Physica Medica xxx (2014) 1e8
Contents lists available at ScienceDirect
Physica Medica journal homepage: http://www.physicamedica.com
Original paper
The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer Seong Soon Jang a, *, Gil Ja Huh a, Suk Young Park b, Po Song Yang c, Eun Youn Cho d a
Department of Radiation Oncology, College of Medicine, The Catholic University of Korea, Seoul, South Korea Department of Internal Medicine, College of Medicine, The Catholic University of Korea, Seoul, South Korea c Department of Radiology, College of Medicine, The Catholic University of Korea, Seoul, South Korea d Department of Radiation Oncology, Daejeon St. Mary's Hospital, Daejeon, South Korea b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 7 March 2014 Received in revised form 24 April 2014 Accepted 7 May 2014 Available online xxx
The purpose of this study was to evaluate the impacts of respiratory gating and different gating windows (GWs) on lung dosimetry in stereotactic body radiotherapy (SBRT) for lung cancer. Gated SBRT plans were developed using the four-dimensional computed tomography data from 17 lung cancer patients treated with SBRT. Using amplitude-based end-exhalation gating, we established 2 fixed GWs with approximate duty cycles of 50% (50% GW) and 25% (25% GW), respectively, for this study. For highly mobile tumors (3D mobility > 10 mm), additional benefits in lung-dose reductions were achieved with the 25% GW, as a result of inadequate mobility and planning target volume reductions obtained with the 50% GW. In these tumors, the absolute differences compared to the non-gated and 50% gated plans, were 0.5 Gy and 0.33 Gy for the mean lung dose and 1.11% and 0.71% for the V20, respectively. Dosimetric benefits were achieved with the 50% GW, compared with the non-gated plan, for tumors with both low mobility and small volume (gross tumor volume 10 cc). Among the identified predictive factors of dosimetric benefits, the lateral distance from midspinal canal and the motion range in anterioreposterior direction might be stronger factors because of their correlations with many of the lung-dose parameters and greater predictive capacity. The results of the present study might facilitate the selection of appropriate patients and the optimal GW according to the tumor characteristics for gated lung SBRT. © 2014 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.
Keywords: Lung cancer Respiratory gating Stereotactic body radiotherapy
Introduction Stereotactic body radiotherapy (SBRT) for early-stage non-small cell lung cancer (NSCLC) and metastatic lung cancer has been reported to yield high local control rates in most studies [1]. This approach involves the delivery of an ablative dose to the tumor using highly conformal and hypofractionated radiation over a short time course, while limiting the doses to the surrounding normal tissues. In radiotherapy for lung cancer, large uncertainties exist regarding target delineation and localization because of respiration-induced tumor motion. These uncertainties are particularly influential in the SBRT technique, which administers high doses (biological effective doses >100 Gy) in small fractions (5
* Corresponding author. Department of Radiation Oncology, The Catholic University of Korea Daejeon St. Mary's Hospital, 64 Daeheung-ro, Jung-gu, Daejeon 301-723, South Korea. Tel.: þ82 42 220 9630; fax: þ82 42 221 9038. E-mail address: [email protected] (S.S. Jang).
fractions) to small target volumes [2]. The magnitude of respiration-induced tumor motion in the lung can exceed 2e3 cm, depending on the tumor sites and the individual patient, if this motion is not actively controlled. In fact, this motion features a patient-specific aspect, as it is difficult to estimate the range of motion before the actual measurements [3e5]. Report 62 from the International Commission on Radiation Units and Measurements (ICRU) introduced the concept of an internal target volume (ITV), which consists of the clinical target volume (CTV) plus an additional internal margin to account for tumor motion [6]. Recently, an adaptation of the four-dimensional computed tomography (4D CT) technique allowed the acquisition of 3D CT images during multiple phases of the respiratory cycle. Using this tumor and organ motion information, 4D CT can be employed for individualized ITV delineation and respiratory gated radiotherapy (RGRT) of moving lung tumors [5,7]. Together with this 4D imaging technique, various methods have been used to reduce the impact of respiratory motion during radiotherapy. These methods have included motionencompassing, respiratory gating, breath-holding, forced shallow
http://dx.doi.org/10.1016/j.ejmp.2014.05.005 1120-1797/© 2014 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
2
S.S. Jang et al. / Physica Medica xxx (2014) 1e8
breathing with abdominal compression, and beam tracking techniques [5,8,9]. Of these methods, respiratory gating involves the delivery of radiation only during a predetermined portion (gating window) of the breathing cycle. Therefore, RGRT could potentially reduce the CTV-planning target volume (PTV) margin, and thereby reduce the doses to the surrounding normal tissues and/or escalations of the tumor dose. In recent studies [10e16] of 4D CT datasets from patients with stages IeIII lung cancer, the use of a patient-specific margin (individualized ITV) or respiratory gating (gated ITV) yielded improved target coverage and significant reductions in the PTV and lungdoses, compared with population-based (conventional) margins. However, the additional dosimetric benefits from respiratory gating were modest or very limited compared with the non-gated ITV approach and, interestingly, some studies have reported variations in these additional benefits according to some tumor parameters, such as location and mobility [10,12,13]. The use of 4D CT scanning data, which constitute the gold standard for ITV delineation, is strongly recommended for SBRT planning in cases of lung cancer [17]. However, there are currently no established criteria regarding the additional use of motion-reducing methods, such as respiratory gating, in SBRT for lung cancer. Despite the highly conformal dose distributions in small target volumes exhibited by this technique, high fractional doses can result in significant toxicity to the surrounding normal tissues [18]. The long treatment time of gated SBRT results in patient discomfort and might thereby induce setup errors due to the movement of uncomfortable patients. Therefore, the use of larger gating windows (GWs) might be more patient friendly, and some patients might be indicated for large GWs or non-gating without significant increases in the lung dose. To date, reports describing the dosimetric effects of respiratory gating in lung SBRT have been limited [15,16], and the optimal portion of respiratory cycle for gating remains unknown and controversial. Controversy also remains with regard to the threshold of tumor motion or other parameters for which RGRT should be considered. The American Association of Physicists in Medicine (AAPM) has produced a report about the methods for reducing the impact of respiratory motion, and it recommends that respiratory management techniques be considered when the range of tumor motion exceeds 5 mm in any direction [8]. Normally, motion-reducing techniques, such respiratory gating, should be considered in cases with considerable tumor motion (e.g., >1 cm). The dosimetric benefits achieved with gating and a threshold of tumor motion or other parameters that could facilitate appropriate patient selection should be defined to allow for additional recommendations for respiratory gating in SBRT for lung cancer. In addition, the determination of an optimal GW according to the tumor characteristics might facilitate more efficient treatment delivery during gated SBRT. The present study was undertaken to evaluate the impacts of respiratory gating and different GWs on target volume and lung dosimetry and to identify the tumor parameters predictive of the dosimetric benefits achieved with gating in lung cancer patients treated with SBRT. Materials and methods Patient characteristics Following the approval of our institutional review board (IRB approval number: DIRB-00109_1-003), 17 patients who had been treated at our institution with SBRT via 4D CT scans for early-stage NSCLC (15 tumors) or for pulmonary metastases (2 tumors) were included in this retrospective study. The median patient age was 68 years old (range, 55e86 years), and the sample included 15 men
and 2 women. The tumors were located in the upper (n ¼ 9), middle (n ¼ 1), and lower lobes (n ¼ 7). Of the 17 tumors, 16 were peripherally located, and 1 was centrally located. 4D CT data acquisition and tumor motion analysis A 4D CT technique using a multi-slice CT scanner (SOMATOM Sensation 64; Siemens Medical Solutions, Erlangen, Germany) was performed for SBRT planning in all of the patients. The patients were immobilized with a Wing board and Vac-Lok body cushion (CIVCO Medical Solutions, Orange City, IA, USA) with the arms placed above the head. The patients were advised to breathe freely and regularly, and abdominal compression to reduce breathing motion was not applied to any patient. A single helical 4D CT scan that included the whole lung was acquired under fixed acquisition parameters (pitch of 0.1, rotation time of 0.5 s, 120 kV, and 400 mA), with a commercially available motion-monitoring system (AZ733V; Anzai Medical, Tokyo, Japan). A pressure sensor (AZ-733V), fixed to the upper abdominal region with an elastic belt, generated the external respiratory signal. A lower signal amplitude (low pressure) corresponded to the exhalation (Ex) phase of the breathing cycle, and a higher amplitude (high pressure) corresponded to the inhalation (In) phase [19]. Using the Syngo software package (Siemens Medical Solutions), the projections were sorted retrospectively according to the corresponding breathing phases (Ex and In) and the relative amplitudes at 25% intervals from 0% to 100%, and the images were reconstructed into 8 respiratory phase bins (100%, Ex 75%, Ex 50%, Ex 25%, 0%, In 25%, In 50%, and In 75%), which were equally distributed throughout the breathing cycle with slice thicknesses of 3.0 mm. Immediately following the 4D CT scan, a modified slow CT scan with the same scanning range and slice thickness was obtained with the same scanner while using the longest possible gantry rotation time (1.0 s) and a reduced pitch factor (0.5) [20]. The tumor motion amplitude was determined by measuring the tumor movement in the 8-phase 4D CT datasets with the InSpace 4D software package (Siemens Medical Solutions). The motion ranges at the tumor centroid in the superioreinferior (SI), anterioreposterior (AP), and lefteright (LR) directions were measured on the transverse, sagittal, and coronal planes with grid spacing of 1 mm for all 8 phase bins registered by this software. In our amplitude-based gating around the end of exhalation (EOE, 0% bin), we established 2 fixed GWs with approximate duty cycles of 50% and 25%, respectively, for this study. The 50% GW and 25% GW were defined as 5 phases (Ex 50%, Ex 25%, 0%, In 25%, and In 50%) and 3 phases (Ex 25%, 0%, and In 25%) around the EOE, respectively. The tumor residual motions within each GW were also determined, together with the full motion range for all 8 phases. Target volume definitions and volumetric analysis All of the CT datasets were transferred to a commercial treatment-planning system (Pinnacle3 version 8.0m; Philips Medical Systems, Fitchburg, WI, USA), and thereafter, the 4D CT and modified slow CT images were superimposed using an automated algorithm from the Syntegra® software package (Philips Medical Systems). The matched results were visually verified by reviewing the alignment of the spinal vertebrae. The gross tumor volumes (GTVs) in each of the 8 phases of the 4D CT images were delineated with the lung window setting by the same radiation oncologist and were projected onto the modified slow CT image of the same slice. The GTV size on the EOE (GTVeoe), the lateral distance from the midspinal canal to the GTVeoe centroid, and the craniocaudal distance from the carina to the GTVeoe centroid were additionally measured on the EOE images for use as tumor parameters. The craniocaudal position was expressed as a positive or negative
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
S.S. Jang et al. / Physica Medica xxx (2014) 1e8
number when the GTVeoe centroid was above or below the carina, respectively. We used the following 3 approaches to define the ITVs, based on the different GWs: (1) the composite volume of the GTVs generated from all 8 phases (non-gating with 100% duty cycle) of the 4D CT images (ITVNG); (2) the composite volume of the 5 GTVs within the 50% GW (ITVGW50); and (3) the composite volume of the 3 GTVs within the 25% GW (ITVGW25). Additionally, an isotropic internal margin of 5 mm was added to expand the GTVeoe to the ITVeoe, thus allowing for volumetric and geometric comparisons between the ITVNG and ITVeoe. The PTVs (PTVNG, PTVGW50, PTVGW25, and PTVeoe) were generated by adding 5 mm isotropic setup margins to the ITVs. Together with the absolute sizes of the ITVs and PTVs, the relative reductions (%) in ITV and PTV were calculated for each comparison (NG vs. GW50, NG vs. GW25, and GW50 vs. GW25). Dosimetric analysis The 3 PTVs (PTVNG, PTVGW50, and PTVGW25) were used to develop 3 conformal SBRT plans for all 17 tumors, based on the different GWs, to assess the dosimetric impacts on normal lung tissue. All of the plans used 10e14 coplanar and/or non-coplanar beams and were normalized such that at least 95% of the PTV received the prescription dose. The dose-fractionation schedules were 48 Gy in 4 fractions (15 tumors), 56 Gy in 4 fractions (1 tumor), and 50 Gy in 5 fractions (1 tumor). The beam energies, weights, and gantry angles remained fixed for each tumor in the same beam configuration used for the actual patient treatment, to allow for meaningful comparison. To ensure a more realistic lung volume during treatment, the dose distributions were calculated on the modified slow CT images for the PTVNG and PTVGW50 and on the EOE images for the PTVGW25; heterogeneity corrections were performed using the Collapsed Cone Convolution Superposition algorithm. The dosimetric effects of SBRT planning on normal lung tissue with the 3 different PTVs were analyzed via lung-dose parameters, such as the mean lung dose (MLD), the percentage volumes of both lungs minus the PTV receiving specific doses of 5, 10, 20, and 25 Gy (V5, V10, V20, and V25), and the absolute volume of healthy lung tissue exceeding a high dose of 40 Gy (AV40) according to doseevolume histogram estimations. Predictive factor analysis of the dosimetric benefits of respiratory gating To identify the predictive factors of the dosimetric benefits of respiratory gating and of the different GWs, the relative lung-dose reductions (%) between the plans were correlated with tumor parameters, such as the GTVeoe, GTVeoe position (lateral distance from the midspinal canal and craniocaudal distance from the carina), and motion range factors (SI, AP, and LR movement; 3D mobility; and the overlap ratio between the 2 extreme bins). Additionally, differences in the relative lung-dose reductions between the plans were also compared in these tumor parameter groups, using a cut-off value. Furthermore, the predictive capacities of the tumor parameters to detect patients who might be expected to exhibit reductions greater than the defined relative lung-dose reduction thresholds were analyzed according to the area under the curve (AUC) values of receiver operating characteristic (ROC) curves. Statistical analysis To compare absolute differences in the lung-dose parameters in each pair of plans, we used Wilcoxon's signed-rank test for each tumor. The correlations between the tumor parameters and the lung-dose reductions between the plans were evaluated with
3
Spearman's correlation analysis. In addition, differences in tumor motion, PTV reduction, and lung-dose reduction between the tumor parameter groups were compared with the ManneWhitney test. All of the statistical analyses were performed using the SPSS software package (version 15.0; SPSS Inc., Chicago, IL, USA). Values of p < 0.05 were considered significant. Results Tumor characteristics The median GTVeoe was 5.9 cc (range, 1.1e23.6 cc), and the tumors (GTVeoe centroid) were located at a median lateral distance of 7.6 cm (range, 4.5e10.0 cm) from the midspinal canal and a median distance of 2.4 cm (range, e9.8e4.5 cm) downward from the carina. The tumor motions in the SI direction were approximately 1.5e1.7 times greater than those in the AP and LR directions. In 8 patients with upper lobe tumors, motion ranges 5 mm were observed in all directions. With gating, residual motions 5 mm in any direction were observed within the 25% GW for 6 tumors and within the 50% GW for 1 tumor. Another 2 tumors exhibited residual motion >5 mm in the SI direction, despite applying the 25% GW. The mean 3D mobility for all 17 of the tumors, which was calculated as (SI2 þ AP2 þ LR2)1/2, was found to be 10.0 ± 7.9 mm, and all of the lower lobe tumors had a 3D mobility > 10 mm. The 3D mobility reduction magnitude (%) in response to gating was increased in the lower lobe tumors and with the 25% GW in comparison to the upper/middle lobe tumors and with the 50% GW, respectively (Table 1). The overlap ratios, which were defined as the ratios between the overlapping and encompassing volumes, between the 2 extreme bins for tumors in the upper/middle lobe and lower lobe were 0.65 ± 0.17 and 0.36 ± 0.23, respectively (p < 0.05). Volumetric analyses Table 2 presents the mean values of the 3 ITVs and 3 PTVs that were generated using different GWs and relative volumetric reductions between their pair. Importantly, significant differences were observed in the volumetric reductions between each pair of PTVs, according to the 3D mobility. The difference in the PTV Table 1 Motion amplitude at the tumor centroid and motion reduction with gating. Motion factors (mean ± SD) All tumors (n ¼ 17) SI (mm) SI within 50% GW SI within 25% GW AP (mm) AP within 50% GW AP within 25% GW LR (mm) LR within 50% GW LR within 25% GW 3D (mm) 3D within 50% GW 3D within 25% GW 3D reduction (%) with 50% GW (NG vs. 50% GW) 3D reduction (%) with 25% GW (NG vs. 25% GW) Additional 3D reduction (%) with 25% GW (50% GW vs. 25% GW)
7.4 5.6 3.2 4.8 3.7 2.2 4.4 3.1 1.8 10.0 7.5 4.5 19.9
± ± ± ± ± ± ± ± ± ± ± ± ±
Upper/middle lobe (n ¼ 10) vs. lower lobe (n ¼ 7) tumor
6.1 3.7 4.0 3.1 2.1 2.3 3.9 2.6 2.7 2.2 1.4 1.6 3.6 2.2 2.1 1.7 1.0 1.1 7.9 5.1 5.1 4.2 2.4 3.1 10.2 16.0
± ± ± ± ± ± ± ± ± ± ± ± ±
2.0 1.4 1.4 1.1 1.1 0.9 1.1 0.6 0.5 2.3 1.7 1.5 8.9
vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs.
12.7 ± 6.0 9.3 ± 3.5 4.6 ± 2.1 8.0 ± 4.2 5.8 ± 3.1 3.1 ± 1.4 7.6 ± 3.6 5.1 ± 1.7 2.7 ± 0.8 16.9 ± 7.9 12.3 ± 4.4 6.4 ± 2.1 25.5 ± 9.8
48.1 ± 18.6 39.6 ± 19.0 vs. 60.2 ± 9.1
p-Value
0.001 0.001 0.027 0.001 0.006 0.026 0.001 0.002 0.001 0.001 0.001 0.003 0.097 0.019
35.7 ± 18.5 28.8 ± 17.6 vs. 45.5 ± 16.0 0.079
Abbreviations: SI; AP; LR, motion ranges in the superioreinferior; anterioreposterior; lefteright directions; 3D, 3D mobility, which was calculated as (SI2 þ AP2 þ LR2)1/2; GW, gating window; NG, non-gating; SD, standard deviation.
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
4
S.S. Jang et al. / Physica Medica xxx (2014) 1e8
Table 2 Comparisons of the internal target volumes (ITVs) and planning target volumes (PTVs) obtained with different gating windows. Absolute volumes and relative reductions (mean ± SD)
NG
50% GW
25% GW
ITV (cc) (n ¼ 17) ITV reduction (%) (n ¼ 17) PTV (cc) (n ¼ 17) PTV reduction (%) (n ¼ 17)
13.6 ± 9.7
12.3 ± 9.0
10.9 ± 8.0
39.2 ± 21.9
36.3 ± 20.7
33.4 ± 19.0
3D 10 mm (n ¼ 10) >10 mm (n ¼ 7) GTVeoe 10 cc (n ¼ 10) >10 cc (n ¼ 7)
NG vs. 50% GW
NG vs. 25% GW
50% GW vs. 25% GW
10.7 ± 6.0
20.0 ± 9.1
10.5 ± 7.1
7.5 ± 4.2
14.3 ± 7.4
7.4 ± 6.1
5.6 ± 2.8
9.5 ± 5.0
4.2 ± 4.0
10.3 ± 4.5 (p ¼ 0.025)
21.1 ± 4.0 (p ¼ 0.001)
11.9 ± 6.0 (p ¼ 0.011)
8.7 ± 4.3
15.4 ± 4.9
7.3 ± 4.8
5.8 ± 3.7 (p ¼ 0.172)
12.7 ± 10.3 (p ¼ 0.380)
7.5 ± 8.1 (p ¼ 0.558)
Abbreviations: 3D, 3D mobility; GTVeoe, gross tumor volume on end of exhalation; GW, gating window; NG, non-gating; SD, standard deviation.
reduction between the groups with 10 mm and >10 mm in 3D mobility was smaller with the 50% GW, compared with the 25% GW or with the additional reduction with the 25% GW (Table 2). Additionally, the excessive volume of ITVeoe (28.5 ± 15.9 cc) for the ITVNG was 53.5 ± 9.8%, and a mean of 3.6% (range, 0.0e20.5%) of the ITVNG was not encompassed in the ITVeoe. For PTV between them, the PTVNG was 41.8 ± 11.2% smaller than the PTVeoe. Dosimetric analyses for the 3 plans, based on the different gating windows The comparisons of the lung-dose parameters for the 3 SBRT plans, based on the different GWs, are summarized in Table 3-1 and -2. Compared with the non-gated plans, the average dose reductions in the normal lung with the 50% and 25% GWs were 0.15 Gy and 0.31 Gy for the MLD and 0.35% and 0.70% for the V20, respectively. However, these dose differences between the plans were further increased in tumors with 3D mobility >10 mm, with corresponding values of 0.17 Gy and 0.50 Gy for the MLD and 0.41% and 1.11% for the V20, respectively. Interestingly, in the subgroup with 3D mobility 10 mm, the differences in the lung-dose parameters between the plans that used the PTVGW50 and PTVGW25 were not significant, in contrast to the significant differences observed in the overall tumor group or in the subgroup with mobility >10 mm. Additionally, the tumors with both 3D mobility 10 mm and GTVeoe >10 cc exhibited no significant differences regarding any of the lung-dose parameters between each pair of plans, whereas no significant differences were observed for only the parameters between the plans that used the PTVGW50 and PTVGW25 for tumors with both 3D mobility 10 mm and GTVeoe 10 cc. Predictive factors of the dosimetric benefits of respiratory gating When comparing the tumor parameters and relative reductions (%) in the lung-dose parameters, the motion-related tumor parameters (GTVeoe position and motion range factors) were found to be correlated significantly with the lung-dose reductions (or additional reductions) with the 25% GW, whereas the GTVeoe exhibited significant negative correlations with only the reductions in the V20 and V25 with the 50% GW. In particular, the lateral distance from midspinal canal and the motion range in AP direction exhibited significant correlations with many of the lung-dose parameters, such as the MLD, V10, V20, and V25. These correlations were also strongly related to the differences in the relative lungdose reductions (%) between the tumor parameter groups (Table 4). To evaluate the predictive capacities of the tumor parameters that were correlated significantly with the lung-dose reductions, we set the thresholds at 10% reductions in the V20 and
V25 with the 25% GW (between the plans that used the PTVNG and PTVGW25) for the motion-related parameters and at 7% reductions in the V20 and V25 with the 50% GW (between the plans that used the PTVNG and PTVGW50) for the GTVeoe. By these thresholds, 6 of the 17 patients (35%) in our study would be selected for gated RT with 25% or 50% GW. Except for the craniocaudal distance from the carina (AUC ¼ 0.242) and the overlap ratio between the 2 extreme bins (AUC ¼ 0.333), the motion-related parameters exhibited AUC values 0.7 (0.788 for 3D mobility, 0.818 for the lateral distance from the midspinal canal, 0.773 for SI motion, 0.848 for AP motion, and 0.818 for LR motion), whereas the AUC value for the GTVeoe was 0.743 (Fig. 1). Discussion In radiotherapy for lung cancer, patient-specific tumor motion patterns suggest the need for individualized margins that consider motion within each patient's breathing cycle, rather than the application of population-based margins, and this individual approach to target delineation is considered especially important regarding the SBRT technique [2e5,21]. Our study also indicated the importance of a patient-specific approach to target delineation. Compared with the ITVNG as the current standard ITV delineation method in SBRT planning, more than 50% of the ITVeoe was found to be excess volume, thus resulting in a large difference between the PTVs for SBRT planning. More importantly, the ITVeoe could not completely cover the ITVNG in 8 patients, which might have led to misses in tumor dose coverage. The individual determination of the GW or duty cycle is usually performed by evaluating the tumor position at adjacent phases using 4D CT data and by assessing of the residual motions (e.g., for 5 mm) within the GWs. However, in our study, 2 fixed GWs around the EOE were used to evaluate the optimal GW according to the tumor characteristics or parameters. Because the PTV determines the apertures of the beams that enter the lung, this volume is a potential main factor that could affect the adjacent normal lung under identical beam settings. With a limitation of 3D dose calculation technique, the planned dose distribution for small lung tumors can change under different respiratory conditions [22]. To ensure more realistic lung status during beam delivery according to each gating condition, we used the modified slow CT images for the PTVNG and PTVGW50 and the EOE images for the PTVGW25 with narrow GW around the EOE, respectively, as reference images for dose calculation. Our modified slow CT scan provides images with a more representative geometry for the entire respiratory cycle, as would occur during treatment. Therefore, this reference image may be proper for dose calculations of the non-gated plan and 50% gated plan with larger GW than conventional GWs. Given the relationships among the tumor mobility, PTV, and lung-dose reductions,
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
S.S. Jang et al. / Physica Medica xxx (2014) 1e8
5
Table 3 1. Absolute differences in the lung-dose parameters for the 3 SBRT plans (n ¼ 17), based on different gating windows. 2. Absolute differences in the lung-dose parameters of the subgroups, according to the 3D mobility and GTVeoe. Parameters
MLD (Gy) V5 (%) V10 (%) V20 (%) V25 (%) AV40 (cc)
PTVNG
PTVGW50
PTVGW25
PTVNG vs. PTVGW50
(Mean ± SD)
(Mean ± SD)
(Mean ± SD)
Mean (range)
p
Mean (range)
p
Mean (range)
p
0.15 0.53 0.51 0.35 0.24 2.4
0.000 0.001 0.000 0.000 0.000 0.001
0.31 1.15 0.99 0.70 0.51 6.5
0.000 0.002 0.002 0.000 0.000 0.000
0.16 0.63 0.47 0.36 0.26 4.1
0.006 0.044 0.039 0.001 0.004 0.001
4.32 19.59 13.99 6.74 4.60 48.3
± ± ± ± ± ±
1.15 5.16 3.87 2.66 1.90 19.4
Subgroups and parameters
3D mobility MLD (Gy) V5 (%) V10 (%) V20 (%) V25 (%) AV40 (cc) 3D mobility MLD (Gy) V5 (%) V10 (%) V20 (%) V25 (%) AV40 (cc) 3D mobility MLD (Gy) V5 (%) V10 (%) V20 (%) V25 (%) AV40 (cc) 3D mobility MLD (Gy) V5 (%) V10 (%) V20 (%) V25 (%) AV40 (cc)
4.17 19.06 13.48 6.39 4.36 45.9
± ± ± ± ± ±
1.13 5.03 3.75 2.59 1.86 19.7
4.01 18.43 13.00 6.04 4.09 41.8
± ± ± ± ± ±
1.02 4.41 3.31 2.32 1.66 17.9
(0.01e0.35) (0.26e1.21) (0.05e1.34) (0.01e0.77) (0.00e0.59) (0.4e7.3)
PTVNG vs. PTVGW50
PTVNG vs. PTVGW25
(0.02e1.29) (1.02e5.37) (0.86e4.77) (0.18e3.24) (0.13e2.28) (2.0e15.8)
PTVNG vs. PTVGW25
PTVGW50 vs. PTVGW25
(0.11e1.0) (1.15e4.33) (0.94e3.79) (0.14e2.46) (0.07e1.69) (0.3e11.4)
PTVGW50 vs. PTVGW25
Mean (range)
p
Mean (range)
p
Mean (range)
p
0.17 0.58 0.64 0.41 0.27 2.6
0.018 0.043 0.028 0.018 0.018 0.063
0.50 1.90 1.77 1.11 0.78 7.9
(0.13e1.29) (0.24e5.37) (0.14e4.77) (0.22e3.24) (0.20e2.28) (2.0e15.8)
0.018 0.018 0.018 0.018 0.018 0.018
0.33 1.32 1.13 0.71 0.51 5.3
(0.11e1.0) (0.36e4.33) (0.18e3.79) (0.19e2.46) (0.18e1.69) (2.1e11.4)
0.018 0.018 0.018 0.018 0.018 0.018
0.005 0.005 0.005 0.005 0.005 0.007
0.18 0.64 0.44 0.42 0.32 5.5
(0.02e0.46) (1.02e2.94) (0.86e1.32) (0.18e0.73) (0.13e0.48) (2.9e10.1)
0.007 0.047 0.047 0.005 0.005 0.005
0.04 0.14 0.01 0.11 0.09 3.2
(0.11e0.26) (1.15e1.82) (0.94e1.05) (0.14e0.52) (0.07e0.33) (0.3e9.8)
0.386 0.878 0.799 0.059 0.203 0.017
0.028 0.028 0.028 0.028 0.028 0.028
0.23 0.99 0.71 0.50 0.33 4.7
(0.12e0.46) (0.06e2.94) (0.22e1.32) (0.25e0.73) (0.13e0.48) (2.9e9.9)
0.028 0.028 0.028 0.028 0.028 0.028
0.06 0.37 0.16 0.08 0.03 1.9
(0.11e0.26) (0.59e1.82) (0.48e1.05) (0.14e0.52) (0.07e0.29) (0.3e7.2)
0.345 0.600 0.463 0.345 0.917 0.173
0.068 0.068 0.068 0.068 0.068 0.144
0.09 0.11 0.03 0.30 0.30 6.6
(0.02e0.18) (1.02e0.73) (0.86e0.57) (0.18e0.49) (0.23e0.40) (5.1e10.1)
0.144 0.715 0.715 0.068 0.068 0.068
0.01 0.20 0.20 0.15 0.19 5.2
(0.06e0.09) (1.15e0.50) (0.94e0.26) (0.00e0.32) (0.01e0.33) (0.9e9.8)
0.465 0.715 0.715 0.068 0.144 0.068
> 10 mm (n ¼ 7) (0.02e0.35) (0.26e1.21) (0.05e1.34) (0.03e0.77) (0.00e0.59) (0.4e7.3)
10 mm (n ¼ 10) 0.13 (0.01e0.30) 0.49 (0.04e1.12) 0.43 (0.02e1.14) 0.31 (0.01e0.70) 0.22 (0.01e0.51) 2.3 (0.1e5.1) 10 mm and GTVeoe 10 cc (n ¼ 6) 0.17 (0.07e0.30) 0.61 (0.11e1.12) 0.55 (0.10e1.14) 0.41 (0.21e0.70) 0.30 (0.13e0.51) 2.8 (1.0e4.7) 10 mm and GTVeoe > 10 cc (n ¼ 4) 0.07 (0.01e0.24) 0.31 (0.04e0.84) 0.24 (0.02e0.80) 0.15 (0.01e0.49) 0.11 (0.01e0.33) 1.4 (0.1e5.1)
the use of a 50% GW in tumors with high mobility might result in inadequate mobility reductions and continued large residual motion, leading to the inadequate reductions in the PTV and lung-dose parameters. In lower lobe tumors with high mobility (3D mobility > 10 mm), the average residual 3D mobility within the 50% GW was still large (0.73-fold of the full 3D mobility), in contrast to the reduced value (0.38-fold) within the 25% GW. Additionally, in response to the additional PTV reduction with the 25% GW (PTVGW50 vs. PTVGW25), these tumors exhibited a significantly increased value (11.9%), compared with the tumors (4.2%) with low mobility. Dosimetrically, these results led to significant differences in the lung-dose parameters between the plans based on the 50% and 25% GWs for tumors with high mobility but no significant differences for those with low mobility. Additionally, the motionrelated tumor parameters were correlated with the relative lungdose reductions only with the 25% GW, and the differences (approximately 1.9-fold) in the V20 and V25 between the tumor groups according to 3D mobility were significant with the 25% GW but nearly unchanged (approximately 1.1-fold) with the 50% GW. When the tumor volume factor was added to the analysis, no dosimetric benefits were observed in any of the lung-dose parameters when gated SBRT planning was used for tumors with both low mobility and large volumes (GTVeoe > 10 cc), despite the small patient number in this subgroup. However, dosimetric benefits were obtained when the 50% GW was used for tumors with both low mobility and small volumes, although no additional benefits
were obtained with the 25% GW. Regarding the initial motion analysis, although 5 mm motion in all directions without gating was observed in 8 patients, 4 of these patients obtained dosimetric benefits from the 50% GW, compared with the non-gated plan, along with an additional small (10 cc) GTVeoe. In our analyses, based on the GTVeoe (10 cc vs. >10 cc), the differences in the PTV reduction with the 50% GW (8.7 ± 4.3% vs. 5.8 ± 3.7%) were relatively greater, when compared with the reductions with the 25% GW (15.4 ± 4.9% vs. 12.7 ± 10.3%) and when comparing the 50% and 25% GWs (7.3 ± 4.8% vs. 7.5 ± 8.1%). Along with the lack of correlation between motion-related tumor factors, the GTVeoe was correlated significantly with only the relative reductions in V20 and V25 with the 50% GW. In GTVeoe-based comparisons of the tumor groups, the relative reductions in the V20 and V25 with the 50% GW were significantly greater (approximately 2.7e2.9 fold) in tumors with volumes 10 cc, but the differences between these parameters were not significant with the 25% GW. However, despite these significant differences in the lung doses according to the tumor mobility and volume, the absolute difference in magnitudes might be small for all of the dosimetric parameters. Given the dosimetric results in our study, when the 25% GW was used for tumors with high mobility, the absolute differences in the MLD relative to nongating and the 50% GW were 0.5 Gy and 0.33 Gy, respectively, and the differences in the V20 were 1.11% and 0.71%, respectively. For tumors with both low mobility and a small volume, when the 50% GW was selected, the absolute differences in the MLD and V20
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
6
S.S. Jang et al. / Physica Medica xxx (2014) 1e8
Table 4 Significant differences (p < 0.05) in the relative lung-dose reductions (%) between the tumor parameter groups. Tumor parameter groups
MLD (%, mean ± SD) GW50
GW25
GTVeoe 10 cc (n ¼ 10) >10 cc (n ¼ 7) Lateral distance from midspinal canal 8.5 cm (n ¼ 11) 5.0 ± 3.6 >8.5 cm (n ¼ 6) 10.4 ± 5.6 SI motion 7.5 mm (n ¼ 11) 5.2 ± 4.1 >7.5 mm (n ¼ 6) 10.0 ± 5.4 AP motion 5 mm (n ¼ 12) 5.1 ± 4.0 >5 mm (n ¼ 5) 11.2 ± 5.0 LR motion 5 mm (n ¼ 11) 5.2 ± 4.1 >5 mm (n ¼ 6) 10.0 ± 5.4 3D mobility 10 mm (n ¼ 10) 4.8 ± 4.0 >10 mm (n ¼ 7) 10.0 ± 5.0
V10 (%, mean ± SD) Add
GW50
GW25
V20 (%, mean ± SD) Add
GW50
GW25
V25 (%, mean ± SD) Add
7.2 ± 3.6 2.7 ± 3.1 1.1 ± 4.6 7.0 ± 5.8
2.4 ± 3.9 10.3 ± 6.3
2.1 ± 4.3 6.1 ± 4.2
1.7 ± 5.6 5.9 ± 5.2
3.1 ± 5.3 9.0 ± 6.0
4.5 ± 5.4 12.0 ± 5.7
1.2 ± 4.7 7.8 ± 5.5
2.1 ± 4.3 6.1 ± 4.2
1.7 ± 5.6 5.9 ± 5.2
1.4 ± 3.7 6.6 ± 4.1
0.7 ± 5.0 6.6 ± 5.1
GW25
Add
8.1 ± 3.4 15.7 ± 8.6
2.7 ± 4.6 10.5 ± 6.1
7.6 ± 3.6 2.6 ± 3.2
1.7 ± 3.6 6.9 ± 4.5
1.9 ± 3.6 7.5 ± 4.5
GW50
7.8 ± 5.1 16.2 ± 7.6
2.8 ± 4.6 10.8 ± 6.0
3.2 ± 5.3 9.8 ± 6.1 8.3 ± 4.1 16.6 ± 8.5
3.1 ± 5.3 9.0 ± 6.0 7.4 ± 5.1 14.3 ± 7.5
2.3 ± 4.8 9.3 ± 5.5
3.1 ± 4.6 11.3 ± 6.4 3.2 ± 5.3 9.8 ± 6.1
7.9 ± 3.9 14.9 ± 7.9
2.3 ± 4.6 10.1 ± 5.6
Abbreviations: GW50, lung-dose reduction by 50% gating window (PTVNG vs. PTVGW50); GW25, lung-dose reduction by 25% gating window (PTVNG vs. PTVGW25); Add, additional lung-dose reduction by 25% gating window (PTVGW50 vs. PTVGW25).
relative to the non-gated plan were 0.17 Gy and 0.41%, respectively. For stage III or node positive NSCLC with larger target volumes, the lung-dose reductions from gating, which have been reported in previous studies [10,12], were slightly greater than those obtained with our SBRT planning method. In the previous studies, which
used gated plans with 3e4 phases around the EOE and conventional techniques, the absolute differences in the MLD and V20 relative to the non-gated plans were 0.9e1.6 Gy and 0.6e1.9%, respectively. In a study of SBRT plans, Wu et al. [16] reported results similar to ours in 12 patients with stage I/II NSCLC and tumor
Figure 1. Receiver operating characteristic (ROC) curves used to detect patients with reductions >10% (7%) in the V20 and V25, with the 25% (50%) gating window for tumor parameters AeE (parameter F). AUC: area under the curve.
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
S.S. Jang et al. / Physica Medica xxx (2014) 1e8
motion of 1 cm in most cases. The authors found that the absolute differences in the MLD and V20 between the gated (with 30% duty cycle at EOE) and non-gated plans were 0.6 Gy and 1.2%, respectively. Recent studies have suggested that the MLD and the percentage of the total lung volume receiving a specific dose could serve as the main dosimetric predictors of symptomatic radiation pneumonitis after SBRT. For MLD 5 Gy and/or a V20 5e10%, the risk of grade 2 radiation pneumonitis was reported to be only 10e15% in most studies [23e26]. Therefore, when selecting appropriate patients and the optimal GW for gated lung SBRT, the relationships between the dosimetric benefits from gating and the clinical risk of radiation pneumonitis should also be considered. In addition, when considering several disadvantages and patient compliance regarding the implementation of respiratory gating, great care is necessary when selecting the appropriate patients to receive gated SBRT, although the potential errors associated with this technique can be minimized with the use of breath coaching and/or image-guided techniques, and the treatment times might be reduced with higher dose rates and/or volumetric modulated arc therapy (VMAT) [27,28]. The benefits from respiratory gating or predictors of benefits should be dosimetrically verified using comparisons between gated and non-gated plans. However, it would be laborious and impractical to perform multiple plans in each patient to enable such comparisons. The tumor parameters and the thresholds that predict the benefits from respiratory gating might facilitate the selection of appropriate patients for this technique in the preplanning conditions. Regarding the predictive factors, some studies of conventional techniques have suggested that the tumor location within the lung (upper vs. middle/lower lobe), 3D CTV motion, craniocaudal CTV motion and position or GTV could predict the dosimetric benefits of respiratory gating [12,13,29]. To date, no studies have reported dosimetric benefits according to the tumor volume during gated SBRT for lung cancer. However, regarding the tumor volume among these predictors, Starkschall et al. [29] reported that patients with large GTVs (>100 cm3) exhibited only small lung dose reductions under gating with breath-hold 3D CT images and conventional planning. Our results demonstrated that, together with GTVeoe as a predictor of dosimetric benefits from gating with the 50% GW, several motionrelated factors were predictive of the benefits obtained with the 25% GW. Although we used 3D mobility as a reference factor with which to differentiate low and high mobility, given its ability to include motion information in all directions, 3D tumor mobility evaluations might be time consuming. Therefore, among the identified predictive factors, some factors, such as the lateral distance from the midspinal canal and the motion range in a single direction, might be more easily assessable predictors. In particular, the lateral distance from the midspinal canal and the motion range in the AP direction might be stronger factors because of their correlations with many of the lung-dose parameters and their higher AUC values. In contrast, the craniocaudal distance from the carina and the overlap ratio between 2 extreme bins were not correlated with the lung-dose parameters or with AUC values of 10 mm), additional benefits in lung-dose reductions were achieved with the 25% GW, as a result of inadequate mobility and PTV reductions achieved with the 50% GW. Although no dosimetric benefits were achieved from respiratory gating for tumors with both low mobility and a large GTV (GTVeoe > 10 cc), benefits were achieved with the 50% GW, compared with the non-gated plan, for tumors with both low mobility and a small GTV. Among the predictive factors identified in our study, GTVeoe was correlated with dosimetric benefits when the 50% GW was used, and motion-related factors were correlated with the dosimetric benefits when the 25% GW was used. In particular, tumor parameters, such as the lateral distance from the midspinal canal and the motion range in the AP direction, might facilitate the selection of appropriate patients for gated lung SBRT, as these parameters are more easily assessable strong predictors. Conflict of interest statement The authors declare that they have no competing interests. References [1] Chi A, Liao Z, Nguyen NP, Xu J, Stea B, Komaki R. Systemic review of the patterns of failure following stereotactic body radiation therapy in early-stage non-small-cell lung cancer: clinical implications. Radiother Oncol 2010;94: 1e11. [2] Chang JY, Cox JD. Improving radiation conformality in the treatment of nonsmall cell lung cancer. Semin Radiat Oncol 2010;20:171e7. [3] Stevens CW, Munden RF, Forster KM, Kelly JF, Liao Z, Starkschall G, et al. Respiratory-driven lung tumor motion is independent of tumor size, tumor location, and pulmonary function. Int J Radiat Oncol Biol Phys 2001;51:62e8. [4] Seppenwoolde Y, Shirato H, Kitamura K, Shimizu S, van Herk M, Lebesque JV, et al. Precise and real-time measurement of 3D tumor motion in lung due to breathing and heartbeat, measured during radiotherapy. Int J Radiat Oncol Biol Phys 2002;53:822e34. [5] Loo Jr BW, Kavanagh BD, Meyer JL. Motion management and image guidance for thoracic tumor radiotherapy: clinical treatment programs. Front Radiat Ther Oncol 2011;43:271e91. [6] ICRU. Report 62: prescribing, recording, and reporting photon beam therapy (supplement to ICRU report 50). Bethesda, MD: International Commission on Radiation Units and Measurement; 1999. [7] Jiang SB. Radiotherapy of mobile tumors. Semin Radiat Oncol 2006;16: 239e48. [8] Keall PJ, Mageras GS, Balter JM, Emery RS, Forster KM, Jiang SB, et al. The management of respiratory motion in radiation oncology report of AAPM Task Group 76. Med Phys 2006;33:3874e900. [9] Bouilhol G, Ayadi M, Rit S, Thengumpallil S, Schaerer J, Vandemeulebroucke J, et al. Is abdominal compression useful in lung stereotactic body radiation therapy? A 4DCT and dosimetric lobe-dependent study. Phys Med 2013;29: 333e40. [10] Muirhead R, Featherstone C, Duffton A, Moore K, McNee S. The potential clinical benefit of respiratory gated radiotherapy (RGRT) in non-small cell lung cancer (NSCLC). Radiother Oncol 2010;95:172e7. [11] Rietzel E, Liu AK, Doppke KP, Wolfgang JA, Chen AB, Chen GT, et al. Design of 4D treatment planning target volumes. Int J Radiat Oncol Biol Phys 2006;66: 287e95. €rnsen de Koste JR, Lagerwaard FJ, Vincent A, Slotman BJ, [12] Underberg RW, van So Senan S. A dosimetric analysis of respiration-gated radiotherapy in patients with stage III lung cancer. Radiat Oncol 2006;1:8. [13] Saito T, Sakamoto T, Oya N. Predictive factors for lung dose reduction by respiratory gating at radiotherapy for lung cancer. J Radiat Res 2010;51: 691e8. [14] Underberg RW, Lagerwaard FJ, Slotman BJ, Cuijpers JP, Senan S. Benefit of respiration-gated stereotactic radiotherapy for stage I lung cancer: an analysis of 4DCT datasets. Int J Radiat Oncol Biol Phys 2005;62:554e60. [15] Zhao B, Yang Y, Li T, Li X, Heron DE, Huq MS. Image-guided respiratory-gated lung stereotactic body radiotherapy: which target definition is optimal? Med Phys 2009;36:2248e57. [16] Wu J, Li H, Shekhar R, Suntharalingam M, D'Souza W. An evaluation of planning techniques for stereotactic body radiation therapy in lung tumors. Radiother Oncol 2008;87:35e43.
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
8
S.S. Jang et al. / Physica Medica xxx (2014) 1e8
choux C, Price A, [17] De Ruysscher D, Faivre-Finn C, Nestle U, Hurkmans CW, Le Pe et al. European Organisation for Research and Treatment of Cancer recommendations for planning and delivery of high-dose, high-precision radiotherapy for lung cancer. J Clin Oncol 2010;28:5301e10. re O, Enachescu C, Sotton MP, Bouffard-Vercelli J, Jalade P, [18] Udrescu C, Rouvie et al. Potential interest of developing an integrated boost dose escalation for stereotactic irradiation of primary prostate cancer. Phys Med 2014;30:320e5. [19] Li XA, Stepaniak C, Gore E. Technical and dosimetric aspects of respiratory gating using a pressure-sensor motion monitoring system. Med Phys 2006;33:145e54. [20] Chinneck CD, McJury M, Hounsell AR. The potential for undertaking slow CT using a modern CT scanner. Br J Radiol 2010;83:687e93. [21] Liu HH, Balter P, Tutt T, Choi B, Zhang J, Wang C, et al. Assessing respirationinduced tumor motion and internal target volume using four-dimensional computed tomography for radiotherapy of lung cancer. Int J Radiat Oncol Biol Phys 2007;68:531e40. [22] Kunieda E, Deloar HM, Kishitani N, Fujisaki T, Kawase T, Seki S, et al. Variation of dose distribution of stereotactic radiotherapy for small-volume lung tumors under different respiratory conditions. Phys Med 2008;24:204e11. [23] Borst GR, Ishikawa M, Nijkamp J, Hauptmann M, Shirato H, Onimaru R, et al. Radiation pneumonitis in patients treated for malignant pulmonary lesions with hypofractionated radiation therapy. Radiother Oncol 2009;91:307e13.
[24] Matsuo Y, Shibuya K, Nakamura M, Narabayashi M, Sakanaka K, Ueki N, et al. Doseevolume metrics associated with radiation pneumonitis after stereotactic body radiation therapy for lung cancer. Int J Radiat Oncol Biol Phys 2012;83:e545e9. [25] Barriger RB, Forquer JA, Brabham JG, Andolino DL, Shapiro RH, Henderson MA, et al. A dose-volume analysis of radiation pneumonitis in non-small cell lung cancer patients treated with stereotactic body radiation therapy. Int J Radiat Oncol Biol Phys 2012;82:457e62. [26] Baker R, Han G, Sarangkasiri S, DeMarco M, Turke C, Stevens CW, et al. Clinical and dosimetric predictors of radiation pneumonitis in a large series of patients treated with stereotactic body radiation therapy to the lung. Int J Radiat Oncol Biol Phys 2013;85:190e5. [27] Mageras GS, Yorke E. Deep inspiration breath hold and respiratory gating strategies for reducing organ motion in radiation treatment. Semin Radiat Oncol 2004;14:65e75. [28] McGrath SD, Matuszak MM, Yan D, Kestin LL, Martinez AA, Grills IS. Volumetric modulated arc therapy for delivery of hypofractionated stereotactic lung radiotherapy: a dosimetric and treatment efficiency analysis. Radiother Oncol 2010;95:153e7. [29] Starkschall G, Forster KM, Kitamura K, Cardenas A, Tucker SL, Stevens CW. Correlation of gross tumor volume excursion with potential benefits of respiratory gating. Int J Radiat Oncol Biol Phys 2004;60:1291e7.
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
Contents lists available at ScienceDirect
Physica Medica journal homepage: http://www.physicamedica.com
Original paper
The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer Seong Soon Jang a, *, Gil Ja Huh a, Suk Young Park b, Po Song Yang c, Eun Youn Cho d a
Department of Radiation Oncology, College of Medicine, The Catholic University of Korea, Seoul, South Korea Department of Internal Medicine, College of Medicine, The Catholic University of Korea, Seoul, South Korea c Department of Radiology, College of Medicine, The Catholic University of Korea, Seoul, South Korea d Department of Radiation Oncology, Daejeon St. Mary's Hospital, Daejeon, South Korea b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 7 March 2014 Received in revised form 24 April 2014 Accepted 7 May 2014 Available online xxx
The purpose of this study was to evaluate the impacts of respiratory gating and different gating windows (GWs) on lung dosimetry in stereotactic body radiotherapy (SBRT) for lung cancer. Gated SBRT plans were developed using the four-dimensional computed tomography data from 17 lung cancer patients treated with SBRT. Using amplitude-based end-exhalation gating, we established 2 fixed GWs with approximate duty cycles of 50% (50% GW) and 25% (25% GW), respectively, for this study. For highly mobile tumors (3D mobility > 10 mm), additional benefits in lung-dose reductions were achieved with the 25% GW, as a result of inadequate mobility and planning target volume reductions obtained with the 50% GW. In these tumors, the absolute differences compared to the non-gated and 50% gated plans, were 0.5 Gy and 0.33 Gy for the mean lung dose and 1.11% and 0.71% for the V20, respectively. Dosimetric benefits were achieved with the 50% GW, compared with the non-gated plan, for tumors with both low mobility and small volume (gross tumor volume 10 cc). Among the identified predictive factors of dosimetric benefits, the lateral distance from midspinal canal and the motion range in anterioreposterior direction might be stronger factors because of their correlations with many of the lung-dose parameters and greater predictive capacity. The results of the present study might facilitate the selection of appropriate patients and the optimal GW according to the tumor characteristics for gated lung SBRT. © 2014 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.
Keywords: Lung cancer Respiratory gating Stereotactic body radiotherapy
Introduction Stereotactic body radiotherapy (SBRT) for early-stage non-small cell lung cancer (NSCLC) and metastatic lung cancer has been reported to yield high local control rates in most studies [1]. This approach involves the delivery of an ablative dose to the tumor using highly conformal and hypofractionated radiation over a short time course, while limiting the doses to the surrounding normal tissues. In radiotherapy for lung cancer, large uncertainties exist regarding target delineation and localization because of respiration-induced tumor motion. These uncertainties are particularly influential in the SBRT technique, which administers high doses (biological effective doses >100 Gy) in small fractions (5
* Corresponding author. Department of Radiation Oncology, The Catholic University of Korea Daejeon St. Mary's Hospital, 64 Daeheung-ro, Jung-gu, Daejeon 301-723, South Korea. Tel.: þ82 42 220 9630; fax: þ82 42 221 9038. E-mail address: [email protected] (S.S. Jang).
fractions) to small target volumes [2]. The magnitude of respiration-induced tumor motion in the lung can exceed 2e3 cm, depending on the tumor sites and the individual patient, if this motion is not actively controlled. In fact, this motion features a patient-specific aspect, as it is difficult to estimate the range of motion before the actual measurements [3e5]. Report 62 from the International Commission on Radiation Units and Measurements (ICRU) introduced the concept of an internal target volume (ITV), which consists of the clinical target volume (CTV) plus an additional internal margin to account for tumor motion [6]. Recently, an adaptation of the four-dimensional computed tomography (4D CT) technique allowed the acquisition of 3D CT images during multiple phases of the respiratory cycle. Using this tumor and organ motion information, 4D CT can be employed for individualized ITV delineation and respiratory gated radiotherapy (RGRT) of moving lung tumors [5,7]. Together with this 4D imaging technique, various methods have been used to reduce the impact of respiratory motion during radiotherapy. These methods have included motionencompassing, respiratory gating, breath-holding, forced shallow
http://dx.doi.org/10.1016/j.ejmp.2014.05.005 1120-1797/© 2014 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
2
S.S. Jang et al. / Physica Medica xxx (2014) 1e8
breathing with abdominal compression, and beam tracking techniques [5,8,9]. Of these methods, respiratory gating involves the delivery of radiation only during a predetermined portion (gating window) of the breathing cycle. Therefore, RGRT could potentially reduce the CTV-planning target volume (PTV) margin, and thereby reduce the doses to the surrounding normal tissues and/or escalations of the tumor dose. In recent studies [10e16] of 4D CT datasets from patients with stages IeIII lung cancer, the use of a patient-specific margin (individualized ITV) or respiratory gating (gated ITV) yielded improved target coverage and significant reductions in the PTV and lungdoses, compared with population-based (conventional) margins. However, the additional dosimetric benefits from respiratory gating were modest or very limited compared with the non-gated ITV approach and, interestingly, some studies have reported variations in these additional benefits according to some tumor parameters, such as location and mobility [10,12,13]. The use of 4D CT scanning data, which constitute the gold standard for ITV delineation, is strongly recommended for SBRT planning in cases of lung cancer [17]. However, there are currently no established criteria regarding the additional use of motion-reducing methods, such as respiratory gating, in SBRT for lung cancer. Despite the highly conformal dose distributions in small target volumes exhibited by this technique, high fractional doses can result in significant toxicity to the surrounding normal tissues [18]. The long treatment time of gated SBRT results in patient discomfort and might thereby induce setup errors due to the movement of uncomfortable patients. Therefore, the use of larger gating windows (GWs) might be more patient friendly, and some patients might be indicated for large GWs or non-gating without significant increases in the lung dose. To date, reports describing the dosimetric effects of respiratory gating in lung SBRT have been limited [15,16], and the optimal portion of respiratory cycle for gating remains unknown and controversial. Controversy also remains with regard to the threshold of tumor motion or other parameters for which RGRT should be considered. The American Association of Physicists in Medicine (AAPM) has produced a report about the methods for reducing the impact of respiratory motion, and it recommends that respiratory management techniques be considered when the range of tumor motion exceeds 5 mm in any direction [8]. Normally, motion-reducing techniques, such respiratory gating, should be considered in cases with considerable tumor motion (e.g., >1 cm). The dosimetric benefits achieved with gating and a threshold of tumor motion or other parameters that could facilitate appropriate patient selection should be defined to allow for additional recommendations for respiratory gating in SBRT for lung cancer. In addition, the determination of an optimal GW according to the tumor characteristics might facilitate more efficient treatment delivery during gated SBRT. The present study was undertaken to evaluate the impacts of respiratory gating and different GWs on target volume and lung dosimetry and to identify the tumor parameters predictive of the dosimetric benefits achieved with gating in lung cancer patients treated with SBRT. Materials and methods Patient characteristics Following the approval of our institutional review board (IRB approval number: DIRB-00109_1-003), 17 patients who had been treated at our institution with SBRT via 4D CT scans for early-stage NSCLC (15 tumors) or for pulmonary metastases (2 tumors) were included in this retrospective study. The median patient age was 68 years old (range, 55e86 years), and the sample included 15 men
and 2 women. The tumors were located in the upper (n ¼ 9), middle (n ¼ 1), and lower lobes (n ¼ 7). Of the 17 tumors, 16 were peripherally located, and 1 was centrally located. 4D CT data acquisition and tumor motion analysis A 4D CT technique using a multi-slice CT scanner (SOMATOM Sensation 64; Siemens Medical Solutions, Erlangen, Germany) was performed for SBRT planning in all of the patients. The patients were immobilized with a Wing board and Vac-Lok body cushion (CIVCO Medical Solutions, Orange City, IA, USA) with the arms placed above the head. The patients were advised to breathe freely and regularly, and abdominal compression to reduce breathing motion was not applied to any patient. A single helical 4D CT scan that included the whole lung was acquired under fixed acquisition parameters (pitch of 0.1, rotation time of 0.5 s, 120 kV, and 400 mA), with a commercially available motion-monitoring system (AZ733V; Anzai Medical, Tokyo, Japan). A pressure sensor (AZ-733V), fixed to the upper abdominal region with an elastic belt, generated the external respiratory signal. A lower signal amplitude (low pressure) corresponded to the exhalation (Ex) phase of the breathing cycle, and a higher amplitude (high pressure) corresponded to the inhalation (In) phase [19]. Using the Syngo software package (Siemens Medical Solutions), the projections were sorted retrospectively according to the corresponding breathing phases (Ex and In) and the relative amplitudes at 25% intervals from 0% to 100%, and the images were reconstructed into 8 respiratory phase bins (100%, Ex 75%, Ex 50%, Ex 25%, 0%, In 25%, In 50%, and In 75%), which were equally distributed throughout the breathing cycle with slice thicknesses of 3.0 mm. Immediately following the 4D CT scan, a modified slow CT scan with the same scanning range and slice thickness was obtained with the same scanner while using the longest possible gantry rotation time (1.0 s) and a reduced pitch factor (0.5) [20]. The tumor motion amplitude was determined by measuring the tumor movement in the 8-phase 4D CT datasets with the InSpace 4D software package (Siemens Medical Solutions). The motion ranges at the tumor centroid in the superioreinferior (SI), anterioreposterior (AP), and lefteright (LR) directions were measured on the transverse, sagittal, and coronal planes with grid spacing of 1 mm for all 8 phase bins registered by this software. In our amplitude-based gating around the end of exhalation (EOE, 0% bin), we established 2 fixed GWs with approximate duty cycles of 50% and 25%, respectively, for this study. The 50% GW and 25% GW were defined as 5 phases (Ex 50%, Ex 25%, 0%, In 25%, and In 50%) and 3 phases (Ex 25%, 0%, and In 25%) around the EOE, respectively. The tumor residual motions within each GW were also determined, together with the full motion range for all 8 phases. Target volume definitions and volumetric analysis All of the CT datasets were transferred to a commercial treatment-planning system (Pinnacle3 version 8.0m; Philips Medical Systems, Fitchburg, WI, USA), and thereafter, the 4D CT and modified slow CT images were superimposed using an automated algorithm from the Syntegra® software package (Philips Medical Systems). The matched results were visually verified by reviewing the alignment of the spinal vertebrae. The gross tumor volumes (GTVs) in each of the 8 phases of the 4D CT images were delineated with the lung window setting by the same radiation oncologist and were projected onto the modified slow CT image of the same slice. The GTV size on the EOE (GTVeoe), the lateral distance from the midspinal canal to the GTVeoe centroid, and the craniocaudal distance from the carina to the GTVeoe centroid were additionally measured on the EOE images for use as tumor parameters. The craniocaudal position was expressed as a positive or negative
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
S.S. Jang et al. / Physica Medica xxx (2014) 1e8
number when the GTVeoe centroid was above or below the carina, respectively. We used the following 3 approaches to define the ITVs, based on the different GWs: (1) the composite volume of the GTVs generated from all 8 phases (non-gating with 100% duty cycle) of the 4D CT images (ITVNG); (2) the composite volume of the 5 GTVs within the 50% GW (ITVGW50); and (3) the composite volume of the 3 GTVs within the 25% GW (ITVGW25). Additionally, an isotropic internal margin of 5 mm was added to expand the GTVeoe to the ITVeoe, thus allowing for volumetric and geometric comparisons between the ITVNG and ITVeoe. The PTVs (PTVNG, PTVGW50, PTVGW25, and PTVeoe) were generated by adding 5 mm isotropic setup margins to the ITVs. Together with the absolute sizes of the ITVs and PTVs, the relative reductions (%) in ITV and PTV were calculated for each comparison (NG vs. GW50, NG vs. GW25, and GW50 vs. GW25). Dosimetric analysis The 3 PTVs (PTVNG, PTVGW50, and PTVGW25) were used to develop 3 conformal SBRT plans for all 17 tumors, based on the different GWs, to assess the dosimetric impacts on normal lung tissue. All of the plans used 10e14 coplanar and/or non-coplanar beams and were normalized such that at least 95% of the PTV received the prescription dose. The dose-fractionation schedules were 48 Gy in 4 fractions (15 tumors), 56 Gy in 4 fractions (1 tumor), and 50 Gy in 5 fractions (1 tumor). The beam energies, weights, and gantry angles remained fixed for each tumor in the same beam configuration used for the actual patient treatment, to allow for meaningful comparison. To ensure a more realistic lung volume during treatment, the dose distributions were calculated on the modified slow CT images for the PTVNG and PTVGW50 and on the EOE images for the PTVGW25; heterogeneity corrections were performed using the Collapsed Cone Convolution Superposition algorithm. The dosimetric effects of SBRT planning on normal lung tissue with the 3 different PTVs were analyzed via lung-dose parameters, such as the mean lung dose (MLD), the percentage volumes of both lungs minus the PTV receiving specific doses of 5, 10, 20, and 25 Gy (V5, V10, V20, and V25), and the absolute volume of healthy lung tissue exceeding a high dose of 40 Gy (AV40) according to doseevolume histogram estimations. Predictive factor analysis of the dosimetric benefits of respiratory gating To identify the predictive factors of the dosimetric benefits of respiratory gating and of the different GWs, the relative lung-dose reductions (%) between the plans were correlated with tumor parameters, such as the GTVeoe, GTVeoe position (lateral distance from the midspinal canal and craniocaudal distance from the carina), and motion range factors (SI, AP, and LR movement; 3D mobility; and the overlap ratio between the 2 extreme bins). Additionally, differences in the relative lung-dose reductions between the plans were also compared in these tumor parameter groups, using a cut-off value. Furthermore, the predictive capacities of the tumor parameters to detect patients who might be expected to exhibit reductions greater than the defined relative lung-dose reduction thresholds were analyzed according to the area under the curve (AUC) values of receiver operating characteristic (ROC) curves. Statistical analysis To compare absolute differences in the lung-dose parameters in each pair of plans, we used Wilcoxon's signed-rank test for each tumor. The correlations between the tumor parameters and the lung-dose reductions between the plans were evaluated with
3
Spearman's correlation analysis. In addition, differences in tumor motion, PTV reduction, and lung-dose reduction between the tumor parameter groups were compared with the ManneWhitney test. All of the statistical analyses were performed using the SPSS software package (version 15.0; SPSS Inc., Chicago, IL, USA). Values of p < 0.05 were considered significant. Results Tumor characteristics The median GTVeoe was 5.9 cc (range, 1.1e23.6 cc), and the tumors (GTVeoe centroid) were located at a median lateral distance of 7.6 cm (range, 4.5e10.0 cm) from the midspinal canal and a median distance of 2.4 cm (range, e9.8e4.5 cm) downward from the carina. The tumor motions in the SI direction were approximately 1.5e1.7 times greater than those in the AP and LR directions. In 8 patients with upper lobe tumors, motion ranges 5 mm were observed in all directions. With gating, residual motions 5 mm in any direction were observed within the 25% GW for 6 tumors and within the 50% GW for 1 tumor. Another 2 tumors exhibited residual motion >5 mm in the SI direction, despite applying the 25% GW. The mean 3D mobility for all 17 of the tumors, which was calculated as (SI2 þ AP2 þ LR2)1/2, was found to be 10.0 ± 7.9 mm, and all of the lower lobe tumors had a 3D mobility > 10 mm. The 3D mobility reduction magnitude (%) in response to gating was increased in the lower lobe tumors and with the 25% GW in comparison to the upper/middle lobe tumors and with the 50% GW, respectively (Table 1). The overlap ratios, which were defined as the ratios between the overlapping and encompassing volumes, between the 2 extreme bins for tumors in the upper/middle lobe and lower lobe were 0.65 ± 0.17 and 0.36 ± 0.23, respectively (p < 0.05). Volumetric analyses Table 2 presents the mean values of the 3 ITVs and 3 PTVs that were generated using different GWs and relative volumetric reductions between their pair. Importantly, significant differences were observed in the volumetric reductions between each pair of PTVs, according to the 3D mobility. The difference in the PTV Table 1 Motion amplitude at the tumor centroid and motion reduction with gating. Motion factors (mean ± SD) All tumors (n ¼ 17) SI (mm) SI within 50% GW SI within 25% GW AP (mm) AP within 50% GW AP within 25% GW LR (mm) LR within 50% GW LR within 25% GW 3D (mm) 3D within 50% GW 3D within 25% GW 3D reduction (%) with 50% GW (NG vs. 50% GW) 3D reduction (%) with 25% GW (NG vs. 25% GW) Additional 3D reduction (%) with 25% GW (50% GW vs. 25% GW)
7.4 5.6 3.2 4.8 3.7 2.2 4.4 3.1 1.8 10.0 7.5 4.5 19.9
± ± ± ± ± ± ± ± ± ± ± ± ±
Upper/middle lobe (n ¼ 10) vs. lower lobe (n ¼ 7) tumor
6.1 3.7 4.0 3.1 2.1 2.3 3.9 2.6 2.7 2.2 1.4 1.6 3.6 2.2 2.1 1.7 1.0 1.1 7.9 5.1 5.1 4.2 2.4 3.1 10.2 16.0
± ± ± ± ± ± ± ± ± ± ± ± ±
2.0 1.4 1.4 1.1 1.1 0.9 1.1 0.6 0.5 2.3 1.7 1.5 8.9
vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs.
12.7 ± 6.0 9.3 ± 3.5 4.6 ± 2.1 8.0 ± 4.2 5.8 ± 3.1 3.1 ± 1.4 7.6 ± 3.6 5.1 ± 1.7 2.7 ± 0.8 16.9 ± 7.9 12.3 ± 4.4 6.4 ± 2.1 25.5 ± 9.8
48.1 ± 18.6 39.6 ± 19.0 vs. 60.2 ± 9.1
p-Value
0.001 0.001 0.027 0.001 0.006 0.026 0.001 0.002 0.001 0.001 0.001 0.003 0.097 0.019
35.7 ± 18.5 28.8 ± 17.6 vs. 45.5 ± 16.0 0.079
Abbreviations: SI; AP; LR, motion ranges in the superioreinferior; anterioreposterior; lefteright directions; 3D, 3D mobility, which was calculated as (SI2 þ AP2 þ LR2)1/2; GW, gating window; NG, non-gating; SD, standard deviation.
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
4
S.S. Jang et al. / Physica Medica xxx (2014) 1e8
Table 2 Comparisons of the internal target volumes (ITVs) and planning target volumes (PTVs) obtained with different gating windows. Absolute volumes and relative reductions (mean ± SD)
NG
50% GW
25% GW
ITV (cc) (n ¼ 17) ITV reduction (%) (n ¼ 17) PTV (cc) (n ¼ 17) PTV reduction (%) (n ¼ 17)
13.6 ± 9.7
12.3 ± 9.0
10.9 ± 8.0
39.2 ± 21.9
36.3 ± 20.7
33.4 ± 19.0
3D 10 mm (n ¼ 10) >10 mm (n ¼ 7) GTVeoe 10 cc (n ¼ 10) >10 cc (n ¼ 7)
NG vs. 50% GW
NG vs. 25% GW
50% GW vs. 25% GW
10.7 ± 6.0
20.0 ± 9.1
10.5 ± 7.1
7.5 ± 4.2
14.3 ± 7.4
7.4 ± 6.1
5.6 ± 2.8
9.5 ± 5.0
4.2 ± 4.0
10.3 ± 4.5 (p ¼ 0.025)
21.1 ± 4.0 (p ¼ 0.001)
11.9 ± 6.0 (p ¼ 0.011)
8.7 ± 4.3
15.4 ± 4.9
7.3 ± 4.8
5.8 ± 3.7 (p ¼ 0.172)
12.7 ± 10.3 (p ¼ 0.380)
7.5 ± 8.1 (p ¼ 0.558)
Abbreviations: 3D, 3D mobility; GTVeoe, gross tumor volume on end of exhalation; GW, gating window; NG, non-gating; SD, standard deviation.
reduction between the groups with 10 mm and >10 mm in 3D mobility was smaller with the 50% GW, compared with the 25% GW or with the additional reduction with the 25% GW (Table 2). Additionally, the excessive volume of ITVeoe (28.5 ± 15.9 cc) for the ITVNG was 53.5 ± 9.8%, and a mean of 3.6% (range, 0.0e20.5%) of the ITVNG was not encompassed in the ITVeoe. For PTV between them, the PTVNG was 41.8 ± 11.2% smaller than the PTVeoe. Dosimetric analyses for the 3 plans, based on the different gating windows The comparisons of the lung-dose parameters for the 3 SBRT plans, based on the different GWs, are summarized in Table 3-1 and -2. Compared with the non-gated plans, the average dose reductions in the normal lung with the 50% and 25% GWs were 0.15 Gy and 0.31 Gy for the MLD and 0.35% and 0.70% for the V20, respectively. However, these dose differences between the plans were further increased in tumors with 3D mobility >10 mm, with corresponding values of 0.17 Gy and 0.50 Gy for the MLD and 0.41% and 1.11% for the V20, respectively. Interestingly, in the subgroup with 3D mobility 10 mm, the differences in the lung-dose parameters between the plans that used the PTVGW50 and PTVGW25 were not significant, in contrast to the significant differences observed in the overall tumor group or in the subgroup with mobility >10 mm. Additionally, the tumors with both 3D mobility 10 mm and GTVeoe >10 cc exhibited no significant differences regarding any of the lung-dose parameters between each pair of plans, whereas no significant differences were observed for only the parameters between the plans that used the PTVGW50 and PTVGW25 for tumors with both 3D mobility 10 mm and GTVeoe 10 cc. Predictive factors of the dosimetric benefits of respiratory gating When comparing the tumor parameters and relative reductions (%) in the lung-dose parameters, the motion-related tumor parameters (GTVeoe position and motion range factors) were found to be correlated significantly with the lung-dose reductions (or additional reductions) with the 25% GW, whereas the GTVeoe exhibited significant negative correlations with only the reductions in the V20 and V25 with the 50% GW. In particular, the lateral distance from midspinal canal and the motion range in AP direction exhibited significant correlations with many of the lung-dose parameters, such as the MLD, V10, V20, and V25. These correlations were also strongly related to the differences in the relative lungdose reductions (%) between the tumor parameter groups (Table 4). To evaluate the predictive capacities of the tumor parameters that were correlated significantly with the lung-dose reductions, we set the thresholds at 10% reductions in the V20 and
V25 with the 25% GW (between the plans that used the PTVNG and PTVGW25) for the motion-related parameters and at 7% reductions in the V20 and V25 with the 50% GW (between the plans that used the PTVNG and PTVGW50) for the GTVeoe. By these thresholds, 6 of the 17 patients (35%) in our study would be selected for gated RT with 25% or 50% GW. Except for the craniocaudal distance from the carina (AUC ¼ 0.242) and the overlap ratio between the 2 extreme bins (AUC ¼ 0.333), the motion-related parameters exhibited AUC values 0.7 (0.788 for 3D mobility, 0.818 for the lateral distance from the midspinal canal, 0.773 for SI motion, 0.848 for AP motion, and 0.818 for LR motion), whereas the AUC value for the GTVeoe was 0.743 (Fig. 1). Discussion In radiotherapy for lung cancer, patient-specific tumor motion patterns suggest the need for individualized margins that consider motion within each patient's breathing cycle, rather than the application of population-based margins, and this individual approach to target delineation is considered especially important regarding the SBRT technique [2e5,21]. Our study also indicated the importance of a patient-specific approach to target delineation. Compared with the ITVNG as the current standard ITV delineation method in SBRT planning, more than 50% of the ITVeoe was found to be excess volume, thus resulting in a large difference between the PTVs for SBRT planning. More importantly, the ITVeoe could not completely cover the ITVNG in 8 patients, which might have led to misses in tumor dose coverage. The individual determination of the GW or duty cycle is usually performed by evaluating the tumor position at adjacent phases using 4D CT data and by assessing of the residual motions (e.g., for 5 mm) within the GWs. However, in our study, 2 fixed GWs around the EOE were used to evaluate the optimal GW according to the tumor characteristics or parameters. Because the PTV determines the apertures of the beams that enter the lung, this volume is a potential main factor that could affect the adjacent normal lung under identical beam settings. With a limitation of 3D dose calculation technique, the planned dose distribution for small lung tumors can change under different respiratory conditions [22]. To ensure more realistic lung status during beam delivery according to each gating condition, we used the modified slow CT images for the PTVNG and PTVGW50 and the EOE images for the PTVGW25 with narrow GW around the EOE, respectively, as reference images for dose calculation. Our modified slow CT scan provides images with a more representative geometry for the entire respiratory cycle, as would occur during treatment. Therefore, this reference image may be proper for dose calculations of the non-gated plan and 50% gated plan with larger GW than conventional GWs. Given the relationships among the tumor mobility, PTV, and lung-dose reductions,
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
S.S. Jang et al. / Physica Medica xxx (2014) 1e8
5
Table 3 1. Absolute differences in the lung-dose parameters for the 3 SBRT plans (n ¼ 17), based on different gating windows. 2. Absolute differences in the lung-dose parameters of the subgroups, according to the 3D mobility and GTVeoe. Parameters
MLD (Gy) V5 (%) V10 (%) V20 (%) V25 (%) AV40 (cc)
PTVNG
PTVGW50
PTVGW25
PTVNG vs. PTVGW50
(Mean ± SD)
(Mean ± SD)
(Mean ± SD)
Mean (range)
p
Mean (range)
p
Mean (range)
p
0.15 0.53 0.51 0.35 0.24 2.4
0.000 0.001 0.000 0.000 0.000 0.001
0.31 1.15 0.99 0.70 0.51 6.5
0.000 0.002 0.002 0.000 0.000 0.000
0.16 0.63 0.47 0.36 0.26 4.1
0.006 0.044 0.039 0.001 0.004 0.001
4.32 19.59 13.99 6.74 4.60 48.3
± ± ± ± ± ±
1.15 5.16 3.87 2.66 1.90 19.4
Subgroups and parameters
3D mobility MLD (Gy) V5 (%) V10 (%) V20 (%) V25 (%) AV40 (cc) 3D mobility MLD (Gy) V5 (%) V10 (%) V20 (%) V25 (%) AV40 (cc) 3D mobility MLD (Gy) V5 (%) V10 (%) V20 (%) V25 (%) AV40 (cc) 3D mobility MLD (Gy) V5 (%) V10 (%) V20 (%) V25 (%) AV40 (cc)
4.17 19.06 13.48 6.39 4.36 45.9
± ± ± ± ± ±
1.13 5.03 3.75 2.59 1.86 19.7
4.01 18.43 13.00 6.04 4.09 41.8
± ± ± ± ± ±
1.02 4.41 3.31 2.32 1.66 17.9
(0.01e0.35) (0.26e1.21) (0.05e1.34) (0.01e0.77) (0.00e0.59) (0.4e7.3)
PTVNG vs. PTVGW50
PTVNG vs. PTVGW25
(0.02e1.29) (1.02e5.37) (0.86e4.77) (0.18e3.24) (0.13e2.28) (2.0e15.8)
PTVNG vs. PTVGW25
PTVGW50 vs. PTVGW25
(0.11e1.0) (1.15e4.33) (0.94e3.79) (0.14e2.46) (0.07e1.69) (0.3e11.4)
PTVGW50 vs. PTVGW25
Mean (range)
p
Mean (range)
p
Mean (range)
p
0.17 0.58 0.64 0.41 0.27 2.6
0.018 0.043 0.028 0.018 0.018 0.063
0.50 1.90 1.77 1.11 0.78 7.9
(0.13e1.29) (0.24e5.37) (0.14e4.77) (0.22e3.24) (0.20e2.28) (2.0e15.8)
0.018 0.018 0.018 0.018 0.018 0.018
0.33 1.32 1.13 0.71 0.51 5.3
(0.11e1.0) (0.36e4.33) (0.18e3.79) (0.19e2.46) (0.18e1.69) (2.1e11.4)
0.018 0.018 0.018 0.018 0.018 0.018
0.005 0.005 0.005 0.005 0.005 0.007
0.18 0.64 0.44 0.42 0.32 5.5
(0.02e0.46) (1.02e2.94) (0.86e1.32) (0.18e0.73) (0.13e0.48) (2.9e10.1)
0.007 0.047 0.047 0.005 0.005 0.005
0.04 0.14 0.01 0.11 0.09 3.2
(0.11e0.26) (1.15e1.82) (0.94e1.05) (0.14e0.52) (0.07e0.33) (0.3e9.8)
0.386 0.878 0.799 0.059 0.203 0.017
0.028 0.028 0.028 0.028 0.028 0.028
0.23 0.99 0.71 0.50 0.33 4.7
(0.12e0.46) (0.06e2.94) (0.22e1.32) (0.25e0.73) (0.13e0.48) (2.9e9.9)
0.028 0.028 0.028 0.028 0.028 0.028
0.06 0.37 0.16 0.08 0.03 1.9
(0.11e0.26) (0.59e1.82) (0.48e1.05) (0.14e0.52) (0.07e0.29) (0.3e7.2)
0.345 0.600 0.463 0.345 0.917 0.173
0.068 0.068 0.068 0.068 0.068 0.144
0.09 0.11 0.03 0.30 0.30 6.6
(0.02e0.18) (1.02e0.73) (0.86e0.57) (0.18e0.49) (0.23e0.40) (5.1e10.1)
0.144 0.715 0.715 0.068 0.068 0.068
0.01 0.20 0.20 0.15 0.19 5.2
(0.06e0.09) (1.15e0.50) (0.94e0.26) (0.00e0.32) (0.01e0.33) (0.9e9.8)
0.465 0.715 0.715 0.068 0.144 0.068
> 10 mm (n ¼ 7) (0.02e0.35) (0.26e1.21) (0.05e1.34) (0.03e0.77) (0.00e0.59) (0.4e7.3)
10 mm (n ¼ 10) 0.13 (0.01e0.30) 0.49 (0.04e1.12) 0.43 (0.02e1.14) 0.31 (0.01e0.70) 0.22 (0.01e0.51) 2.3 (0.1e5.1) 10 mm and GTVeoe 10 cc (n ¼ 6) 0.17 (0.07e0.30) 0.61 (0.11e1.12) 0.55 (0.10e1.14) 0.41 (0.21e0.70) 0.30 (0.13e0.51) 2.8 (1.0e4.7) 10 mm and GTVeoe > 10 cc (n ¼ 4) 0.07 (0.01e0.24) 0.31 (0.04e0.84) 0.24 (0.02e0.80) 0.15 (0.01e0.49) 0.11 (0.01e0.33) 1.4 (0.1e5.1)
the use of a 50% GW in tumors with high mobility might result in inadequate mobility reductions and continued large residual motion, leading to the inadequate reductions in the PTV and lung-dose parameters. In lower lobe tumors with high mobility (3D mobility > 10 mm), the average residual 3D mobility within the 50% GW was still large (0.73-fold of the full 3D mobility), in contrast to the reduced value (0.38-fold) within the 25% GW. Additionally, in response to the additional PTV reduction with the 25% GW (PTVGW50 vs. PTVGW25), these tumors exhibited a significantly increased value (11.9%), compared with the tumors (4.2%) with low mobility. Dosimetrically, these results led to significant differences in the lung-dose parameters between the plans based on the 50% and 25% GWs for tumors with high mobility but no significant differences for those with low mobility. Additionally, the motionrelated tumor parameters were correlated with the relative lungdose reductions only with the 25% GW, and the differences (approximately 1.9-fold) in the V20 and V25 between the tumor groups according to 3D mobility were significant with the 25% GW but nearly unchanged (approximately 1.1-fold) with the 50% GW. When the tumor volume factor was added to the analysis, no dosimetric benefits were observed in any of the lung-dose parameters when gated SBRT planning was used for tumors with both low mobility and large volumes (GTVeoe > 10 cc), despite the small patient number in this subgroup. However, dosimetric benefits were obtained when the 50% GW was used for tumors with both low mobility and small volumes, although no additional benefits
were obtained with the 25% GW. Regarding the initial motion analysis, although 5 mm motion in all directions without gating was observed in 8 patients, 4 of these patients obtained dosimetric benefits from the 50% GW, compared with the non-gated plan, along with an additional small (10 cc) GTVeoe. In our analyses, based on the GTVeoe (10 cc vs. >10 cc), the differences in the PTV reduction with the 50% GW (8.7 ± 4.3% vs. 5.8 ± 3.7%) were relatively greater, when compared with the reductions with the 25% GW (15.4 ± 4.9% vs. 12.7 ± 10.3%) and when comparing the 50% and 25% GWs (7.3 ± 4.8% vs. 7.5 ± 8.1%). Along with the lack of correlation between motion-related tumor factors, the GTVeoe was correlated significantly with only the relative reductions in V20 and V25 with the 50% GW. In GTVeoe-based comparisons of the tumor groups, the relative reductions in the V20 and V25 with the 50% GW were significantly greater (approximately 2.7e2.9 fold) in tumors with volumes 10 cc, but the differences between these parameters were not significant with the 25% GW. However, despite these significant differences in the lung doses according to the tumor mobility and volume, the absolute difference in magnitudes might be small for all of the dosimetric parameters. Given the dosimetric results in our study, when the 25% GW was used for tumors with high mobility, the absolute differences in the MLD relative to nongating and the 50% GW were 0.5 Gy and 0.33 Gy, respectively, and the differences in the V20 were 1.11% and 0.71%, respectively. For tumors with both low mobility and a small volume, when the 50% GW was selected, the absolute differences in the MLD and V20
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
6
S.S. Jang et al. / Physica Medica xxx (2014) 1e8
Table 4 Significant differences (p < 0.05) in the relative lung-dose reductions (%) between the tumor parameter groups. Tumor parameter groups
MLD (%, mean ± SD) GW50
GW25
GTVeoe 10 cc (n ¼ 10) >10 cc (n ¼ 7) Lateral distance from midspinal canal 8.5 cm (n ¼ 11) 5.0 ± 3.6 >8.5 cm (n ¼ 6) 10.4 ± 5.6 SI motion 7.5 mm (n ¼ 11) 5.2 ± 4.1 >7.5 mm (n ¼ 6) 10.0 ± 5.4 AP motion 5 mm (n ¼ 12) 5.1 ± 4.0 >5 mm (n ¼ 5) 11.2 ± 5.0 LR motion 5 mm (n ¼ 11) 5.2 ± 4.1 >5 mm (n ¼ 6) 10.0 ± 5.4 3D mobility 10 mm (n ¼ 10) 4.8 ± 4.0 >10 mm (n ¼ 7) 10.0 ± 5.0
V10 (%, mean ± SD) Add
GW50
GW25
V20 (%, mean ± SD) Add
GW50
GW25
V25 (%, mean ± SD) Add
7.2 ± 3.6 2.7 ± 3.1 1.1 ± 4.6 7.0 ± 5.8
2.4 ± 3.9 10.3 ± 6.3
2.1 ± 4.3 6.1 ± 4.2
1.7 ± 5.6 5.9 ± 5.2
3.1 ± 5.3 9.0 ± 6.0
4.5 ± 5.4 12.0 ± 5.7
1.2 ± 4.7 7.8 ± 5.5
2.1 ± 4.3 6.1 ± 4.2
1.7 ± 5.6 5.9 ± 5.2
1.4 ± 3.7 6.6 ± 4.1
0.7 ± 5.0 6.6 ± 5.1
GW25
Add
8.1 ± 3.4 15.7 ± 8.6
2.7 ± 4.6 10.5 ± 6.1
7.6 ± 3.6 2.6 ± 3.2
1.7 ± 3.6 6.9 ± 4.5
1.9 ± 3.6 7.5 ± 4.5
GW50
7.8 ± 5.1 16.2 ± 7.6
2.8 ± 4.6 10.8 ± 6.0
3.2 ± 5.3 9.8 ± 6.1 8.3 ± 4.1 16.6 ± 8.5
3.1 ± 5.3 9.0 ± 6.0 7.4 ± 5.1 14.3 ± 7.5
2.3 ± 4.8 9.3 ± 5.5
3.1 ± 4.6 11.3 ± 6.4 3.2 ± 5.3 9.8 ± 6.1
7.9 ± 3.9 14.9 ± 7.9
2.3 ± 4.6 10.1 ± 5.6
Abbreviations: GW50, lung-dose reduction by 50% gating window (PTVNG vs. PTVGW50); GW25, lung-dose reduction by 25% gating window (PTVNG vs. PTVGW25); Add, additional lung-dose reduction by 25% gating window (PTVGW50 vs. PTVGW25).
relative to the non-gated plan were 0.17 Gy and 0.41%, respectively. For stage III or node positive NSCLC with larger target volumes, the lung-dose reductions from gating, which have been reported in previous studies [10,12], were slightly greater than those obtained with our SBRT planning method. In the previous studies, which
used gated plans with 3e4 phases around the EOE and conventional techniques, the absolute differences in the MLD and V20 relative to the non-gated plans were 0.9e1.6 Gy and 0.6e1.9%, respectively. In a study of SBRT plans, Wu et al. [16] reported results similar to ours in 12 patients with stage I/II NSCLC and tumor
Figure 1. Receiver operating characteristic (ROC) curves used to detect patients with reductions >10% (7%) in the V20 and V25, with the 25% (50%) gating window for tumor parameters AeE (parameter F). AUC: area under the curve.
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
S.S. Jang et al. / Physica Medica xxx (2014) 1e8
motion of 1 cm in most cases. The authors found that the absolute differences in the MLD and V20 between the gated (with 30% duty cycle at EOE) and non-gated plans were 0.6 Gy and 1.2%, respectively. Recent studies have suggested that the MLD and the percentage of the total lung volume receiving a specific dose could serve as the main dosimetric predictors of symptomatic radiation pneumonitis after SBRT. For MLD 5 Gy and/or a V20 5e10%, the risk of grade 2 radiation pneumonitis was reported to be only 10e15% in most studies [23e26]. Therefore, when selecting appropriate patients and the optimal GW for gated lung SBRT, the relationships between the dosimetric benefits from gating and the clinical risk of radiation pneumonitis should also be considered. In addition, when considering several disadvantages and patient compliance regarding the implementation of respiratory gating, great care is necessary when selecting the appropriate patients to receive gated SBRT, although the potential errors associated with this technique can be minimized with the use of breath coaching and/or image-guided techniques, and the treatment times might be reduced with higher dose rates and/or volumetric modulated arc therapy (VMAT) [27,28]. The benefits from respiratory gating or predictors of benefits should be dosimetrically verified using comparisons between gated and non-gated plans. However, it would be laborious and impractical to perform multiple plans in each patient to enable such comparisons. The tumor parameters and the thresholds that predict the benefits from respiratory gating might facilitate the selection of appropriate patients for this technique in the preplanning conditions. Regarding the predictive factors, some studies of conventional techniques have suggested that the tumor location within the lung (upper vs. middle/lower lobe), 3D CTV motion, craniocaudal CTV motion and position or GTV could predict the dosimetric benefits of respiratory gating [12,13,29]. To date, no studies have reported dosimetric benefits according to the tumor volume during gated SBRT for lung cancer. However, regarding the tumor volume among these predictors, Starkschall et al. [29] reported that patients with large GTVs (>100 cm3) exhibited only small lung dose reductions under gating with breath-hold 3D CT images and conventional planning. Our results demonstrated that, together with GTVeoe as a predictor of dosimetric benefits from gating with the 50% GW, several motionrelated factors were predictive of the benefits obtained with the 25% GW. Although we used 3D mobility as a reference factor with which to differentiate low and high mobility, given its ability to include motion information in all directions, 3D tumor mobility evaluations might be time consuming. Therefore, among the identified predictive factors, some factors, such as the lateral distance from the midspinal canal and the motion range in a single direction, might be more easily assessable predictors. In particular, the lateral distance from the midspinal canal and the motion range in the AP direction might be stronger factors because of their correlations with many of the lung-dose parameters and their higher AUC values. In contrast, the craniocaudal distance from the carina and the overlap ratio between 2 extreme bins were not correlated with the lung-dose parameters or with AUC values of 10 mm), additional benefits in lung-dose reductions were achieved with the 25% GW, as a result of inadequate mobility and PTV reductions achieved with the 50% GW. Although no dosimetric benefits were achieved from respiratory gating for tumors with both low mobility and a large GTV (GTVeoe > 10 cc), benefits were achieved with the 50% GW, compared with the non-gated plan, for tumors with both low mobility and a small GTV. Among the predictive factors identified in our study, GTVeoe was correlated with dosimetric benefits when the 50% GW was used, and motion-related factors were correlated with the dosimetric benefits when the 25% GW was used. In particular, tumor parameters, such as the lateral distance from the midspinal canal and the motion range in the AP direction, might facilitate the selection of appropriate patients for gated lung SBRT, as these parameters are more easily assessable strong predictors. Conflict of interest statement The authors declare that they have no competing interests. References [1] Chi A, Liao Z, Nguyen NP, Xu J, Stea B, Komaki R. Systemic review of the patterns of failure following stereotactic body radiation therapy in early-stage non-small-cell lung cancer: clinical implications. Radiother Oncol 2010;94: 1e11. [2] Chang JY, Cox JD. Improving radiation conformality in the treatment of nonsmall cell lung cancer. Semin Radiat Oncol 2010;20:171e7. [3] Stevens CW, Munden RF, Forster KM, Kelly JF, Liao Z, Starkschall G, et al. Respiratory-driven lung tumor motion is independent of tumor size, tumor location, and pulmonary function. Int J Radiat Oncol Biol Phys 2001;51:62e8. [4] Seppenwoolde Y, Shirato H, Kitamura K, Shimizu S, van Herk M, Lebesque JV, et al. Precise and real-time measurement of 3D tumor motion in lung due to breathing and heartbeat, measured during radiotherapy. Int J Radiat Oncol Biol Phys 2002;53:822e34. [5] Loo Jr BW, Kavanagh BD, Meyer JL. Motion management and image guidance for thoracic tumor radiotherapy: clinical treatment programs. Front Radiat Ther Oncol 2011;43:271e91. [6] ICRU. Report 62: prescribing, recording, and reporting photon beam therapy (supplement to ICRU report 50). Bethesda, MD: International Commission on Radiation Units and Measurement; 1999. [7] Jiang SB. Radiotherapy of mobile tumors. Semin Radiat Oncol 2006;16: 239e48. [8] Keall PJ, Mageras GS, Balter JM, Emery RS, Forster KM, Jiang SB, et al. The management of respiratory motion in radiation oncology report of AAPM Task Group 76. Med Phys 2006;33:3874e900. [9] Bouilhol G, Ayadi M, Rit S, Thengumpallil S, Schaerer J, Vandemeulebroucke J, et al. Is abdominal compression useful in lung stereotactic body radiation therapy? A 4DCT and dosimetric lobe-dependent study. Phys Med 2013;29: 333e40. [10] Muirhead R, Featherstone C, Duffton A, Moore K, McNee S. The potential clinical benefit of respiratory gated radiotherapy (RGRT) in non-small cell lung cancer (NSCLC). Radiother Oncol 2010;95:172e7. [11] Rietzel E, Liu AK, Doppke KP, Wolfgang JA, Chen AB, Chen GT, et al. Design of 4D treatment planning target volumes. Int J Radiat Oncol Biol Phys 2006;66: 287e95. €rnsen de Koste JR, Lagerwaard FJ, Vincent A, Slotman BJ, [12] Underberg RW, van So Senan S. A dosimetric analysis of respiration-gated radiotherapy in patients with stage III lung cancer. Radiat Oncol 2006;1:8. [13] Saito T, Sakamoto T, Oya N. Predictive factors for lung dose reduction by respiratory gating at radiotherapy for lung cancer. J Radiat Res 2010;51: 691e8. [14] Underberg RW, Lagerwaard FJ, Slotman BJ, Cuijpers JP, Senan S. Benefit of respiration-gated stereotactic radiotherapy for stage I lung cancer: an analysis of 4DCT datasets. Int J Radiat Oncol Biol Phys 2005;62:554e60. [15] Zhao B, Yang Y, Li T, Li X, Heron DE, Huq MS. Image-guided respiratory-gated lung stereotactic body radiotherapy: which target definition is optimal? Med Phys 2009;36:2248e57. [16] Wu J, Li H, Shekhar R, Suntharalingam M, D'Souza W. An evaluation of planning techniques for stereotactic body radiation therapy in lung tumors. Radiother Oncol 2008;87:35e43.
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
8
S.S. Jang et al. / Physica Medica xxx (2014) 1e8
choux C, Price A, [17] De Ruysscher D, Faivre-Finn C, Nestle U, Hurkmans CW, Le Pe et al. European Organisation for Research and Treatment of Cancer recommendations for planning and delivery of high-dose, high-precision radiotherapy for lung cancer. J Clin Oncol 2010;28:5301e10. re O, Enachescu C, Sotton MP, Bouffard-Vercelli J, Jalade P, [18] Udrescu C, Rouvie et al. Potential interest of developing an integrated boost dose escalation for stereotactic irradiation of primary prostate cancer. Phys Med 2014;30:320e5. [19] Li XA, Stepaniak C, Gore E. Technical and dosimetric aspects of respiratory gating using a pressure-sensor motion monitoring system. Med Phys 2006;33:145e54. [20] Chinneck CD, McJury M, Hounsell AR. The potential for undertaking slow CT using a modern CT scanner. Br J Radiol 2010;83:687e93. [21] Liu HH, Balter P, Tutt T, Choi B, Zhang J, Wang C, et al. Assessing respirationinduced tumor motion and internal target volume using four-dimensional computed tomography for radiotherapy of lung cancer. Int J Radiat Oncol Biol Phys 2007;68:531e40. [22] Kunieda E, Deloar HM, Kishitani N, Fujisaki T, Kawase T, Seki S, et al. Variation of dose distribution of stereotactic radiotherapy for small-volume lung tumors under different respiratory conditions. Phys Med 2008;24:204e11. [23] Borst GR, Ishikawa M, Nijkamp J, Hauptmann M, Shirato H, Onimaru R, et al. Radiation pneumonitis in patients treated for malignant pulmonary lesions with hypofractionated radiation therapy. Radiother Oncol 2009;91:307e13.
[24] Matsuo Y, Shibuya K, Nakamura M, Narabayashi M, Sakanaka K, Ueki N, et al. Doseevolume metrics associated with radiation pneumonitis after stereotactic body radiation therapy for lung cancer. Int J Radiat Oncol Biol Phys 2012;83:e545e9. [25] Barriger RB, Forquer JA, Brabham JG, Andolino DL, Shapiro RH, Henderson MA, et al. A dose-volume analysis of radiation pneumonitis in non-small cell lung cancer patients treated with stereotactic body radiation therapy. Int J Radiat Oncol Biol Phys 2012;82:457e62. [26] Baker R, Han G, Sarangkasiri S, DeMarco M, Turke C, Stevens CW, et al. Clinical and dosimetric predictors of radiation pneumonitis in a large series of patients treated with stereotactic body radiation therapy to the lung. Int J Radiat Oncol Biol Phys 2013;85:190e5. [27] Mageras GS, Yorke E. Deep inspiration breath hold and respiratory gating strategies for reducing organ motion in radiation treatment. Semin Radiat Oncol 2004;14:65e75. [28] McGrath SD, Matuszak MM, Yan D, Kestin LL, Martinez AA, Grills IS. Volumetric modulated arc therapy for delivery of hypofractionated stereotactic lung radiotherapy: a dosimetric and treatment efficiency analysis. Radiother Oncol 2010;95:153e7. [29] Starkschall G, Forster KM, Kitamura K, Cardenas A, Tucker SL, Stevens CW. Correlation of gross tumor volume excursion with potential benefits of respiratory gating. Int J Radiat Oncol Biol Phys 2004;60:1291e7.
Please cite this article in press as: Jang SS, et al., The impact of respiratory gating on lung dosimetry in stereotactic body radiotherapy for lung cancer, Physica Medica (2014), http://dx.doi.org/10.1016/j.ejmp.2014.05.005
