- Email: [email protected]
Evaluation of the robustness of 3-dimensional conformal technique with MLC position control into the planning target volume in stereotactic body radiotherapy for lung cancer
ARTICLE IN PRESS
JID: MDO
[mUS5Gb;May 25, 2019;11:46]
Medical Dosimetry xxx (xxxx) xxx
Medical Dosimetry journal homepage: www.meddos.org
Evaluation of the robustness of 3-dimensional conformal technique with MLC position control into the planning target volume in stereotactic body radiotherapy for lung cancer Masahide Saito, MS∗, Naoki Sano, PhD, Kengo Kuriyama, MD, PhD, Takafumi Komiyama, MD, PhD, Kan Marino, MD, Shinichi Aoki, MD, Yoshiyasu Maehata, MD, PhD, Ryo Saito, MD, Hidekazu Suzuki, RT, Yuki Shibata, RT, Koji Ueda, RT, Hiroshi Onishi, MD, PhD Department of Radiology, University of Yamanashi, Yamanashi 409-3898, Japan
a r t i c l e
i n f o
Article history: Received 25 March 2019 Accepted 24 April 2019 Available online xxx Keywords: Lung SBRT 3DCRT Robustness
a b s t r a c t The purpose of this study was to evaluate the robustness of 3-dimensional conformal technique with MLC position control into the planning target volume (PTV) in stereotactic body radiotherapy for lung cancer. Two techniques using fixed beams were compared; one technique involved setting the MLC position outside the PTV and was referred to as Plan “O.” Another technique involved setting the MLC position inside the PTV and was referred to as Plan “I.” Two tumor motions were simulated: (1) tumor motion on the internal target volume (ITV) boundary and (2) tumor motion on the PTV boundary. Ten-phase CT images that captured the tumor in respiratory motion were generated for 2 simulations. Then, 4-dimensional (4D) treatment planning was performed by using deformable image registration. The gross tumor volume (GTV) dose changes between the 4D accumulated dose and treatment planning dose were evaluated for Plan “O” and Plan “I,” respectively. For the simulation of tumor motion on the ITV boundary, the changes in GTV D50% were −0.10 ± 0.31% and −0.22 ± 0.26% (p < 0.05) for Plan “O” and Plan “I,” respectively. In the same manner, for the simulation of tumor motion on the PTV boundary, the changes in GTV D50% were −3.37 ± 2.16% and −3.68 ± 1.71% (p < 0.05). Our result suggested that the dose change would be negligible in a clinical situation where the tumor moves within the ITV margin for both techniques, while Plan “O” showed better robustness. © 2019 American Association of Medical Dosimetrists. Published by Elsevier Inc. All rights reserved.
Introduction Stereotactic body radiation therapy (SBRT) is commonly used for therapeutic management of patients with non–small cell lung cancer (NSCLC). A few earlier clinical trials that focused on SBRT to treat NSCLC have shown high local control with limited toxicity, and outcomes comparable to that of surgery.1,2 Most clinical studies have traditionally assessed the importance of 3D conformal radiotherapy (3DCRT), while intensity-modulated radiation therapy (IMRT) and volumetric-modulated arc therapy (VMAT) are still in the process of development. Although there are some reports that VMAT and IMRT usually show better dose distribution to help spare critical structures than 3DCRT,3,4 an interplay effect between respiration-induced tumor motion and multileaf colli-
∗ Reprint requests to Masahide Saito, MS, Department of Radiology, University of Yamanashi, 1110 Shimokato, Chuo-city, Yamanashi 409-3898, Japan. E-mail address: [email protected] (M. Saito).
mator (MLC) motion may affect the accuracy of the VMAT dose prescription.5-9 Furthermore, the accuracy of delivery using VMAT is affected by not only patient movement during treatment, but also MLC-positioning errors and limited accuracy of MLC modeling in the treatment planning system.10 Namely, VMAT has multiple layers of uncertainties in the treatment process when compared to 3DCRT. 3DCRT is a more simple technique than VMAT in SBRT for treatment of lung cancer, because this technique is not susceptible to the interplay effect between MLC leaves and tumor motion.10 Therefore, from this point of view, 3DCRT is more robust (=safe treatment technique) to these errors than VMAT. However, skill and experience of the clinician play an important role in the 3DCRT treatment planning. Generally, in order to improve dose coverage of the tumor, an MLC margin of several millimeters (that is “plus MLC margin”) is applied to each leaf that fits the shape of the planning target volume (PTV). Furthermore, some planners insert the MLC into the PTV (that is “minus MLC margin”) to avoid a hot spot for chest wall or to increase the GTV dose with decreasing the surrounding lung dose. Although this technique is
https://doi.org/10.1016/j.meddos.2019.04.007 0958-3947/© 2019 American Association of Medical Dosimetrists. Published by Elsevier Inc. All rights reserved.
ARTICLE IN PRESS
JID: MDO 2
[mUS5Gb;May 25, 2019;11:46]
M. Saito, N. Sano and K. Kuriyama et al. / Medical Dosimetry xxx (xxxx) xxx
Table 1 Target characteristics for all patient (n = 25) PTV margin (mm) ITV margin (mm) Tumor (GTV) volume (cc) Tumor density (g/cm3 ) PTV volume (cc) PT V/GT V ratio
All is 3.00 2.47 ± 1.47 (range: 0.00 to 6.00) 9.53 ± 5.63 (range: 1.74 to 21.50) 0.64 ± 0.22 (range: 0.27 to 1.01) 37.92 ± 18.10 (range: 11.55 to 79.30) 4.59 ± 1.50 (range: 8.19 to 2.52)
usually used in clinical situations, there have been no previous reports on whether the dose coverage of the gross tumor volume (GTV) is acceptable using “minus MLC margin.” The Radiation Therapy Oncology Group 0618 trial does not mention the need to use minus MLC margin. They simply recommend that field aperture size and shape should correspond nearly identically to the projection of the PTV along the beam’s eye view.11 On the other hand, the Japan Clinical Oncology Group (JCOG) 1408 trial permits to use the minus MLC margin,12 while it was not permitted in previous JCOG0403 trials.13 Although 3DCRT is important technique, there have been no reports for the robustness when we use the minus MLC margin. Therefore, the purpose of this study was to determine the robustness of treatment planning for SBRT of lung cancer using a technique with MLC position control into the PTV using 4D treatment planning. Materials and Methods Study design and treatment planning This is a retrospective study approved by the institutional review board. Twenty-five NSCLC patients treated with SBRT in our institute were enrolled. All patients were given breath-hold treatment with an Abches device, which is respiratory monitoring device with 2 respiratory indicators on the chest and abdomen.14 Although all treatment plans were initially created by Pinnacle3 treatment planning systems (Philips Radiation Oncology Systems, Madison, WI) with 3DCRT using 7 to 11 coplanar or noncoplanar beams, we modified all the plans in RayStation ver. 6.0 (RaySearch Laboratories, Stockholm, Sweden) to make the research progress smoothly. For all cases, we did not use a beam through the contralateral lung. All plans were generated for Elekta Synergy with Agility gantry head, which has 160 MLC leaves of 5 mm (Elekta AB, Stockholm, Sweden). The dose calculation algorithm was a collapsed cone convolution and the dose grid size was 2 mm for all dimensions. Dose constraints of organs at risk were set according to the JCOG0702 protocol.15 The patients received a dose covering 95% volume (D95) prescription of 50 Gy for PTV in 4 fractions. The internal target volume (ITV) margin was defined to use multiple CT images with breath holding. The PTV margin and the MLC margin were defined as 3 mm for all dimensions (note that the MLC was modified to the extent that it does not enter PTV region). Target characteristics for all patients are summarized in Table 1. In this study, two 3DCRT techniques were compared; one technique involved setting the MLC position outside the PTV and was referred to as Plan “O.” Another technique involved setting the MLC position inside the PTV and was referred to as Plan “I,” which is shown in Fig. 1. Each technique was applied to all beams for all patients; however, Plan “I” avoided the ITV region. Namely, for Plan “I,” the length between the MLC position and PTV edge was 1 mm for almost cases (maximum 3 mm). Tumor motion simulation and 4D treatment planning Our simulation was based on the hypothesis that the tumor moves randomly during the breath-holding treatment. Therefore, for the evaluation of robustness of the above two 3DCRT techniques, the motion of the 2 tumors was simulated by an in-house program. These types of motions included the following: (1) random motion on the ITV boundary and (2) random motion on the PTV boundary as the worst case. Figure 2 shows the typical simulation of the tumor motion. Overall, 10phase CT images that captured the tumor in random motion, were generated for 2 simulations. Each image was created using a deformable image registration (DIR) with a hybrid deformation algorithm which used image intensity and a region of interest (ROI) of the GTV referred to as the controlling ROI. These simulations were performed by RayStation and MATLAB 2016a (Mathworks, Natick, MA). Then, 4D treatment planning was performed by RayStation. First, the dose distribution was calculated for 10-phase images, respectively. Next, the dose of all phases was accumulated into the original CT image by using DIR. We used the Hybrid algorithm with an image intensity that focused on the ROI (PTV). For the dose accumulation, the weight of the dose distribution on each CT image was equal.
Fig. 1. Concept of two 3DCRT techniques. Part (a) is Plan “O” that involved setting the MLC position outside the PTV. Part (b) is Plan “I” that involved setting the MLC position inside the PTV. MLC shapes and dose distributions are shown for each plan.
Evaluation of the robustness The dose changes between the 4D accumulated dose and treatment planning dose were evaluated for Plan “O” and Plan “I,” respectively. For the dose change of the GTV, the maximum dose (Dmax ), mean dose (Dmean ), minimum dose (Dmin ) and each dose volume index (D98%, D50%, and D2%) were calculated. In the same manner, for the dose change of the PTV, maximum dose (Dmax ), mean dose (Dmean ), minimum dose (Dmin ) and each dose volume index (D95%, and D2%) were calculated. In addition, the change in conformity index was calculated for the 2 methods. Conformity index was calculated by the following equation; CI =
VRI TV
(1)
where VRI = reference isodose (50 Gy) volume and TV = target volume (PTV). All statistical analyses were performed using JMP Pro software ver. 13 (SAS Institute Inc., NC). Results Table 2 shows the absolute value of treatment planning dose and 4D accumulated dose for 2 planning techniques (Plan “O” and Plan “I”) for each dose index and conformity index. Figure 3 shows the percent change of target dose between each planning value and 4D accumulated dose. There was a significant difference in target dose changes between Plan “O” and Plan “I.” For all simulations, the PTV dose was more affected by the tumor motion than the GTV dose. For the simulation of random motion on the IT V boundary, the changes in GT V D50% were −0.10 ± 0.31% and −0.22 ± 0.26% (p < 0.05) for Plan “O” and Plan “I,” respectively. The changes in PTV D95% were −1.12 ± 0.98% and −1.48 ± 1.29% (p < 0.05) for Plan “O” and Plan “I,” respectively. In the same manner, for the simulation of random motion on the PTV boundary, the change in GTV D50% was −0.47 ± 0.43% and −0.88 ± 1.04% (p < 0.05) for Plan “O” and Plan “I,” respectively, and that of PTV D95% was −3.37 ± 2.16% and −3.68 ± 1.71% (p < 0.05) for Plan “O” and Plan “I,” respectively. Figure 4 shows a typical case (Patient 1) of the dose difference between the 4D accumulated dose and treatment planning dose for the 2 techniques. For the simulation of random motion on the ITV boundary, only small dose differences within ±1% were observed for both techniques. However, if the tumor moved onto the PTV boundary, the dose decreased 3% for both techniques.
Discussion The robustness of the SBRT to lung cancer using VMAT has already been investigated by several authors. Rao et al. investigated the impact of the intrafractional respiratory motion on the dose distribution of free-breathing lung SBRT using VMAT and fixedfield IMRT.6 They showed that the dose to the GTV was slightly decreased on average (1% of the prescription) in the 4D calculation compared with the 3D calculation. Ong et al. investigated the dosimetric impact of the interplay effect during RapidArc SBRT for lung cancers using flattening filter-free beams with different dose rates.7 Their result indicated that multiple arcs and more than 2 fractions reduced the interplay effect to a level that appeared
ARTICLE IN PRESS
JID: MDO
[mUS5Gb;May 25, 2019;11:46]
M. Saito, N. Sano and K. Kuriyama et al. / Medical Dosimetry xxx (xxxx) xxx
3
Fig. 2. The typical simulation of the tumor motion. Part (a) shows the random motion on the ITV boundary and part (b) shows random motion on the PTV boundary as the worst case. Ten-phase CT images that captured the tumor in random motion were generated for 2 simulations, respectively. Each image was created using a deformable image registration with hybrid deformation algorithm which used image intensity and a region of interest (ROI) of the GTV referred to as the controlling ROI. Table 2 The absolute value of treatment planning dose and 4D accumulated dose for 2 planning techniques (Plan “O” and Plan “I”) for each dose index. Furthermore, the conformity indexes for all doses are shown in the below of the table Absolute dose (Gy) of Plan “O”
GTV
Dmax Dmean Dmin D98% D50% D2% PTV Dmax Dmean Dmin D95% D2% Conformity index
Absolute dose (Gy) of Plan “I”
Planning dose
4D accumulated dose (Inside ITV)
4D accumulated dose (Outside ITV)
Planning dose
4D accumulated dose (Inside ITV)
70.80 ± 6.11 66.84 ± 4.35 57.74 ± 2.97 61.39 ± 2.59 67.16 ± 4.45 70.37 ± 5.90 70.83 ± 5.96 62.65 ± 2.76 44.79 ± 5.62 55.00 ± 0.00 69.83 ± 5.58 0.66 ± 0.09
70.54 ± 5.96 66.74 ± 4.32 57.57 ± 2.72 61.37 ± 2.46 67.09 ± 4.45 70.19 ± 5.83 70.60 ± 5.92 62.33 ± 2.73 43.82 ± 5.76 54.38 ± 0.54 69.66 ± 5.54 0.67 ± 0.09
70.40 ± 5.90 66.52 ± 4.25 56.63 ± 2.97 60.75 ± 2.43 66.84 ± 4.42 70.08 ± 5.77 70.40 ± 5.90 61.67 ± 2.66 41.98 ± 5.81 53.15 ± 1.19 69.53 ± 5.57 0.71 ± 0.09
73.29 ± 6.40 68.99 ± 4.18 58.15 ± 3.49 63.09 ± 2.44 69.32 ± 4.26 72.93 ± 6.26 73.29 ± 6.40 63.66 ± 2.45 43.17 ± 4.21 55.00 ± 0.02 72.21 ± 5.87 0.72 ± 0.09
73.02 ± 6.32 68.85 ± 4.13 57.73 ± 3.20 63.03 ± 2.19 69.17 ± 4.23 72.69 ± 6.16 73.02 ± 6.32 63.20 ± 2.50 42.33 ± 4.32 54.19 ± 0.70 71.90 ± 5.89 0.73 ± 0.09
unlikely to be clinically significant. For 3DCRT, although their effects may be reduced because of no MLC movement, there have been no reports on them. Furthermore, it is unclear whether Plan “O,” which involves setting the MLC position outside the PTV is a more robust technique when compared to Plan “I,” the technique that involves setting the MLC position inside the PTV. Therefore, we investigated this using the original 4D simulation for the retrospective SBRT cases in our institute. First, we created Plan “O” and Plan “I” for all patients. From the results of Table 2, the negative MLC margin technique (Plan “I”) increase the conformity and absolute dose for the GTV. Then, we simulated the random tumor motion on the ITV boundary for each patient. For the simulation, the dose change of Plan “I” was significantly larger than that of Plan “O.” It is likely that the decreased field boundary dose that derived from tumor motion was affected by the number of MLC leaves that shielded the PTV boundary for each port, indicating that Plan “O” is more robust than Plan “I.” In particular, the minimum dose mostly affected the tumor motion; this was similar to the finding in a previous study.9 It was thought that the minimum dose exists in the field boundary for most cases, which can be affected by the tumor motion. However, both changes would be negligible in clinical situations (the dose decreased to −0.69 % and −2.22% for the GTV and the PTV, respectively as shown in Fig. 3).
4D accumulated dose (Outside ITV) 72.78 ± 6.36 68.35 ± 4.25 56.40 ± 3.47 62.06 ± 2.17 68.72 ± 4.36 72.40 ± 6.23 72.78 ± 6.35 62.36 ± 2.60 40.61 ± 4.73 52.98 ± 0.93 71.69 ± 5.81 0.76 ± 0.09
We also simulated the random tumor motion on the PTV boundary for each patient to represent the worst case. The change of each dose index of this simulation was larger than that of the simulation of tumor motion on the ITV boundary for both techniques. It is probable that the simulation of tumor motion on the PTV boundary is more likely to have tumor tissue at the boundary of the MLC leaves. For this simulation, the dose decreased to −2.99 % and −6.27% for the GTV and the PTV, respectively as shown in Table 2. Therefore, our study indicated that it is acceptable to use the adjusted MLC margin on the clinical condition that the tumor moves within the ITV margin. However, it should be noted that there is a nonignorable impact to the target dose when the tumor has large motion beyond the ITV margin even using both 3DCRT techniques. Although it is common knowledge that the setting of appropriate ITV margin is essential for lung SBRT, we emphasize it again to reduce the influence. Our study had some limitations. First, a limited number of patients were enrolled in this study. Second, this study was an only a simulation using a treatment planning system, indicating that the results depended on the accuracy of a dose calculation algorithm and DIR. Finally, motion simulation was performed in only a few of the tumors. Future studies that include larger cohorts and assess other simulations such as those that use a respiration model with a baseline shift or respiration drift should be investigated.
JID: MDO 4
ARTICLE IN PRESS
[mUS5Gb;May 25, 2019;11:46]
M. Saito, N. Sano and K. Kuriyama et al. / Medical Dosimetry xxx (xxxx) xxx
Fig. 3. The percent difference between planning dose and 4D accumulated dose for each tumor simulation. The vertical axis is the percent difference, and lateral axis is each dose index. Parts (a) and (b) show the result of GTV dose index, and parts (c) and (d) show the result of PTV dose index. Parts (a) and (c) show the result of tumor simulation within ITV, and parts (b) and (d) show the result of tumor simulation without ITV. The significant change (p value < 0.05) in target dose changes between Plan “O” and Plan “I” is represented by ∗ .
References
Fig. 4. A typical case (Patient 1) of the dose difference between the 4D accumulated dose and treatment planning dose for the 2 techniques. Plan “O” (left side) and Plan “I” (right side) are shown. For the simulation of random motion on the ITV boundary (a), only small dose differences within ±1% were observed for both techniques. However, if the tumor moved onto the PTV boundary (b), the dose decreased 3% for both techniques.
Conclusions In this study, we compared the robustness of two 3DCRT techniques for lung SBRT (Plan “O” and Plan “I”). The dose change would be negligible in a clinical situation where the tumor moves within the ITV margin for both techniques, while Plan “O” showed better robustness. Our results provide useful information about 3DCRT field shape for a clinical SBRT trial of lung cancer. Conflicts of Interest There is no conflict of interest with regard to this manuscript.
1. Onishi, H; Shirato, H; Nagata, Y; et al. Stereotactic body radiotherapy (SBRT) for operable stage I non-small-cell lung cancer: Can SBRT be comparable to surgery. Int J Radiat Oncol Biol Phys 81:1352–8; 2011. 2. Nagata, Y; Takayama, K; Matsuo, Y; et al. Clinical outcomes of a phase I/II study of 48 Gy of stereotactic body radiotherapy in 4 fractions for primary lung cancer using a stereotactic body frame. Int J Radiat Oncol Biol Phys 63:1427–31; 2005. 3. Ong, CL; Verbakel, WF; Cuijpers, JP; et al. Stereotactic radiotherapy for peripheral lung tumors: A comparison of volumetric modulated arc therapy with 3 other delivery techniques. Radiother Oncol 97:437–42; 2010. 4. Zhang, GG; Ku, L; Dilling, TJ; et al. Volumetric modulated arc planning for lung stereotactic body radiotherapy using conventional and unflattened photon beams: A dosimetric comparison with 3D technique. Radiat Oncol 6:152; 2011. 5. Riley, C; Yang, Y; Li, T; et al. Dosimetric evaluation of the interplay effect in respiratory-gated RapidArc radiation therapy. Med Phys 41:011715; 2014. 6. Rao, M; Wu, J; Cao, D; et al. Dosimetric impact of breathing motion in lung stereotactic body radiotherapy treatment using intensity modulated radiotherapy and volumetric modulated arc therapy [corrected]. Int J Radiat Oncol Biol Phys 83:e251–6; 2012. 7. Ong, CL; Dahele, M; Slotman, BJ; et al. Dosimetric impact of the interplay effect during stereotactic lung radiation therapy delivery using flattening filter-free beams and volumetric modulated arc therapy. Int J Radiat Oncol Biol Phys 86:743–8; 2013. 8. Ceberg, S; Ceberg, C; Falk, M; et al. Evaluation of breathing interplay effects during VMAT by using 3D gel measurements. J Phys 444:012098; 2013. 9. Stambaugh, C; Nelms, BE; Dilling, T; et al. Experimentally studied dynamic dose interplay does not meaningfully affect target dose in VMAT SBRT lung treatments. Med Phys 40:091710; 2013. 10. Cai, J; Malhotra, HK; Orton, CG Point/Counterpoint. A 3D-conformal technique is better than IMRT or VMAT for lung SBRT. Med Phys 41:040601; 2014. 11. Timmerman, R; Galvin, J; Michalski, J; et al. Accreditation and quality assurance for Radiation Therapy Oncology Group: Multicenter clinical trials using Stereotactic Body Radiation Therapy in lung cancer. Acta Oncol 45:779–86; 2006. 12. Kimura, T; Nagata, Y; Eba, J; et al. A randomized Phase III trial of comparing two dose-fractionations stereotactic body radiotherapy (SBRT) for medically inoperable Stage IA non-small cell lung cancer or small lung lesions clinically diagnosed as primary lung cancer: Japan Clinical Oncology Group Study JCOG1408 (J-SBRT trial). Jpn J Clin Oncol 47(3):277–81; 2017.
JID: MDO
ARTICLE IN PRESS M. Saito, N. Sano and K. Kuriyama et al. / Medical Dosimetry xxx (xxxx) xxx
13. Nagata, Y; Hiraoka, M; Shibata, T; et al. Prospective trial of stereotactic body radiation therapy for both operable and inoperable T1N0M0 non-small cell lung cancer: Japan Clinical Oncology Group Study JCOG0403. Int J Radiat Oncol Biol Phys 93:989–96; 2015. 14. Onishi, H; Kawakami, H; Marino, K; et al. A simple respiratory indicator for irradiation during voluntary breath holding: A one-touch device without electronic materials. Radiology 255:917–23; 2010.
[mUS5Gb;May 25, 2019;11:46] 5
15. Onimaru, R; Shirato, H; Shibata, T; et al. Phase I study of stereotactic body radiation therapy for peripheral T2N0M0 non-small cell lung cancer with PTV<100 cc using a continual reassessment method (JCOG0702). Radiother Oncol 116:276–80; 2015.
JID: MDO
[mUS5Gb;May 25, 2019;11:46]
Medical Dosimetry xxx (xxxx) xxx
Medical Dosimetry journal homepage: www.meddos.org
Evaluation of the robustness of 3-dimensional conformal technique with MLC position control into the planning target volume in stereotactic body radiotherapy for lung cancer Masahide Saito, MS∗, Naoki Sano, PhD, Kengo Kuriyama, MD, PhD, Takafumi Komiyama, MD, PhD, Kan Marino, MD, Shinichi Aoki, MD, Yoshiyasu Maehata, MD, PhD, Ryo Saito, MD, Hidekazu Suzuki, RT, Yuki Shibata, RT, Koji Ueda, RT, Hiroshi Onishi, MD, PhD Department of Radiology, University of Yamanashi, Yamanashi 409-3898, Japan
a r t i c l e
i n f o
Article history: Received 25 March 2019 Accepted 24 April 2019 Available online xxx Keywords: Lung SBRT 3DCRT Robustness
a b s t r a c t The purpose of this study was to evaluate the robustness of 3-dimensional conformal technique with MLC position control into the planning target volume (PTV) in stereotactic body radiotherapy for lung cancer. Two techniques using fixed beams were compared; one technique involved setting the MLC position outside the PTV and was referred to as Plan “O.” Another technique involved setting the MLC position inside the PTV and was referred to as Plan “I.” Two tumor motions were simulated: (1) tumor motion on the internal target volume (ITV) boundary and (2) tumor motion on the PTV boundary. Ten-phase CT images that captured the tumor in respiratory motion were generated for 2 simulations. Then, 4-dimensional (4D) treatment planning was performed by using deformable image registration. The gross tumor volume (GTV) dose changes between the 4D accumulated dose and treatment planning dose were evaluated for Plan “O” and Plan “I,” respectively. For the simulation of tumor motion on the ITV boundary, the changes in GTV D50% were −0.10 ± 0.31% and −0.22 ± 0.26% (p < 0.05) for Plan “O” and Plan “I,” respectively. In the same manner, for the simulation of tumor motion on the PTV boundary, the changes in GTV D50% were −3.37 ± 2.16% and −3.68 ± 1.71% (p < 0.05). Our result suggested that the dose change would be negligible in a clinical situation where the tumor moves within the ITV margin for both techniques, while Plan “O” showed better robustness. © 2019 American Association of Medical Dosimetrists. Published by Elsevier Inc. All rights reserved.
Introduction Stereotactic body radiation therapy (SBRT) is commonly used for therapeutic management of patients with non–small cell lung cancer (NSCLC). A few earlier clinical trials that focused on SBRT to treat NSCLC have shown high local control with limited toxicity, and outcomes comparable to that of surgery.1,2 Most clinical studies have traditionally assessed the importance of 3D conformal radiotherapy (3DCRT), while intensity-modulated radiation therapy (IMRT) and volumetric-modulated arc therapy (VMAT) are still in the process of development. Although there are some reports that VMAT and IMRT usually show better dose distribution to help spare critical structures than 3DCRT,3,4 an interplay effect between respiration-induced tumor motion and multileaf colli-
∗ Reprint requests to Masahide Saito, MS, Department of Radiology, University of Yamanashi, 1110 Shimokato, Chuo-city, Yamanashi 409-3898, Japan. E-mail address: [email protected] (M. Saito).
mator (MLC) motion may affect the accuracy of the VMAT dose prescription.5-9 Furthermore, the accuracy of delivery using VMAT is affected by not only patient movement during treatment, but also MLC-positioning errors and limited accuracy of MLC modeling in the treatment planning system.10 Namely, VMAT has multiple layers of uncertainties in the treatment process when compared to 3DCRT. 3DCRT is a more simple technique than VMAT in SBRT for treatment of lung cancer, because this technique is not susceptible to the interplay effect between MLC leaves and tumor motion.10 Therefore, from this point of view, 3DCRT is more robust (=safe treatment technique) to these errors than VMAT. However, skill and experience of the clinician play an important role in the 3DCRT treatment planning. Generally, in order to improve dose coverage of the tumor, an MLC margin of several millimeters (that is “plus MLC margin”) is applied to each leaf that fits the shape of the planning target volume (PTV). Furthermore, some planners insert the MLC into the PTV (that is “minus MLC margin”) to avoid a hot spot for chest wall or to increase the GTV dose with decreasing the surrounding lung dose. Although this technique is
https://doi.org/10.1016/j.meddos.2019.04.007 0958-3947/© 2019 American Association of Medical Dosimetrists. Published by Elsevier Inc. All rights reserved.
ARTICLE IN PRESS
JID: MDO 2
[mUS5Gb;May 25, 2019;11:46]
M. Saito, N. Sano and K. Kuriyama et al. / Medical Dosimetry xxx (xxxx) xxx
Table 1 Target characteristics for all patient (n = 25) PTV margin (mm) ITV margin (mm) Tumor (GTV) volume (cc) Tumor density (g/cm3 ) PTV volume (cc) PT V/GT V ratio
All is 3.00 2.47 ± 1.47 (range: 0.00 to 6.00) 9.53 ± 5.63 (range: 1.74 to 21.50) 0.64 ± 0.22 (range: 0.27 to 1.01) 37.92 ± 18.10 (range: 11.55 to 79.30) 4.59 ± 1.50 (range: 8.19 to 2.52)
usually used in clinical situations, there have been no previous reports on whether the dose coverage of the gross tumor volume (GTV) is acceptable using “minus MLC margin.” The Radiation Therapy Oncology Group 0618 trial does not mention the need to use minus MLC margin. They simply recommend that field aperture size and shape should correspond nearly identically to the projection of the PTV along the beam’s eye view.11 On the other hand, the Japan Clinical Oncology Group (JCOG) 1408 trial permits to use the minus MLC margin,12 while it was not permitted in previous JCOG0403 trials.13 Although 3DCRT is important technique, there have been no reports for the robustness when we use the minus MLC margin. Therefore, the purpose of this study was to determine the robustness of treatment planning for SBRT of lung cancer using a technique with MLC position control into the PTV using 4D treatment planning. Materials and Methods Study design and treatment planning This is a retrospective study approved by the institutional review board. Twenty-five NSCLC patients treated with SBRT in our institute were enrolled. All patients were given breath-hold treatment with an Abches device, which is respiratory monitoring device with 2 respiratory indicators on the chest and abdomen.14 Although all treatment plans were initially created by Pinnacle3 treatment planning systems (Philips Radiation Oncology Systems, Madison, WI) with 3DCRT using 7 to 11 coplanar or noncoplanar beams, we modified all the plans in RayStation ver. 6.0 (RaySearch Laboratories, Stockholm, Sweden) to make the research progress smoothly. For all cases, we did not use a beam through the contralateral lung. All plans were generated for Elekta Synergy with Agility gantry head, which has 160 MLC leaves of 5 mm (Elekta AB, Stockholm, Sweden). The dose calculation algorithm was a collapsed cone convolution and the dose grid size was 2 mm for all dimensions. Dose constraints of organs at risk were set according to the JCOG0702 protocol.15 The patients received a dose covering 95% volume (D95) prescription of 50 Gy for PTV in 4 fractions. The internal target volume (ITV) margin was defined to use multiple CT images with breath holding. The PTV margin and the MLC margin were defined as 3 mm for all dimensions (note that the MLC was modified to the extent that it does not enter PTV region). Target characteristics for all patients are summarized in Table 1. In this study, two 3DCRT techniques were compared; one technique involved setting the MLC position outside the PTV and was referred to as Plan “O.” Another technique involved setting the MLC position inside the PTV and was referred to as Plan “I,” which is shown in Fig. 1. Each technique was applied to all beams for all patients; however, Plan “I” avoided the ITV region. Namely, for Plan “I,” the length between the MLC position and PTV edge was 1 mm for almost cases (maximum 3 mm). Tumor motion simulation and 4D treatment planning Our simulation was based on the hypothesis that the tumor moves randomly during the breath-holding treatment. Therefore, for the evaluation of robustness of the above two 3DCRT techniques, the motion of the 2 tumors was simulated by an in-house program. These types of motions included the following: (1) random motion on the ITV boundary and (2) random motion on the PTV boundary as the worst case. Figure 2 shows the typical simulation of the tumor motion. Overall, 10phase CT images that captured the tumor in random motion, were generated for 2 simulations. Each image was created using a deformable image registration (DIR) with a hybrid deformation algorithm which used image intensity and a region of interest (ROI) of the GTV referred to as the controlling ROI. These simulations were performed by RayStation and MATLAB 2016a (Mathworks, Natick, MA). Then, 4D treatment planning was performed by RayStation. First, the dose distribution was calculated for 10-phase images, respectively. Next, the dose of all phases was accumulated into the original CT image by using DIR. We used the Hybrid algorithm with an image intensity that focused on the ROI (PTV). For the dose accumulation, the weight of the dose distribution on each CT image was equal.
Fig. 1. Concept of two 3DCRT techniques. Part (a) is Plan “O” that involved setting the MLC position outside the PTV. Part (b) is Plan “I” that involved setting the MLC position inside the PTV. MLC shapes and dose distributions are shown for each plan.
Evaluation of the robustness The dose changes between the 4D accumulated dose and treatment planning dose were evaluated for Plan “O” and Plan “I,” respectively. For the dose change of the GTV, the maximum dose (Dmax ), mean dose (Dmean ), minimum dose (Dmin ) and each dose volume index (D98%, D50%, and D2%) were calculated. In the same manner, for the dose change of the PTV, maximum dose (Dmax ), mean dose (Dmean ), minimum dose (Dmin ) and each dose volume index (D95%, and D2%) were calculated. In addition, the change in conformity index was calculated for the 2 methods. Conformity index was calculated by the following equation; CI =
VRI TV
(1)
where VRI = reference isodose (50 Gy) volume and TV = target volume (PTV). All statistical analyses were performed using JMP Pro software ver. 13 (SAS Institute Inc., NC). Results Table 2 shows the absolute value of treatment planning dose and 4D accumulated dose for 2 planning techniques (Plan “O” and Plan “I”) for each dose index and conformity index. Figure 3 shows the percent change of target dose between each planning value and 4D accumulated dose. There was a significant difference in target dose changes between Plan “O” and Plan “I.” For all simulations, the PTV dose was more affected by the tumor motion than the GTV dose. For the simulation of random motion on the IT V boundary, the changes in GT V D50% were −0.10 ± 0.31% and −0.22 ± 0.26% (p < 0.05) for Plan “O” and Plan “I,” respectively. The changes in PTV D95% were −1.12 ± 0.98% and −1.48 ± 1.29% (p < 0.05) for Plan “O” and Plan “I,” respectively. In the same manner, for the simulation of random motion on the PTV boundary, the change in GTV D50% was −0.47 ± 0.43% and −0.88 ± 1.04% (p < 0.05) for Plan “O” and Plan “I,” respectively, and that of PTV D95% was −3.37 ± 2.16% and −3.68 ± 1.71% (p < 0.05) for Plan “O” and Plan “I,” respectively. Figure 4 shows a typical case (Patient 1) of the dose difference between the 4D accumulated dose and treatment planning dose for the 2 techniques. For the simulation of random motion on the ITV boundary, only small dose differences within ±1% were observed for both techniques. However, if the tumor moved onto the PTV boundary, the dose decreased 3% for both techniques.
Discussion The robustness of the SBRT to lung cancer using VMAT has already been investigated by several authors. Rao et al. investigated the impact of the intrafractional respiratory motion on the dose distribution of free-breathing lung SBRT using VMAT and fixedfield IMRT.6 They showed that the dose to the GTV was slightly decreased on average (1% of the prescription) in the 4D calculation compared with the 3D calculation. Ong et al. investigated the dosimetric impact of the interplay effect during RapidArc SBRT for lung cancers using flattening filter-free beams with different dose rates.7 Their result indicated that multiple arcs and more than 2 fractions reduced the interplay effect to a level that appeared
ARTICLE IN PRESS
JID: MDO
[mUS5Gb;May 25, 2019;11:46]
M. Saito, N. Sano and K. Kuriyama et al. / Medical Dosimetry xxx (xxxx) xxx
3
Fig. 2. The typical simulation of the tumor motion. Part (a) shows the random motion on the ITV boundary and part (b) shows random motion on the PTV boundary as the worst case. Ten-phase CT images that captured the tumor in random motion were generated for 2 simulations, respectively. Each image was created using a deformable image registration with hybrid deformation algorithm which used image intensity and a region of interest (ROI) of the GTV referred to as the controlling ROI. Table 2 The absolute value of treatment planning dose and 4D accumulated dose for 2 planning techniques (Plan “O” and Plan “I”) for each dose index. Furthermore, the conformity indexes for all doses are shown in the below of the table Absolute dose (Gy) of Plan “O”
GTV
Dmax Dmean Dmin D98% D50% D2% PTV Dmax Dmean Dmin D95% D2% Conformity index
Absolute dose (Gy) of Plan “I”
Planning dose
4D accumulated dose (Inside ITV)
4D accumulated dose (Outside ITV)
Planning dose
4D accumulated dose (Inside ITV)
70.80 ± 6.11 66.84 ± 4.35 57.74 ± 2.97 61.39 ± 2.59 67.16 ± 4.45 70.37 ± 5.90 70.83 ± 5.96 62.65 ± 2.76 44.79 ± 5.62 55.00 ± 0.00 69.83 ± 5.58 0.66 ± 0.09
70.54 ± 5.96 66.74 ± 4.32 57.57 ± 2.72 61.37 ± 2.46 67.09 ± 4.45 70.19 ± 5.83 70.60 ± 5.92 62.33 ± 2.73 43.82 ± 5.76 54.38 ± 0.54 69.66 ± 5.54 0.67 ± 0.09
70.40 ± 5.90 66.52 ± 4.25 56.63 ± 2.97 60.75 ± 2.43 66.84 ± 4.42 70.08 ± 5.77 70.40 ± 5.90 61.67 ± 2.66 41.98 ± 5.81 53.15 ± 1.19 69.53 ± 5.57 0.71 ± 0.09
73.29 ± 6.40 68.99 ± 4.18 58.15 ± 3.49 63.09 ± 2.44 69.32 ± 4.26 72.93 ± 6.26 73.29 ± 6.40 63.66 ± 2.45 43.17 ± 4.21 55.00 ± 0.02 72.21 ± 5.87 0.72 ± 0.09
73.02 ± 6.32 68.85 ± 4.13 57.73 ± 3.20 63.03 ± 2.19 69.17 ± 4.23 72.69 ± 6.16 73.02 ± 6.32 63.20 ± 2.50 42.33 ± 4.32 54.19 ± 0.70 71.90 ± 5.89 0.73 ± 0.09
unlikely to be clinically significant. For 3DCRT, although their effects may be reduced because of no MLC movement, there have been no reports on them. Furthermore, it is unclear whether Plan “O,” which involves setting the MLC position outside the PTV is a more robust technique when compared to Plan “I,” the technique that involves setting the MLC position inside the PTV. Therefore, we investigated this using the original 4D simulation for the retrospective SBRT cases in our institute. First, we created Plan “O” and Plan “I” for all patients. From the results of Table 2, the negative MLC margin technique (Plan “I”) increase the conformity and absolute dose for the GTV. Then, we simulated the random tumor motion on the ITV boundary for each patient. For the simulation, the dose change of Plan “I” was significantly larger than that of Plan “O.” It is likely that the decreased field boundary dose that derived from tumor motion was affected by the number of MLC leaves that shielded the PTV boundary for each port, indicating that Plan “O” is more robust than Plan “I.” In particular, the minimum dose mostly affected the tumor motion; this was similar to the finding in a previous study.9 It was thought that the minimum dose exists in the field boundary for most cases, which can be affected by the tumor motion. However, both changes would be negligible in clinical situations (the dose decreased to −0.69 % and −2.22% for the GTV and the PTV, respectively as shown in Fig. 3).
4D accumulated dose (Outside ITV) 72.78 ± 6.36 68.35 ± 4.25 56.40 ± 3.47 62.06 ± 2.17 68.72 ± 4.36 72.40 ± 6.23 72.78 ± 6.35 62.36 ± 2.60 40.61 ± 4.73 52.98 ± 0.93 71.69 ± 5.81 0.76 ± 0.09
We also simulated the random tumor motion on the PTV boundary for each patient to represent the worst case. The change of each dose index of this simulation was larger than that of the simulation of tumor motion on the ITV boundary for both techniques. It is probable that the simulation of tumor motion on the PTV boundary is more likely to have tumor tissue at the boundary of the MLC leaves. For this simulation, the dose decreased to −2.99 % and −6.27% for the GTV and the PTV, respectively as shown in Table 2. Therefore, our study indicated that it is acceptable to use the adjusted MLC margin on the clinical condition that the tumor moves within the ITV margin. However, it should be noted that there is a nonignorable impact to the target dose when the tumor has large motion beyond the ITV margin even using both 3DCRT techniques. Although it is common knowledge that the setting of appropriate ITV margin is essential for lung SBRT, we emphasize it again to reduce the influence. Our study had some limitations. First, a limited number of patients were enrolled in this study. Second, this study was an only a simulation using a treatment planning system, indicating that the results depended on the accuracy of a dose calculation algorithm and DIR. Finally, motion simulation was performed in only a few of the tumors. Future studies that include larger cohorts and assess other simulations such as those that use a respiration model with a baseline shift or respiration drift should be investigated.
JID: MDO 4
ARTICLE IN PRESS
[mUS5Gb;May 25, 2019;11:46]
M. Saito, N. Sano and K. Kuriyama et al. / Medical Dosimetry xxx (xxxx) xxx
Fig. 3. The percent difference between planning dose and 4D accumulated dose for each tumor simulation. The vertical axis is the percent difference, and lateral axis is each dose index. Parts (a) and (b) show the result of GTV dose index, and parts (c) and (d) show the result of PTV dose index. Parts (a) and (c) show the result of tumor simulation within ITV, and parts (b) and (d) show the result of tumor simulation without ITV. The significant change (p value < 0.05) in target dose changes between Plan “O” and Plan “I” is represented by ∗ .
References
Fig. 4. A typical case (Patient 1) of the dose difference between the 4D accumulated dose and treatment planning dose for the 2 techniques. Plan “O” (left side) and Plan “I” (right side) are shown. For the simulation of random motion on the ITV boundary (a), only small dose differences within ±1% were observed for both techniques. However, if the tumor moved onto the PTV boundary (b), the dose decreased 3% for both techniques.
Conclusions In this study, we compared the robustness of two 3DCRT techniques for lung SBRT (Plan “O” and Plan “I”). The dose change would be negligible in a clinical situation where the tumor moves within the ITV margin for both techniques, while Plan “O” showed better robustness. Our results provide useful information about 3DCRT field shape for a clinical SBRT trial of lung cancer. Conflicts of Interest There is no conflict of interest with regard to this manuscript.
1. Onishi, H; Shirato, H; Nagata, Y; et al. Stereotactic body radiotherapy (SBRT) for operable stage I non-small-cell lung cancer: Can SBRT be comparable to surgery. Int J Radiat Oncol Biol Phys 81:1352–8; 2011. 2. Nagata, Y; Takayama, K; Matsuo, Y; et al. Clinical outcomes of a phase I/II study of 48 Gy of stereotactic body radiotherapy in 4 fractions for primary lung cancer using a stereotactic body frame. Int J Radiat Oncol Biol Phys 63:1427–31; 2005. 3. Ong, CL; Verbakel, WF; Cuijpers, JP; et al. Stereotactic radiotherapy for peripheral lung tumors: A comparison of volumetric modulated arc therapy with 3 other delivery techniques. Radiother Oncol 97:437–42; 2010. 4. Zhang, GG; Ku, L; Dilling, TJ; et al. Volumetric modulated arc planning for lung stereotactic body radiotherapy using conventional and unflattened photon beams: A dosimetric comparison with 3D technique. Radiat Oncol 6:152; 2011. 5. Riley, C; Yang, Y; Li, T; et al. Dosimetric evaluation of the interplay effect in respiratory-gated RapidArc radiation therapy. Med Phys 41:011715; 2014. 6. Rao, M; Wu, J; Cao, D; et al. Dosimetric impact of breathing motion in lung stereotactic body radiotherapy treatment using intensity modulated radiotherapy and volumetric modulated arc therapy [corrected]. Int J Radiat Oncol Biol Phys 83:e251–6; 2012. 7. Ong, CL; Dahele, M; Slotman, BJ; et al. Dosimetric impact of the interplay effect during stereotactic lung radiation therapy delivery using flattening filter-free beams and volumetric modulated arc therapy. Int J Radiat Oncol Biol Phys 86:743–8; 2013. 8. Ceberg, S; Ceberg, C; Falk, M; et al. Evaluation of breathing interplay effects during VMAT by using 3D gel measurements. J Phys 444:012098; 2013. 9. Stambaugh, C; Nelms, BE; Dilling, T; et al. Experimentally studied dynamic dose interplay does not meaningfully affect target dose in VMAT SBRT lung treatments. Med Phys 40:091710; 2013. 10. Cai, J; Malhotra, HK; Orton, CG Point/Counterpoint. A 3D-conformal technique is better than IMRT or VMAT for lung SBRT. Med Phys 41:040601; 2014. 11. Timmerman, R; Galvin, J; Michalski, J; et al. Accreditation and quality assurance for Radiation Therapy Oncology Group: Multicenter clinical trials using Stereotactic Body Radiation Therapy in lung cancer. Acta Oncol 45:779–86; 2006. 12. Kimura, T; Nagata, Y; Eba, J; et al. A randomized Phase III trial of comparing two dose-fractionations stereotactic body radiotherapy (SBRT) for medically inoperable Stage IA non-small cell lung cancer or small lung lesions clinically diagnosed as primary lung cancer: Japan Clinical Oncology Group Study JCOG1408 (J-SBRT trial). Jpn J Clin Oncol 47(3):277–81; 2017.
JID: MDO
ARTICLE IN PRESS M. Saito, N. Sano and K. Kuriyama et al. / Medical Dosimetry xxx (xxxx) xxx
13. Nagata, Y; Hiraoka, M; Shibata, T; et al. Prospective trial of stereotactic body radiation therapy for both operable and inoperable T1N0M0 non-small cell lung cancer: Japan Clinical Oncology Group Study JCOG0403. Int J Radiat Oncol Biol Phys 93:989–96; 2015. 14. Onishi, H; Kawakami, H; Marino, K; et al. A simple respiratory indicator for irradiation during voluntary breath holding: A one-touch device without electronic materials. Radiology 255:917–23; 2010.
[mUS5Gb;May 25, 2019;11:46] 5
15. Onimaru, R; Shirato, H; Shibata, T; et al. Phase I study of stereotactic body radiation therapy for peripheral T2N0M0 non-small cell lung cancer with PTV<100 cc using a continual reassessment method (JCOG0702). Radiother Oncol 116:276–80; 2015.