applicators The INTRABEAM system has a finite range of spherical applicators that range from 15 to 50 mm in increments of 5 mm. These applicators are constructed of a water-equivalent plastic known as polyetherimide (ULTEM) with a hollow internal cavity. An aluminum filter is embedded between the applicator body and x-ray source (XRS) probe for the smaller spherical applicators (≤ 30 mm) to intentionally harden the beam by removing very low energy photons from the treatment spectrum.2 For larger applicators (> 30 mm), beam hardening is sufficiently performed by the applicator body and no aluminum attenuator is required.14 We performed a phantom study which considered a comprehensive range of clinically available applicators for the INTRABEAM system. INTRABEAM calibration verification The INTRABEAM x-ray source is calibrated by the manufacturer annually during servicing, and the vendor provides measured output data that will be used in this study to calculate dosimetric coverage. Therefore, we independently validated the manufacturer’s calibration by mounting the x-ray source on a precise 3-dimensional translational stage and making measurements using a fixed PTW model 34013 (Physikalisch Technische Werkstaetten, Freiburg, Germany) parallelplate ionization chamber. The aforementioned chamber has a width of 0.9 mm in
the direction of the dose fall-off, cavity radius of 1.45 mm and a sensitive volume of 0.005 cm3 , and is the only choice for verifying the output characteristics of the xray source when using the Zeiss water phantom.15 Figure 2 shows the water phantom setup used to validate the vendor-provided depth dose from 5 to 30 mm in intervals of 5 mm. The vendor provided calibration dataset extends to a depth of 45 mm, but the water phantom setup is limited to a depth of 30 mm. Dosimetry Figure 3 represents the spherical applicator positioned in the breast tissue and the diagram on the left represents the position of the applicator relative to the PTV. A prescription of 20 Gy is delivered to the surface of the applicator, and due to the steep dose gradients, the PTV receives a nonhomogenous dose. For the PTV, DVH metrics [V90 , V80 , V50 , D1mm , Dmin, and HI] were used to quantify the dose distribution and coverage. Vx is the volume of target tissue receiving x percent of the prescription dose, D1mm is absorbed dose in Gray to the tissue at 1 mm depth, Dmin is the minimum dose to the PTV at 10 mm depth, and HI is the ratio of the maximum and minimum dose in the PTV. The maximum PTV dose occurs at the surface of the implanted device, while the minimum dose occurs at a depth of 10 mm. The product of the bare source output and the applicator transfer function is used to generate a relative percent depth dose profile. Using the applicator specific radial percent depth dose, the therapeutic treatment depth in millimeters is calculated
Fig. 4. Vendor calibrated x-ray beam dose rate in water for a 50 kV x-ray source operated at 40 μA and validated with ionization chamber measurements in a specialized water phantom. The calibrated x-ray beam data was modeled by a power fit trend line using Microsoft Excel.
JID: MDO 4
ARTICLE IN PRESS
[mUS5Gb;October 9, 2019;11:9]
M.Y. Shaikh, A. Nalichowski and M.C. Joiner et al. / Medical Dosimetry xxx (xxxx) xxx
Fig. 5. (a) Relative percent depth dose from the surface of 15 to 50 mm in diameter spherical applicators with 5 mm intervals. (b) Comparison of relative percent depth dose from the surface of 30, 35, and 40 mm diameter spherical applicators.
using the 80 and 90% isodose line [R80, R90 ]. Furthermore, we fit the data to a power series using the trend line functionality of Microsoft Excel (2017, Redmond, Washington). The power series function format is given by Eq. (1). Where A and B are model parameters calculated to fit the dataset, x will represent depth in millimeters, and y represents the output dose rate in Gy/min. y = AeBx
(1)
Results and discussion Figure 4 shows the vendor-provided calibrated dose rate (Gy/min) in water for the bare x-ray source (i.e., no spherical applicator), ionization chamber measurements and a trendline of best fit. For the 5 mm depth measurement point, the difference between the calibrated dose rate and measured dose rate is less
than 5%, and for deeper depths, the difference was less than 3%. At the 5 mm depth, small positional devaitions have an amplified impact and deviations exceeding 3% are expected.5 The intention behind the measurements was to validate the vendor-provided calibration before performing the dosimetry study. Additionally, the dataset was fit to a power series function and the line of best fit was expressed mathematically on the figure (R2 = 0.9994). We found the change in dose rate with depth to be more rapid than the inverse square because of beam attenuation. Given the rapid change in dose with depth, a logarithmic scale is better suited to highlight the subtle differences between measurements and the factory calibration. Figure 5a shows the percent depth dose in water normalized to the applicator surface on a logarithmic scale, as differences between applicators are easily obscured on a linear scale. For smaller spherical applicators, the percent depth dose changes rapidly with depth because the impact of beam attention is more pronounced closer to the x-ray source. When evaluating the percent depth dose for a
ARTICLE IN PRESS
JID: MDO
[mUS5Gb;October 9, 2019;11:9]
M.Y. Shaikh, A. Nalichowski and M.C. Joiner et al. / Medical Dosimetry xxx (xxxx) xxx Table 2 Dosimetric metrics for the spherical INTRABEAM applicators
Table 1 A summary of the applicator size and dose rate Applicator diameter (mm) 15 20 25 30 35 40 45 50
5
TARGIT: Dose rate on applicator surface (Gy/min) 2.755 1.764 1.163 0.775 1.023 0.751 0.546 0.399
full spectrum of applicator sizes, the impact of the aluminum filter for the smaller applicators (≤ 30 mm) needs to be considered. Figure 5b compares the normalized percent depth dose of the 30, 35, and 40 mm diameter spherical applicators. The 30 mm spherical applicator is unique in that it features an aluminum filter not present in the 35 and 40 mm diameter spherical applicator. The aluminum attenuator increases penetration depth and offsets some of the effects of beam attention, such that the 30 and 40 mm spherical applicators have a similar normalized percent depth dose. Table 1 presents the dose rate at the surface of the applicator. The full range of spherical applicators can be separated into 2 groups: applicators less than 30 mm which feature an aluminum filter, and applicators greater than 30 mm without an aluminum filter. While analyzing each applicator group separately we observed that the dose rate on the applicator surface decreases with increasing applicator diameter. However, when we compare all applicators as a single group we observe that the surface dose rates of the 35 mm applicators are higher than the 30 mm spherical applicator. This effect can be explained by the presence of the aluminum filter, which increases penetration depth [R80, R90 ] but reduces the dose rate at the surface of the applicator. Figure 6 shows the DVH for the PTV according to spherical applicator diameter, and Table 2 summarizes the dose-volume metrics (V90 , V80 , V50 , HI, Dmin and D1mm ). For a prescription of 20 Gy to the surface of the applicator and applicators between 15 to 50 mm, the Dmin to the PTV ranges between 3.15 to 6.84 Gy, and D1mm ranges between 16.2 to 17.6 Gy. Smaller applicators have reduced Dmin and D1mm doses. The presence of the aluminum filter increases the Dmin and D1mm dose and can be observed by comparing the adjacent 30 and 35 mm diameter spherical applicators. The range of dose-volume parameters across all applicator sizes was: 2.6 to 6.9% for V90 , 5.5 to 15.1% for V80 , 21.1 to 55.6% for V50 . Additionally, for the applicators 15 to 50 mm HI ranged from 6.35 to 2.9, with smaller applicators exhibiting increased HI values. Armoogum et al.16 conducted a functional comparison
Applicator diameter (mm)
V90 (%)
V80 (%)
V50 (%)
HI
Dmin (Gy)
D1mm (Gy)
15 20 25 30 35 40 45 50
2.6 3.6 4.2 5.0 4.7 5.2 6.8 6.9
5.5 7.4 8.9 10.7 10.2 11.5 14.0 15.1
21.1 27.0 33.1 41.0 37.2 43.5 49.4 55.6
6.35 5.03 4.32 3.59 4.03 3.61 3.22 2.90
3.15 3.97 4.66 5.54 4.97 5.55 6.20 6.84
16.2 16.6 16.9 17.3 16.9 17.2 17.6 17.6
of four INTRABEAM x-ray sources and found similar output characteristics. Therefore, the dosimetric comparison in this study is relevant to all INTRABEAM users. The concept of therapeutic treatment depth comes from electron beam radiotherapy and has not been previously calculated for IORT treatments with a 50 kV X-ray source. Figure 7 presents the R90 , R80 and D1mm for the full range of spherical applicators. For applicators between 15 and 50 mm, R90 values ranged from 0.53 to 0.85 mm and R80 values ranged from 1.18 to 1.85 mm. When we separate the applicators into 2 groups: applicators less than 30 mm which feature an aluminum filter, and applicators greater than 30 mm without an aluminum filter, we observed that values for R90 , R80 and D1mm increase with increasing applicator size. When comparing adjacent applicator sizes of 30 and 35 mm from different groups, we observe that the presence of an aluminum filter increases R90 , R80, and D1mm . If the aluminum filter was removed we would expect a linearly decreasing trend in the aforementioned values. Numerous studies have shown that patients with a small margin width (< 1mm) benefit from postoperative radiation therapy,12 , 13 indicating the relevance of dose to the proximal tissue. This study is the first to compute the dose R90 , R80 , and D1mm as a function of applicator size and design. The delivery of IORT is unique in that it is performed within the OR without the use of CT-based planning, which limits the ability to produce a DVH for the lumpectomy cavity. The presented dosimetric dataset provides the framework for clinicians to estimate coverage for homogeneous water-equivalent breast tissue, and to our knowledge, a similar dosimetric comparison has not been reported for the INTRABEAM system. A limitation of our study is that we have modeled breast tissue as being water-equivalent, however, for low-energy photon sources, breast tissue is not perfectly water-equivalent. Ebert and Carruthers reported a reduction in dose of 4% for 40 mm depth when comparing INTRABEAM dose calculations in water to breast tissue.17 Additionally, the introduction of heterogeneities such as the bone and lung would modify the dose-volume coverage and is beyond the scope of this paper.
Fig. 6. DVH for the PTV according to applicator size.
ARTICLE IN PRESS
JID: MDO 6
[mUS5Gb;October 9, 2019;11:9]
M.Y. Shaikh, A. Nalichowski and M.C. Joiner et al. / Medical Dosimetry xxx (xxxx) xxx
Fig. 7. For each applicator size we report the therapeutic treatment depth in millimeters as defined by the 80 and 90% [R80, R90] isodose line and the dose to 1 mm depth [D1mm ] in gray.
Conclusions Accurate knowledge of the absorbed dose to water is crucial for achieving optimal treatment outcomes and managing normal tissue toxicities. To overcome the lack of CT-imaging data in IORT treatments, we have generated sample DVH’s to document dose to the PTV and report a range of dosimetric parameters (V90 , V80 , V50 , D1mm , and Dmin ) for the full range of spherical breast applicators (15 to 50 mm). The PTV receives a more homogeneous dose with increasing spherical applicator size because the impact of beam attenuation with depth is less pronounced for larger applicators. The presence of the aluminum filter in the smaller applicator increases minimum PTV dose but reduces the surface dose rate. Declaration of Competing Interest The authors declare no conflicts of interest. Acknowledgments The authors thank Theresa Kwiatkowski and Robin Scott for their assistance with technical editing and comments that greatly improved the manuscript. References 1. Siegel, R.L.; Miller, K.D.; Jemal, A. Cancer statistics, 2019. CA. Cancer. J. Clin. 69:7– 34; 2019. doi:10.3322/caac.21551. 2. Eaton, D.J. Quality assurance and independent dosimetry for an intraoperative x-ray device. Med. Phys. 39:6908–20; 2012. doi:10.1118/1.4761865. 3. Beatty, J.; Biggs, P.J.; Gall, K.; et al. A new miniature x-ray device for interstitial radiosurgery: Dosimetry. Med. Phys. 23:53–62; 1996. doi:10.1118/1.597791. 4. Liu, Q.; Schneider, F.; Ma, L.; et al. Relative Biologic Effectiveness (RBE) of 50 kV X-rays measured in a phantom for intraoperative tumor-bed irradiation. Int. J. Radiat. Oncol. Biol. Phys. 85:1127–33; 2013. doi:10.1016/j.ijrobp.2012.08.005.
5. Watson, P.G.F.; Bekerat, H.; Papaconstadopoulos, P.; et al. An investigation into the INTRABEAM miniature x-ray source dosimetry using ionization chamber and radiochromic film measurements. Med. Phys. 45:4274–86; 2018. doi:10. 1002/mp.13059. 6. Sethi, A.; Emami, B.; Small, W.; et al. Intraoperative radiotherapy with INTRABEAM: Technical and dosimetric considerations. Front. Ancol. 8:74; 2018. doi:10.3389/fonc.2018.0 0 074. 7. Yanch, J.C.; Harte, K.J. Monte Carlo simulation of a miniature, radiosurgery xray tube using the ITS 3.0 coupled electron-photon transport code. Med. Phys. 23:1551–8; 1996. doi:10.1118/1.597885. 8. Vaidya, J.S.; Joseph, D.J.; Tobias, J.S.; et al. Targeted intraoperative radiotherapy versus whole breast radiotherapy for breast cancer (TARGIT-A trial): An international, prospective, randomised, non-inferiority phase 3 trial. Lancet (London, England) 376:91–102; 2010. doi:10.1016/S0140- 6736(10)60837- 9. 9. Vaidya, J.S.; Baum, M.; Tobias, J.S.; et al. The novel technique of delivering targeted intraoperative radiotherapy (Targit) for early breast cancer. Eur. J. Surg. Oncol. 28:447–54; 2002. doi:10.1053/ejso.2002.1275. 10. Vaidya, J.S.; Wenz, F.; Bulsara, M.; et al. An international randomised controlled trial to compare TARGeted Intraoperative radioTherapy (TARGIT) with conventional postoperative radiotherapy after breast-conserving surgery for women with early-stage breast cancer (the TARGIT-A trial). Health. Technol. Assess. 20:13; 2016. https://njl-admin.nihr.ac.uk/document/download/2003454. 11. Ibbott, G.S Radiation Dosimetry: Electron beams with energies between 1 and 50 MeV (ICRU Report No. 35). Med. Phys. 12:813; 1985. doi:10.1118/1.595780. 12. Silverstein, M.J.; Lagios, M.D.; Groshen, S.; et al. The influence of margin width on local control of ductal carcinoma in situ of the breast. N. Engl. J. Med. 340:1455–61; 1999. doi:10.1056/NEJM199905133401902. 13. Dunne, C.; Burke, J.P.; Morrow, M.; et al. Effect of margin status on local recurrence after breast conservation and radiation therapy for ductal carcinoma in situ. J. Clin. Oncol. 27:1615–20; 2009. doi:10.120 0/JCO.20 08.17.5182. 14. Keshtgar, M.R.S.; Vaidya, J.S.; Tobias, J.S.; et al. Targeted intraoperative radiotherapy for breast cancer in patients in whom external beam radiation is not possible. Int. J. Radiat. Oncol. Biol. Phys. 80:31–8; 2011. doi:10.1016/j.ijrobp.2010. 01.045. 15. Watson, P.G.F.; Popovic, M.; Seuntjens, J. Determination of absorbed dose to water from a miniature kilovoltage x-ray source using a parallel-plate ionization chamber. Phys. Med. Biol. 63:15016; 2017. doi:10.1088/1361-6560/aa9560. 16. Armoogum, K.S.; Parry, J.M.; Souliman, S.K.; et al. Functional intercomparison of intraoperative radiotherapy equipment – Photon Radiosurgery System. Radiat. Oncol. 2:11; 2007. doi:10.1186/1748-717X-2-11. 17. Ebert, M.A.; Carruthers, B. Dosimetric characteristics of a low-kV intra-operative x-ray source: implications for use in a clinical trial for treatment of low-risk breast cancer. Med. Phys. 30:2424–31; 2003. doi:10.1118/1.1595611.