Physica Medica 31 (2015) 414e419
Contents lists available at ScienceDirect
Physica Medica journal homepage: http://www.physicamedica.com
Technical notes
A novel method for dose distribution registration using fiducial marks made by a megavoltage beam in film dosimetry for intensitymodulated radiation therapy quality assurance Shinichi Nakayama a, Hajime Monzen a, b, *, Yuuichi Oonishi a, Rika Mizote c, Hiraku Iramina d, Souichirou Kaneshige e, Takashi Mizowaki f a
Division of Clinical Radiology Service, Okayama Central Hospital, 6-3 Ishimakitacho, kitaku, Okayama 700-0017, Japan Department of Radiation Oncology, Graduate School of Medicine, Kinki University, Osaka, Japan Division of Clinical Radiology Service, Okayama Municipal Hospital, Okayama, Japan d Department of Nuclear Engineering, Graduate School of Engineering, Kyoto University, Kyoto, Japan e Department of Radiology, Okayama Central Hospital, Okayama, Japan f Department of Radiation Oncology and Image-Applied Therapy, Graduate School of Medicine, Kyoto University, Kyoto, Japan b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 October 2014 Received in revised form 28 January 2015 Accepted 2 February 2015 Available online 25 February 2015
Purpose: Photographic film is widely used for the dose distribution verification of intensity-modulated radiation therapy (IMRT). However, analysis for verification of the results is subjective. We present a novel method for marking the isocenter using irradiation from a megavoltage (MV) beam transmitted through slits in a multi-leaf collimator (MLC). Methods: We evaluated the effect of the marking irradiation at 500 monitor units (MU) on the total transmission through the MLC using an ionization chamber and Radiochromic Film. Film dosimetry was performed for quality assurance (QA) of IMRT plans. Three methods of registration were used for each film: marking by irradiating with an MV beam through slits in the MLC (MLC-IC); marking with a fabricated phantom (Phantom-IC); and a subjective method based on isodose lines (Manual). Each method was subjected to local g-analysis. Results: The effect of the marking irradiation on the total transmission was 0.16%, as measured by a ionization chamber at a 10-cm depth in a solid phantom, while the inter-leaf transmission was 0.3%, determined from the film. The mean pass rates for each registration method agreed within ±1% when the criteria used were a distance-to-agreement (DTA) of 3 mm and a dose difference (DD) of 3%. For DTA/DD criteria of 2 mm/3%, the pass rates in the sagittal plane were 96.09 ± 0.631% (MLC-IC), 96.27 ± 0.399% (Phantom-IC), and 95.62 ± 0.988% (Manual). Conclusion: The present method is a versatile and useful method of improving the objectivity of film dosimetry for IMRT QA. © 2015 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.
Keywords: Film dosimetry Radiochromic film Quality assurance Dose distribution
Introduction Photographic film is widely used for the dose distribution verification of intensity-modulated radiation therapy (IMRT). A more recent development is the use of a multi-dimensional detector and an electronic portal imaging device (EPID) [1e2]. For
* Corresponding author. Department of Radiation Oncology, Graduate School of Medicine, Kinki University, 377-2 Oonohigashi Sayama-shi, Osaka 589-8511, Japan. Tel.: þ81 72 366 0221; fax: þ81 72 366 0206. E-mail address: [email protected] (H. Monzen).
example, there are diode detector arrays of MapCheck2 (Sun Nuclear) and ion chamber detectors of MatiXX (IBA Dosimetry) or OCTAVIUS 729 (PTW) in two-dimensional detectors. And there are detector arrays of ArcCHECK (Sun Nuclear) or Delta4 (ScandiDos) in three-dimensional detectors. However, because of its lowintroduction cost and high spatial resolution, photographic film is still useful for dose distribution verification. Moreover, photographic film can be used even in facilities that lack a film processor because radiochromic film (RCF) can be used under lighted conditions and needs no processing. In addition to requiring less knowhow than radiographic films (RGFs), such as Kodak extended dose range 2 (EDR2) and XV2 film, RCF has lower energy dependence,
http://dx.doi.org/10.1016/j.ejmp.2015.02.003 1120-1797/© 2015 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.
S. Nakayama et al. / Physica Medica 31 (2015) 414e419
and its wide energy range and composition are similar to human soft tissue, compared with RGF. The Gafchromic EBT2 and EBT3 could also be applied to the electron and proton beams as well as photon beams [3,4,5]. RCF or RGF has, however, uncertainties of film characteristics or scanning. van Buttum et al. [6] reported RGF has an uncertainty of 2% even when an integrated protocol is used. Also, Borca et al. [7] reported Gafchromic EBT3 has an uncertainty of 1.7%. To verify the obtained dose distribution it is necessary to register the dose distribution from the treatment planning system with the dose distribution of the film. Two common methods include marking the film during measurement to identify the isocenter and manual registration performed by comparing isodose lines and dose profiles. In the former case, the positional location of the film can change with respect to the phantom after marking. In the latter case, subjectivity of the observers can lead to different results of analysis, even using the same data set. Also, there are a couple of methods to make it objective (like e.g. automated registration algorithms). Here, we present a method of isocenter marking that uses a megavoltage (MV) beam to irradiate the film through four small slit-like fields made by a multi-leaf collimator (MLC). This method eliminates the subjectivity of the observers while facilitating marking. Winkler et al. [8] reported that the uncertainty of phantom positioning is 0.5 mm, while the uncertainty of using fiducial marks is 0.4 mm. In this study, the uncertainty of phantom positioning could be decreased because marking with the MV beam was conducted after the phantom was set. This means that the marking was also independent of the quality control of the lasers used for the phantom setup. However, MV beam transmission through the MLC could potentially affect the dose distribution verification. In this study, we evaluated the effect of the marking irradiation on the dose distribution and performed a g-analysis of three registration methods for IMRT QA. Materials and methods Equipment and materials For dose distribution verification, we used a four-dimensional radiation therapy system, Vero4DRT (Mitsubishi Heavy Industries, Tokyo, Japan), with an RT-3000-New phantom (R-Tech, Tokyo, Japan) and Gafchromic EBT3 film (batch number A02061302,
415
International Specialty Products, Wayne, NJ, USA). A flatbed scanner, the Epson ES-10000G (Seiko Epson Corporation, Suwa, Nagano, Japan), was used to read the irradiated film, and the digitized images were analyzed using the DD-system software (R-Tech, Tokyo, Japan). The MLC of the Vero4DRT has 30 pairs of leaves (5-mm width and 110-mm height) under a fixed jaw (15 15 cm2). The MLC is a single-focus type with tongue-and-groove construction. Nakamura et al. [9] reported the maximum inter- and intra-leaf transmission is 0.21% and 0.12%, respectively, and the mean transmission is 0.11%. Fiducial marking using an MLC As illustrated in Figs. 1 and 2, fiducial marks were produced on the film by irradiation through four tiny slits in the MLC before or after the treatment planning irradiation. With this setup, there was no need to pay particular attention to the film orientation because the marking could be performed after the phantom setup. Errors in the laser-mediated setup of the phantom could also be decreased. The fiducial marks make it possible to determine the right-left (RL) axes and superior-inferior (SI) axis coordinate in the coronal plane, or the anterior-posterior (AP) axis and the SI axis in the sagittal plane of MLC independently of the observers’ positions. However, random errors from the variability in MLC positioning remained. Nakamura et al. [9] reported the leaf position accuracy was 0.0 ± 0.1 mm, ranging from 0.3 to 0.2 mm at four gantry angles of 0 , 90 , 180 , and 270 . Effect of the marking irradiation on the total transmission through the MLC Scattering of the marking irradiation may contribute to the total transmission through the MLC and affect the IMRT dose distribution. Therefore, we evaluated the effect of the marking irradiation on the total transmission through the MLC using a Farmer-type ionization chamber and RCF. A PTW 30013-Farmer ionization chamber (PTW, Freiburg, Germany) was set at a source axis distance (SAD) of 100 cm and at a 10-cm depth in a water-equivalent solid phantom (TM phantom, Taisei Medical Co, Osaka, Japan), and then given 500 monitor units (MU) of fiducial marking irradiation. As a control, the measurement was repeated with 500 MU of irradiation through a completely open MLC (15 15 cm2). The percentage of the marking irradiation that reached the phantom was
Figure 1. Leaf sequence for fiducial marking using a multi-leaf collimator (MLC). Slit-like fields (1-mm gap width) were created using two segments, each delivering 250 monitor units (MU). The beam angles for the coronal and sagittal planes were 0 and 90 , respectively.
416
S. Nakayama et al. / Physica Medica 31 (2015) 414e419
Figure 2. Irradiated film with fiducial marks for quality assurance. Four slit-like fiducial marks were produced by irradiation through a multi-leaf collimator (MLC). The isocenter of the radiation is equidistant between the vertical and horizontal fiducial marks. The Vero4DRT (Mitsubishi Heavy Industries, Tokyo, Japan) has 30 leaves in its MLC, and the isocenter is shifted by two leaves in the vertical direction. The MLC of the Vero4DRT does not rotate. The four point-like fiducial marks were derived from the fabricated phantom. Grid lines connecting these point marks intersect the center of the phantom.
then determined as the ratio of the two measurements. We also evaluated the transmission through a completely closed MLC. EBT3 film was set at an SAD of 100 cm at a 10-cm depth in a solid phantom, and then given 500 MU of marking irradiation followed by 19,500 MU (total 20,000 MU) of irradiation through the closed MLC. This was compared with an irradiation of 20,000 MU through the closed MLC and no marking irradiation. The films were analyzed by converting optical density to dose using a dose calibration curve. Incidentally, dose calibration curve was prepared as follows [4,5,10]: A sheet of EBT3 film was cut into 16 pieces of 5 6 cm2. Pieces of EBT3 films were set at an SAD of 100 cm at a 10-cm depth in a solid phantom. The irradiation was performed for dose ranging from 0 to 400 MU (: 0 to approximately 310 cGy) with 14 steps of 10 10 cm2 square field. Especially, in low dose region were performed with a step of 25 MU. Relationships of absorbed dose with each MU were determined from a Farmer-type ionization chamber. Film reading Dose verification using RCF is complicated by a number of factors, including a gradual increase in optical density that continues for several hours after irradiation, sensitivity that is not homogeneous, and dependence on temperature, scan direction, and positioning of the film in the scanner [11,12,13]. To minimize these effects, the RCF was stored at low temperature and humidity before and after the irradiation and was read 24 h after irradiation. Five test scans were conducted before the actual film scanning. The four fiducial marks were aligned with two perpendicular wires affixed to the flatbed scanner to reduce variability in the positioning of the film. Black weights were set around the film to minimize film reflectance and noise from scattered light. Finally, the reading direction was standardized in a verification sequence. The scanner settings were 48-bit color, and three scans. As well, the scan spatial resolution was set to 150 dpi to one-half smaller than the calculated data of 1 mm intervals. The collection data used a red channel, and used a 3 3 median filter to decrease the spike noise.
coronal and sagittal planes through the isocenter were used for verification. For each film, registration of the dose distribution was performed using: 1) fiducial marks (Fig. 2, second panel) generated by a pin-processed phantom (Phantom-IC); 2) fiducial marks (Fig. 2, first panel) generated by irradiation through slits in the MLC (MLCIC); and 3) isodose lines and dose profiles (Manual method). In the manual method, the positioning of the film depended on the subjectivity of the observers. Each registration method was subjected to local g-analysis. gcriteria were a distance-to-agreement (DTA) of 3 mm and a dose difference (DD) of 3%, or a DTA of 2 mm and a DD of 3% [14]. Analyzed areas were the entire film. The dose distributions determined during treatment planning and during actual treatment were normalized to the point dose at the isocenter. Five local ganalyses were conducted for each of the registration methods, and the means and standard deviations of the pass rates were determined. These analyses were conducted by three separate radio technologists. Results Effect of the marking irradiation on the total transmission through the MLC The relative transmission that reached the phantom during marking irradiation was 0.16%, while the relative transmission through a completely closed MLC was 0.14%, as measured by an ionization chamber. Figure 3a compares the inlineedirection profile of film subjected to both marking irradiation (500 MU) and 19,500 MU of irradiation through a closed MLC with the profile of film subjected to 20,000 MU of only irradiation through a closed MLC. The profiles agree within ±0.03%. A representative example of an irradiated film is shown in Fig. 3b. The inter- and intra-leaf transmissions, determined from the film measurements, were ~0.3% and ~0.15%, respectively. Results of g-analysis of each registration method
Relative dose distribution analysis and evaluation Six cases of local prostate IMRT, each with seven-field irradiation, were used for film dosimetry for composite IMRT QA. The
The results of g-analyses of the Phantom-IC, MLC-IC, and Manual methods for each patient are shown in Figs. 4 and 5. A summary of the data, aggregated for all patients, is shown in
S. Nakayama et al. / Physica Medica 31 (2015) 414e419
Figure 3. Effect of marking irradiation on total transmission through a multi-leaf collimator (MLC). (a) Inline-direction profiles of film exposed to 20,000 monitor units (MU) of irradiation through a closed MLC (dashed line) and film exposed to 500 MU of marking irradiation plus 19,500 MU of irradiation through a closed MLC (solid line). (b) Representative example of film exposed to 500 MU of marking irradiation plus 19,500 MU of irradiation through a closed MLC; the inline direction is shown.
Table 1. The standard deviations of the pass rates were generally greater for the sagittal plane than for the coronal plane, particularly with respect to the 2 mm/3% criteria. These analyses were performed by three independent radio technologists. Using the Manual method, some pass rates by these separate observers differed more than 3% using the 2 mm/3% criteria, whereas the pass rates agreed within 1% using the 3 mm/3% criteria. Discussion As measured by an ionization chamber, the difference in total irradiation reaching the phantom during the marking irradiation versus a completely closed MLC was 0.02%. This difference is presumably due to scattered marking radiation. If it was estimated the absorbed dose by total transmission, the absorbed dose at a 10-cm depth due to the scattered 500 MU marking irradiation was 0.6 cGy. This corresponds to 0.3% of the planned treatment of 2 Gy by IMRT. The inlineedirection profiles of film exposed to marking irradiation plus irradiation through a closed MLC and film exposed to irradiation only through a closed MLC agreed within ±0.03%. This agreement extended over most of the profile region, even though transmission through the closed MLC is greater at certain locations.
417
Figure 4. Results of g-analyses of three different registration methods for quality assurance of six cases of local prostate intensity modulated radiation therapy. The dose distribution was registered using fiducial marks generated by either a pin-processed phantom (Phantom-IC), or by irradiation through slits in a multi-leaf collimator (MLC-IC), or manually according to isodose lines and dose profiles (Manual). (a) Coronal plane. (b) Sagittal plane. Five analyses were conducted for each patient with each method by three independent radio technologists. The means and standard deviations of the pass rates are shown. Criteria used: dose difference 3%, distance-to-agreement 3 mm, threshold 40%.
The small difference between the profiles is likely due to signal noise from the scanner and uncertainty arising from the inhomogeneity of the sensitive layer of the film. As determined from the optical density of the irradiated film, the dose received by the phantom from 20,000 MU of irradiation through the closed MLC was 30 cGy. Thus, any blushing of the film caused by scattered marking irradiation can be considered negligible. The Vero4DRT has a maximum field size of 15 15 cm2, with a fixed jaw. The transmission field is shaped by the MLC, the leaf height of which is 11 cm for shielding. The method used here can be used with other radiation therapy systems that use an MLC. Systems that rotate the MLC would move the slit around as needed and may improve the QA. The designs and characteristics of MLCs differ by manufacturer. Hug et al. [15] reported that among MLCs made by Elekta, Siemens, and Varian, the greatest transmission was 2.7%. Although the transmission doses of other MLCs may be greater than that of the Vero4DRT, the method described here may still be used if blushing of the film by the marking irradiation is controlled not only by the MLC but also by the jaw. However, for most of the linear accelerators there is a small shift of CAX when rotating the collimator (between 0.5 and 1.0 mm).
418
S. Nakayama et al. / Physica Medica 31 (2015) 414e419
Figure 5. Results of g-analyses of three different registration methods for quality assurance of six cases of local prostate intensity modulated radiation therapy. Details are the same as for Fig. 4, except that the distance-to-agreement criterion was 2 mm.
The mean pass rates for each registration method agreed within ±1% when a DTA criterion of 3 mm was used for the g-analyses. The standard deviations ranged from 0.2% to 0.4%, with those for the fiducial marking methods (Phantom-IC and MLC-IC) being slightly smaller than the Manual method. Under the criteria used, some position variation was admissible; thus, differences in the determination of the isocenter by laser (e.g., Phantom-IC method), marking by MLC (MLC-IC method), or by observer's positions (Manual method) were relatively insignificant. The means and standard deviations of the pass rates differed slightly under a DTA criterion of 2 mm. The Manual method showed sensible decreasing pass rates and increasing standard deviations. The purpose of the DTA is to detect declinations in the phantom Table 1 Summary of the data shown in Figs. 4 and 5, aggregated for all patient. Shown are the means and standard deviations of the pass rates for each registration method using as a criterion a distance to agreement of 3 or 2 mm. Film plane
Criteria for g-analysis DTAa/DDb
Registration method
Coronal
3 2 3 2
97.67 96.35 97.64 96.27
Sagittal a b
mm/3% mm/3% mm/3% mm/3%
Phantom-IC ± ± ± ±
0.183 0.279 0.274 0.399
MLC-IC 97.67 96.32 97.59 96.09
± ± ± ±
Manual 0.377 0.411 0.256 0.631
97.96 96.48 97.84 95.62
± ± ± ±
0.310 0.702 0.306 0.988
DTA: distance-to-agreement. DD: Dose difference Other criteria: dose difference 3%, dose threshold 40%.
position and geometric field position of the radiation therapy equipment. Apparent declinations due to the subjectivity of observers also have a large effect when the DTA criterion is less than 2 mm. Although the pass rates varied no more than ~3% between observers, they could conceivably differ more if the treatment planning area were greater than the simple round shapes of the dose distributions of local prostate IMRT used in this study. The standard deviations of the pass rates tended to be greater in the sagittal plane than in the coronal plane. Although there are no organs at risk near the prostate in the coronal plane, the dose gradient at the marginal high dose region is steep in the sagittal plane; thus, position errors had a greater effect in the sagittal plane. In head-and-neck IMRT, which requires complex dose distribution planning due to the number of organs at risk, position errors would have large effects on the QA results. With fiducial marking, some variations can occur due to misrecognition of the marks. The slitlike marks generated by the MLC are larger and more variable than the point-like marks from the fabricated phantom. However, for both types of fiducial marks, the standard deviations were less than those of the Manual method, which depends on the subjectivity of the observers. In addition, registration was easier and the g-analyses were performed more rapidly by using fiducial marks. IMRT QA relies heavily on g-analysis, which permits a quantitative evaluation expressed as a pass rate and mean g-value. However, the values can be affected arbitrarily if the registration depends on the subjectivity of the observers, especially under a tight DTA requirement. The purpose of IMRT QA is to ensure the reproducibility of an approved treatment plan and to detect errors arising from the treatment planning system and the linear accelerator. When registration depends on observer subjectivity, position errors are not properly evaluated. Moreover, no standard exists for optimal pass rates of g-analysis, and pass rates may differ among treatment plans, optimization protocols, and systems. When establishing an evaluation protocol, it makes sense to reject observer subjectivity and use fiducial marks for registration. Creating fiducial marks by irradiation with an MV beam is a versatile and useful method for film dosimetry for IMRT QA. Conclusion We evaluated the use of an MV beam to create fiducial marks for film dosimetry. Marking of the film by irradiation through an MLC is a simple method of determining the film coordinates with respect to the radiation therapy equipment without need for laser quality control of phantom positioning. The effects of scattered marking irradiation are negligible, and registration does not depend on the subjectivity of the observers. This facilitates the use of film dosimetry for IMRT QA. References [1] Greer PB, Carmen CP. Dosimetric properties of an amorphous silicon electronic portal imaging device for verification of dynamic intensity modulated radiation therapy. Med Phys 2003;30(7):1618e27. [2] Matsumoto K, Okumura M, Asai Y, Shimomura K, Tamura M, Nishimura Y. Dosimetric properties and clinical application of an a-Si EPID for dynamic IMRT quality assurance. Radiological Phys Technol 2013;6(1):210e8. [3] Arjomandy B, Tailor R, Anand A, Sahoo N, Gillin M, Prado K, et al. Energy dependence and dose response of Gafchromic EBT2 film over a wide range of photon, electron, and proton beam energies. Med Phys 2010;37(5):1942e7. [4] Sorriaux J, Kacperek A, Rossomme S, Lee JA, Bertrand D, Vynckier S, et al. Evaluation of Gafchromic® EBT3 films characteristics in therapy photon, electron and proton beams. Phys Medica 2013;29(6):599e606. [5] Farah Nicolas, Francis Ziad, Abboud Marie. Analysis of the EBT3 Gafchromic film irradiated with 6 MV photons and 6 MeV electrons using reflective mode scanners. Phys Medica 2014;30:708e12. [6] van Battum LJ, Hoffmans D, Piersma H, Heukelom S. Accurate dosimetry with GafChromic EBT film of a 6MV photon beam in water: what level is achievable? Med Phys 2008;35(2):704e16.
S. Nakayama et al. / Physica Medica 31 (2015) 414e419 [7] Casanova Borca V, Pasquino M, Russo G, Grosso P, Cante D, Sciacero P, et al. Dosimetric characterization and use of GAFCHROMIC EBT3 film for IMRT dose verification. J Appl Clin Med Phys 2013;14(2):4111. [8] Winkler P, Zurl B, Guss H, Kindl P, Stuecklschweiger G. Performance analysis of a film dosimetric quality assurance procedure for IMRT with regard to the employment of quantitative evaluation methods. Phys Med Biol 2005;50(4): 643. [9] Nakamura M, Sawada A, Ishihara Y, Takayama K, Mizowaki T, Kaneko S, et al. Dosimetric characterization of a multileaf collimator for a new fourdimensional image-guided radiotherapy system with a gimbaled x-ray head, MHI-TM2000. Med Phys 2010;37(9):4684e91. [10] Avanzo M, Drigo A, Ren Kaiser S, Roggio A, Sartor G, Chiovati P, et al. Dose to the skin in helical tomotherapy: results of in vivo measurements with radiochromic films. Phys Medica 2013;29(3):304e11.
419
[11] Niroomand-Rad A, Blackwell CR, Coursey BM, Gall KP, Galvin JM, McLaughlin WL, et al. Radiochromic film dosimetry: recommendations of AAPM radiation therapy committee task group 55. Med Phys 1998;25(11):2093e115. [12] Lynch BD, Kozelka J, Ranade MK, Li JG, Simon WE, Dempsey JF. Important considerations for radiochromic film dosimetry with flatbed CCD scanners and EBT GAFCHROMIC film. Med Phys 2006;33(12):4551e6. [13] Desroches J, Bouchard H, Lacroix F. Potential errors in optical density measurements due to scanning side in EBT and EBT2 Gafchromic film dosimetry. Med Phys 2010;37(4):1565e70. [14] Stock M, Kroupa B, Georg D. Interpretation and evaluation of the g index and the g index angle for the verification of IMRT hybrid plans. Phys Med Biol 2005;50(3):399. [15] Huq MS, Das IJ, Steinberg T, Galvin JM. A dosimetric comparison of various multileaf collimators. Phys Med Biol 2002;47(12). N159.
Contents lists available at ScienceDirect
Physica Medica journal homepage: http://www.physicamedica.com
Technical notes
A novel method for dose distribution registration using fiducial marks made by a megavoltage beam in film dosimetry for intensitymodulated radiation therapy quality assurance Shinichi Nakayama a, Hajime Monzen a, b, *, Yuuichi Oonishi a, Rika Mizote c, Hiraku Iramina d, Souichirou Kaneshige e, Takashi Mizowaki f a
Division of Clinical Radiology Service, Okayama Central Hospital, 6-3 Ishimakitacho, kitaku, Okayama 700-0017, Japan Department of Radiation Oncology, Graduate School of Medicine, Kinki University, Osaka, Japan Division of Clinical Radiology Service, Okayama Municipal Hospital, Okayama, Japan d Department of Nuclear Engineering, Graduate School of Engineering, Kyoto University, Kyoto, Japan e Department of Radiology, Okayama Central Hospital, Okayama, Japan f Department of Radiation Oncology and Image-Applied Therapy, Graduate School of Medicine, Kyoto University, Kyoto, Japan b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 October 2014 Received in revised form 28 January 2015 Accepted 2 February 2015 Available online 25 February 2015
Purpose: Photographic film is widely used for the dose distribution verification of intensity-modulated radiation therapy (IMRT). However, analysis for verification of the results is subjective. We present a novel method for marking the isocenter using irradiation from a megavoltage (MV) beam transmitted through slits in a multi-leaf collimator (MLC). Methods: We evaluated the effect of the marking irradiation at 500 monitor units (MU) on the total transmission through the MLC using an ionization chamber and Radiochromic Film. Film dosimetry was performed for quality assurance (QA) of IMRT plans. Three methods of registration were used for each film: marking by irradiating with an MV beam through slits in the MLC (MLC-IC); marking with a fabricated phantom (Phantom-IC); and a subjective method based on isodose lines (Manual). Each method was subjected to local g-analysis. Results: The effect of the marking irradiation on the total transmission was 0.16%, as measured by a ionization chamber at a 10-cm depth in a solid phantom, while the inter-leaf transmission was 0.3%, determined from the film. The mean pass rates for each registration method agreed within ±1% when the criteria used were a distance-to-agreement (DTA) of 3 mm and a dose difference (DD) of 3%. For DTA/DD criteria of 2 mm/3%, the pass rates in the sagittal plane were 96.09 ± 0.631% (MLC-IC), 96.27 ± 0.399% (Phantom-IC), and 95.62 ± 0.988% (Manual). Conclusion: The present method is a versatile and useful method of improving the objectivity of film dosimetry for IMRT QA. © 2015 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.
Keywords: Film dosimetry Radiochromic film Quality assurance Dose distribution
Introduction Photographic film is widely used for the dose distribution verification of intensity-modulated radiation therapy (IMRT). A more recent development is the use of a multi-dimensional detector and an electronic portal imaging device (EPID) [1e2]. For
* Corresponding author. Department of Radiation Oncology, Graduate School of Medicine, Kinki University, 377-2 Oonohigashi Sayama-shi, Osaka 589-8511, Japan. Tel.: þ81 72 366 0221; fax: þ81 72 366 0206. E-mail address: [email protected] (H. Monzen).
example, there are diode detector arrays of MapCheck2 (Sun Nuclear) and ion chamber detectors of MatiXX (IBA Dosimetry) or OCTAVIUS 729 (PTW) in two-dimensional detectors. And there are detector arrays of ArcCHECK (Sun Nuclear) or Delta4 (ScandiDos) in three-dimensional detectors. However, because of its lowintroduction cost and high spatial resolution, photographic film is still useful for dose distribution verification. Moreover, photographic film can be used even in facilities that lack a film processor because radiochromic film (RCF) can be used under lighted conditions and needs no processing. In addition to requiring less knowhow than radiographic films (RGFs), such as Kodak extended dose range 2 (EDR2) and XV2 film, RCF has lower energy dependence,
http://dx.doi.org/10.1016/j.ejmp.2015.02.003 1120-1797/© 2015 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.
S. Nakayama et al. / Physica Medica 31 (2015) 414e419
and its wide energy range and composition are similar to human soft tissue, compared with RGF. The Gafchromic EBT2 and EBT3 could also be applied to the electron and proton beams as well as photon beams [3,4,5]. RCF or RGF has, however, uncertainties of film characteristics or scanning. van Buttum et al. [6] reported RGF has an uncertainty of 2% even when an integrated protocol is used. Also, Borca et al. [7] reported Gafchromic EBT3 has an uncertainty of 1.7%. To verify the obtained dose distribution it is necessary to register the dose distribution from the treatment planning system with the dose distribution of the film. Two common methods include marking the film during measurement to identify the isocenter and manual registration performed by comparing isodose lines and dose profiles. In the former case, the positional location of the film can change with respect to the phantom after marking. In the latter case, subjectivity of the observers can lead to different results of analysis, even using the same data set. Also, there are a couple of methods to make it objective (like e.g. automated registration algorithms). Here, we present a method of isocenter marking that uses a megavoltage (MV) beam to irradiate the film through four small slit-like fields made by a multi-leaf collimator (MLC). This method eliminates the subjectivity of the observers while facilitating marking. Winkler et al. [8] reported that the uncertainty of phantom positioning is 0.5 mm, while the uncertainty of using fiducial marks is 0.4 mm. In this study, the uncertainty of phantom positioning could be decreased because marking with the MV beam was conducted after the phantom was set. This means that the marking was also independent of the quality control of the lasers used for the phantom setup. However, MV beam transmission through the MLC could potentially affect the dose distribution verification. In this study, we evaluated the effect of the marking irradiation on the dose distribution and performed a g-analysis of three registration methods for IMRT QA. Materials and methods Equipment and materials For dose distribution verification, we used a four-dimensional radiation therapy system, Vero4DRT (Mitsubishi Heavy Industries, Tokyo, Japan), with an RT-3000-New phantom (R-Tech, Tokyo, Japan) and Gafchromic EBT3 film (batch number A02061302,
415
International Specialty Products, Wayne, NJ, USA). A flatbed scanner, the Epson ES-10000G (Seiko Epson Corporation, Suwa, Nagano, Japan), was used to read the irradiated film, and the digitized images were analyzed using the DD-system software (R-Tech, Tokyo, Japan). The MLC of the Vero4DRT has 30 pairs of leaves (5-mm width and 110-mm height) under a fixed jaw (15 15 cm2). The MLC is a single-focus type with tongue-and-groove construction. Nakamura et al. [9] reported the maximum inter- and intra-leaf transmission is 0.21% and 0.12%, respectively, and the mean transmission is 0.11%. Fiducial marking using an MLC As illustrated in Figs. 1 and 2, fiducial marks were produced on the film by irradiation through four tiny slits in the MLC before or after the treatment planning irradiation. With this setup, there was no need to pay particular attention to the film orientation because the marking could be performed after the phantom setup. Errors in the laser-mediated setup of the phantom could also be decreased. The fiducial marks make it possible to determine the right-left (RL) axes and superior-inferior (SI) axis coordinate in the coronal plane, or the anterior-posterior (AP) axis and the SI axis in the sagittal plane of MLC independently of the observers’ positions. However, random errors from the variability in MLC positioning remained. Nakamura et al. [9] reported the leaf position accuracy was 0.0 ± 0.1 mm, ranging from 0.3 to 0.2 mm at four gantry angles of 0 , 90 , 180 , and 270 . Effect of the marking irradiation on the total transmission through the MLC Scattering of the marking irradiation may contribute to the total transmission through the MLC and affect the IMRT dose distribution. Therefore, we evaluated the effect of the marking irradiation on the total transmission through the MLC using a Farmer-type ionization chamber and RCF. A PTW 30013-Farmer ionization chamber (PTW, Freiburg, Germany) was set at a source axis distance (SAD) of 100 cm and at a 10-cm depth in a water-equivalent solid phantom (TM phantom, Taisei Medical Co, Osaka, Japan), and then given 500 monitor units (MU) of fiducial marking irradiation. As a control, the measurement was repeated with 500 MU of irradiation through a completely open MLC (15 15 cm2). The percentage of the marking irradiation that reached the phantom was
Figure 1. Leaf sequence for fiducial marking using a multi-leaf collimator (MLC). Slit-like fields (1-mm gap width) were created using two segments, each delivering 250 monitor units (MU). The beam angles for the coronal and sagittal planes were 0 and 90 , respectively.
416
S. Nakayama et al. / Physica Medica 31 (2015) 414e419
Figure 2. Irradiated film with fiducial marks for quality assurance. Four slit-like fiducial marks were produced by irradiation through a multi-leaf collimator (MLC). The isocenter of the radiation is equidistant between the vertical and horizontal fiducial marks. The Vero4DRT (Mitsubishi Heavy Industries, Tokyo, Japan) has 30 leaves in its MLC, and the isocenter is shifted by two leaves in the vertical direction. The MLC of the Vero4DRT does not rotate. The four point-like fiducial marks were derived from the fabricated phantom. Grid lines connecting these point marks intersect the center of the phantom.
then determined as the ratio of the two measurements. We also evaluated the transmission through a completely closed MLC. EBT3 film was set at an SAD of 100 cm at a 10-cm depth in a solid phantom, and then given 500 MU of marking irradiation followed by 19,500 MU (total 20,000 MU) of irradiation through the closed MLC. This was compared with an irradiation of 20,000 MU through the closed MLC and no marking irradiation. The films were analyzed by converting optical density to dose using a dose calibration curve. Incidentally, dose calibration curve was prepared as follows [4,5,10]: A sheet of EBT3 film was cut into 16 pieces of 5 6 cm2. Pieces of EBT3 films were set at an SAD of 100 cm at a 10-cm depth in a solid phantom. The irradiation was performed for dose ranging from 0 to 400 MU (: 0 to approximately 310 cGy) with 14 steps of 10 10 cm2 square field. Especially, in low dose region were performed with a step of 25 MU. Relationships of absorbed dose with each MU were determined from a Farmer-type ionization chamber. Film reading Dose verification using RCF is complicated by a number of factors, including a gradual increase in optical density that continues for several hours after irradiation, sensitivity that is not homogeneous, and dependence on temperature, scan direction, and positioning of the film in the scanner [11,12,13]. To minimize these effects, the RCF was stored at low temperature and humidity before and after the irradiation and was read 24 h after irradiation. Five test scans were conducted before the actual film scanning. The four fiducial marks were aligned with two perpendicular wires affixed to the flatbed scanner to reduce variability in the positioning of the film. Black weights were set around the film to minimize film reflectance and noise from scattered light. Finally, the reading direction was standardized in a verification sequence. The scanner settings were 48-bit color, and three scans. As well, the scan spatial resolution was set to 150 dpi to one-half smaller than the calculated data of 1 mm intervals. The collection data used a red channel, and used a 3 3 median filter to decrease the spike noise.
coronal and sagittal planes through the isocenter were used for verification. For each film, registration of the dose distribution was performed using: 1) fiducial marks (Fig. 2, second panel) generated by a pin-processed phantom (Phantom-IC); 2) fiducial marks (Fig. 2, first panel) generated by irradiation through slits in the MLC (MLCIC); and 3) isodose lines and dose profiles (Manual method). In the manual method, the positioning of the film depended on the subjectivity of the observers. Each registration method was subjected to local g-analysis. gcriteria were a distance-to-agreement (DTA) of 3 mm and a dose difference (DD) of 3%, or a DTA of 2 mm and a DD of 3% [14]. Analyzed areas were the entire film. The dose distributions determined during treatment planning and during actual treatment were normalized to the point dose at the isocenter. Five local ganalyses were conducted for each of the registration methods, and the means and standard deviations of the pass rates were determined. These analyses were conducted by three separate radio technologists. Results Effect of the marking irradiation on the total transmission through the MLC The relative transmission that reached the phantom during marking irradiation was 0.16%, while the relative transmission through a completely closed MLC was 0.14%, as measured by an ionization chamber. Figure 3a compares the inlineedirection profile of film subjected to both marking irradiation (500 MU) and 19,500 MU of irradiation through a closed MLC with the profile of film subjected to 20,000 MU of only irradiation through a closed MLC. The profiles agree within ±0.03%. A representative example of an irradiated film is shown in Fig. 3b. The inter- and intra-leaf transmissions, determined from the film measurements, were ~0.3% and ~0.15%, respectively. Results of g-analysis of each registration method
Relative dose distribution analysis and evaluation Six cases of local prostate IMRT, each with seven-field irradiation, were used for film dosimetry for composite IMRT QA. The
The results of g-analyses of the Phantom-IC, MLC-IC, and Manual methods for each patient are shown in Figs. 4 and 5. A summary of the data, aggregated for all patients, is shown in
S. Nakayama et al. / Physica Medica 31 (2015) 414e419
Figure 3. Effect of marking irradiation on total transmission through a multi-leaf collimator (MLC). (a) Inline-direction profiles of film exposed to 20,000 monitor units (MU) of irradiation through a closed MLC (dashed line) and film exposed to 500 MU of marking irradiation plus 19,500 MU of irradiation through a closed MLC (solid line). (b) Representative example of film exposed to 500 MU of marking irradiation plus 19,500 MU of irradiation through a closed MLC; the inline direction is shown.
Table 1. The standard deviations of the pass rates were generally greater for the sagittal plane than for the coronal plane, particularly with respect to the 2 mm/3% criteria. These analyses were performed by three independent radio technologists. Using the Manual method, some pass rates by these separate observers differed more than 3% using the 2 mm/3% criteria, whereas the pass rates agreed within 1% using the 3 mm/3% criteria. Discussion As measured by an ionization chamber, the difference in total irradiation reaching the phantom during the marking irradiation versus a completely closed MLC was 0.02%. This difference is presumably due to scattered marking radiation. If it was estimated the absorbed dose by total transmission, the absorbed dose at a 10-cm depth due to the scattered 500 MU marking irradiation was 0.6 cGy. This corresponds to 0.3% of the planned treatment of 2 Gy by IMRT. The inlineedirection profiles of film exposed to marking irradiation plus irradiation through a closed MLC and film exposed to irradiation only through a closed MLC agreed within ±0.03%. This agreement extended over most of the profile region, even though transmission through the closed MLC is greater at certain locations.
417
Figure 4. Results of g-analyses of three different registration methods for quality assurance of six cases of local prostate intensity modulated radiation therapy. The dose distribution was registered using fiducial marks generated by either a pin-processed phantom (Phantom-IC), or by irradiation through slits in a multi-leaf collimator (MLC-IC), or manually according to isodose lines and dose profiles (Manual). (a) Coronal plane. (b) Sagittal plane. Five analyses were conducted for each patient with each method by three independent radio technologists. The means and standard deviations of the pass rates are shown. Criteria used: dose difference 3%, distance-to-agreement 3 mm, threshold 40%.
The small difference between the profiles is likely due to signal noise from the scanner and uncertainty arising from the inhomogeneity of the sensitive layer of the film. As determined from the optical density of the irradiated film, the dose received by the phantom from 20,000 MU of irradiation through the closed MLC was 30 cGy. Thus, any blushing of the film caused by scattered marking irradiation can be considered negligible. The Vero4DRT has a maximum field size of 15 15 cm2, with a fixed jaw. The transmission field is shaped by the MLC, the leaf height of which is 11 cm for shielding. The method used here can be used with other radiation therapy systems that use an MLC. Systems that rotate the MLC would move the slit around as needed and may improve the QA. The designs and characteristics of MLCs differ by manufacturer. Hug et al. [15] reported that among MLCs made by Elekta, Siemens, and Varian, the greatest transmission was 2.7%. Although the transmission doses of other MLCs may be greater than that of the Vero4DRT, the method described here may still be used if blushing of the film by the marking irradiation is controlled not only by the MLC but also by the jaw. However, for most of the linear accelerators there is a small shift of CAX when rotating the collimator (between 0.5 and 1.0 mm).
418
S. Nakayama et al. / Physica Medica 31 (2015) 414e419
Figure 5. Results of g-analyses of three different registration methods for quality assurance of six cases of local prostate intensity modulated radiation therapy. Details are the same as for Fig. 4, except that the distance-to-agreement criterion was 2 mm.
The mean pass rates for each registration method agreed within ±1% when a DTA criterion of 3 mm was used for the g-analyses. The standard deviations ranged from 0.2% to 0.4%, with those for the fiducial marking methods (Phantom-IC and MLC-IC) being slightly smaller than the Manual method. Under the criteria used, some position variation was admissible; thus, differences in the determination of the isocenter by laser (e.g., Phantom-IC method), marking by MLC (MLC-IC method), or by observer's positions (Manual method) were relatively insignificant. The means and standard deviations of the pass rates differed slightly under a DTA criterion of 2 mm. The Manual method showed sensible decreasing pass rates and increasing standard deviations. The purpose of the DTA is to detect declinations in the phantom Table 1 Summary of the data shown in Figs. 4 and 5, aggregated for all patient. Shown are the means and standard deviations of the pass rates for each registration method using as a criterion a distance to agreement of 3 or 2 mm. Film plane
Criteria for g-analysis DTAa/DDb
Registration method
Coronal
3 2 3 2
97.67 96.35 97.64 96.27
Sagittal a b
mm/3% mm/3% mm/3% mm/3%
Phantom-IC ± ± ± ±
0.183 0.279 0.274 0.399
MLC-IC 97.67 96.32 97.59 96.09
± ± ± ±
Manual 0.377 0.411 0.256 0.631
97.96 96.48 97.84 95.62
± ± ± ±
0.310 0.702 0.306 0.988
DTA: distance-to-agreement. DD: Dose difference Other criteria: dose difference 3%, dose threshold 40%.
position and geometric field position of the radiation therapy equipment. Apparent declinations due to the subjectivity of observers also have a large effect when the DTA criterion is less than 2 mm. Although the pass rates varied no more than ~3% between observers, they could conceivably differ more if the treatment planning area were greater than the simple round shapes of the dose distributions of local prostate IMRT used in this study. The standard deviations of the pass rates tended to be greater in the sagittal plane than in the coronal plane. Although there are no organs at risk near the prostate in the coronal plane, the dose gradient at the marginal high dose region is steep in the sagittal plane; thus, position errors had a greater effect in the sagittal plane. In head-and-neck IMRT, which requires complex dose distribution planning due to the number of organs at risk, position errors would have large effects on the QA results. With fiducial marking, some variations can occur due to misrecognition of the marks. The slitlike marks generated by the MLC are larger and more variable than the point-like marks from the fabricated phantom. However, for both types of fiducial marks, the standard deviations were less than those of the Manual method, which depends on the subjectivity of the observers. In addition, registration was easier and the g-analyses were performed more rapidly by using fiducial marks. IMRT QA relies heavily on g-analysis, which permits a quantitative evaluation expressed as a pass rate and mean g-value. However, the values can be affected arbitrarily if the registration depends on the subjectivity of the observers, especially under a tight DTA requirement. The purpose of IMRT QA is to ensure the reproducibility of an approved treatment plan and to detect errors arising from the treatment planning system and the linear accelerator. When registration depends on observer subjectivity, position errors are not properly evaluated. Moreover, no standard exists for optimal pass rates of g-analysis, and pass rates may differ among treatment plans, optimization protocols, and systems. When establishing an evaluation protocol, it makes sense to reject observer subjectivity and use fiducial marks for registration. Creating fiducial marks by irradiation with an MV beam is a versatile and useful method for film dosimetry for IMRT QA. Conclusion We evaluated the use of an MV beam to create fiducial marks for film dosimetry. Marking of the film by irradiation through an MLC is a simple method of determining the film coordinates with respect to the radiation therapy equipment without need for laser quality control of phantom positioning. The effects of scattered marking irradiation are negligible, and registration does not depend on the subjectivity of the observers. This facilitates the use of film dosimetry for IMRT QA. References [1] Greer PB, Carmen CP. Dosimetric properties of an amorphous silicon electronic portal imaging device for verification of dynamic intensity modulated radiation therapy. Med Phys 2003;30(7):1618e27. [2] Matsumoto K, Okumura M, Asai Y, Shimomura K, Tamura M, Nishimura Y. Dosimetric properties and clinical application of an a-Si EPID for dynamic IMRT quality assurance. Radiological Phys Technol 2013;6(1):210e8. [3] Arjomandy B, Tailor R, Anand A, Sahoo N, Gillin M, Prado K, et al. Energy dependence and dose response of Gafchromic EBT2 film over a wide range of photon, electron, and proton beam energies. Med Phys 2010;37(5):1942e7. [4] Sorriaux J, Kacperek A, Rossomme S, Lee JA, Bertrand D, Vynckier S, et al. Evaluation of Gafchromic® EBT3 films characteristics in therapy photon, electron and proton beams. Phys Medica 2013;29(6):599e606. [5] Farah Nicolas, Francis Ziad, Abboud Marie. Analysis of the EBT3 Gafchromic film irradiated with 6 MV photons and 6 MeV electrons using reflective mode scanners. Phys Medica 2014;30:708e12. [6] van Battum LJ, Hoffmans D, Piersma H, Heukelom S. Accurate dosimetry with GafChromic EBT film of a 6MV photon beam in water: what level is achievable? Med Phys 2008;35(2):704e16.
S. Nakayama et al. / Physica Medica 31 (2015) 414e419 [7] Casanova Borca V, Pasquino M, Russo G, Grosso P, Cante D, Sciacero P, et al. Dosimetric characterization and use of GAFCHROMIC EBT3 film for IMRT dose verification. J Appl Clin Med Phys 2013;14(2):4111. [8] Winkler P, Zurl B, Guss H, Kindl P, Stuecklschweiger G. Performance analysis of a film dosimetric quality assurance procedure for IMRT with regard to the employment of quantitative evaluation methods. Phys Med Biol 2005;50(4): 643. [9] Nakamura M, Sawada A, Ishihara Y, Takayama K, Mizowaki T, Kaneko S, et al. Dosimetric characterization of a multileaf collimator for a new fourdimensional image-guided radiotherapy system with a gimbaled x-ray head, MHI-TM2000. Med Phys 2010;37(9):4684e91. [10] Avanzo M, Drigo A, Ren Kaiser S, Roggio A, Sartor G, Chiovati P, et al. Dose to the skin in helical tomotherapy: results of in vivo measurements with radiochromic films. Phys Medica 2013;29(3):304e11.
419
[11] Niroomand-Rad A, Blackwell CR, Coursey BM, Gall KP, Galvin JM, McLaughlin WL, et al. Radiochromic film dosimetry: recommendations of AAPM radiation therapy committee task group 55. Med Phys 1998;25(11):2093e115. [12] Lynch BD, Kozelka J, Ranade MK, Li JG, Simon WE, Dempsey JF. Important considerations for radiochromic film dosimetry with flatbed CCD scanners and EBT GAFCHROMIC film. Med Phys 2006;33(12):4551e6. [13] Desroches J, Bouchard H, Lacroix F. Potential errors in optical density measurements due to scanning side in EBT and EBT2 Gafchromic film dosimetry. Med Phys 2010;37(4):1565e70. [14] Stock M, Kroupa B, Georg D. Interpretation and evaluation of the g index and the g index angle for the verification of IMRT hybrid plans. Phys Med Biol 2005;50(3):399. [15] Huq MS, Das IJ, Steinberg T, Galvin JM. A dosimetric comparison of various multileaf collimators. Phys Med Biol 2002;47(12). N159.
