- Email: [email protected]
Verification of Planning Target Volume Settings in Volumetric Modulated Arc Therapy for Stereotactic Body Radiation Therapy by Using In-Treatment 4-Dimensional Cone Beam Computed Tomography
International Journal of
Radiation Oncology biology
physics
www.redjournal.org
Physics Contribution
Verification of Planning Target Volume Settings in Volumetric Modulated Arc Therapy for Stereotactic Body Radiation Therapy by Using In-Treatment 4-Dimensional Cone Beam Computed Tomography Wataru Takahashi, MD, Hideomi Yamashita, MD, PhD, Satoshi Kida, MS, Yoshitaka Masutani, PhD, Akira Sakumi, PhD, Kuni Ohtomo, MD, PhD, Keiichi Nakagawa, MD, PhD, and Akihiro Haga, PhD Department of Radiology, The University of Tokyo Hospital, Tokyo, Japan Received Sep 18, 2012, and in revised form Jan 26, 2013. Accepted for publication Feb 15, 2013
Summary Tumor position in actual treatment can be analyzed using in-treatment cone beam computed tomography (CBCT). This 4D version is a direct method for quantitatively assessing the intrafractional location of a moving target. The present study evaluated tumor location during beam delivery and assessed the PTV margin setting for volumetric modulated arc therapy for stereotactic body radiation therapy treatment. We found that the discrepancy between the ITVand the tumor location observed by in-treatment 4D CBCT did not exceed 5 mm in any direction in all phases.
Purpose: To evaluate setup error and tumor motion during beam delivery by using 4dimensional cone beam computed tomography (4D CBCT) and to assess the adequacy of the planning target volume (PTV) margin for lung cancer patients undergoing volumetric modulated arc therapy for stereotactic body radiation therapy (VMAT-SBRT). Methods and Materials: Fifteen lung cancer patients treated by single-arc VMAT-SBRT were selected in this analysis. All patients were treated with an abdominal compressor. The gross tumor volumes were contoured on maximum inspiration and maximum expiration CT datasets from 4D CT respiratory sorting and merged into internal target volumes (ITVs). The PTV margin was isotropically taken as 5 mm. Registration was automatically performed using “pre-3D” CBCT. Treatment was performed with a D95 prescription of 50 Gy delivered in 4 fractions. The 4D tumor locations during beam delivery were determined using in-treatment 4D CBCT images acquired in each fraction. Then, the discrepancy between the actual tumor location and the ITV was evaluated in the lateral, vertical, and longitudinal directions. Results: Overall, 55 4D CBCT sets during VMAT-SBRT were successfully obtained. The amplitude of tumor motion was less than 10 mm in all directions. The average displacements between ITV and actual tumor location during treatment were 0.41 0.93 mm, 0.15 0.58 mm, and 0.60 0.99 mm for the craniocaudal, left-right, and anteroposterior directions, respectively. The discrepancy in each phase did not exceed 5 mm in any direction. Conclusions: With in-treatment 4D CBCT, we confirmed the required PTV margins when the registration for moving target was performed using pre-3D CBCT. In-treatment 4D CBCT is a direct method for quantitatively assessing the intrafractional location of a moving target. Ó 2013 Elsevier Inc.
Reprint requests to: Akihiro Haga, PhD, Department of Radiology, The University of Tokyo Hospital, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan. E-mail: [email protected]
Int J Radiation Oncol Biol Phys, Vol. 86, No. 3, pp. 426e431, 2013 0360-3016/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ijrobp.2013.02.019
This work was partially supported by JSPS KAKENHI 24689048. Conflict of interest: Keiichi Nakagawa received research funding from Elekta K.K. Supplementary material for this article can be found at www.redjournal.org.
Volume 86 Number 3 2013
Verification of PTV using 4D CBCT
validated the planning target volume (PTV) margin used in VMATSBRT when 3D CBCT registration is used for moving targets.
Introduction For early-stage lung cancer, stereotactic body radiation therapy (SBRT) is widely accepted as the standard treatment modality (1-3). Recently, it was reported that volumetric modulated arc therapy (VMAT), which is a novel rotational technique and an extension of intensity modulated radiation therapy (IMRT), is applicable for SBRT for lung tumors (4, 5). This technique achieves treatment plan qualities comparable to the noncoplanar IMRT technique and dramatically decreases the total treatment time for each fraction. The inter- and intrafractional respiratory motions of moving targets such as lung tumors and setup errors are significant concerns even for VMAT-SBRT. Although the internal target volume (ITV) setting accounts for respiratory motion, the breathing patterns, and hence tumor motion, may change between the simulation and treatment sessions. Therefore, the tumor position must be managed similarly during both simulation and treatment. Volumetric images such as a 3-dimensional cone beam computed tomography (3D CBCT) images acquired immediately before treatment could be applicable for the accurate localization of the target. However, the actual location of the tumor during treatment may differ from that of the pre-3D CBCT images. Monitoring and recording of the patient (or target) motion during treatment remain important and challenging topics for radiation therapy. The ideal is to obtain the image volume in the state of delivered beams with gantry rotation using a technique called in-treatment CBCT. In previous studies, it was reported that CBCT-based direct verification during beam delivery is feasible (6, 7). Recently, a system for performing intreatment respiration-correlated CBCT, namely 4D CBCT, was developed by using an image-based recognition technique of the respiration phase (8). These in-treatment 4D CBCT images are most reliable for evaluating displacement during treatment. With this technique, the uncertainties of patient setup and moving targets can be clearly observed. The aim of the present study was to evaluate the location of lung tumor by using in-treatment 4D CBCT. Thus, we
Table 1
427
Methods and Materials Patients After providing written informed consent, 15 patients with lung cancer undergoing VMAT-SBRT were included in this study. Fourteen of the patients were men, and their median age was 66 years old. Patient characteristics are summarized in Table 1. The ITV ranged from 1.6 to 80.1 cc (median, 5.8 cc). All patients underwent computed tomography (CT)-based SBRT planning for IMRT. Four-dimensional CT images for treatment planning were acquired with 2-mm-thick slices using an Aquilion LB model scanner (16-slice; Toshiba). The patients were in the supine position and fitted with an abdominal compressor. The Elekta stereotactic body frame (SBF) was also used to minimize breathing artifacts for both treatment planning CT and intreatment CBCT. Scans were performed using the AZ-733V system (Anzai Medical) as an external respiratory monitoring system. Each respiratory phase scan was transferred to a Pinnacle3, version 9.0, system (Phillips). The gross tumor volume (GTV) was delineated using the lung window (window, 1600 HU; level, 300 HU), on the maximum inspiration and maximum expiration CT datasets from 4D CT respiratory sorting. Then, the ITV was produced using an integration of the GTVs as defined in these 2 phases. In all cases, the PTV was defined by adding a uniform 5-mm margin to the ITV to compensate for setup errors. The tumor moving distance in Table 1 was estimated from the center-of-mass of the GTV in these 2 phases. Patients received a D95 prescription of 50 Gy for PTV in 4 fractions. The single-arc VMAT-SBRT with 6 MV was created by SmartArc (Pinnacle3; Philips). Dose constraints for normal organs at risk for complications were the ipsilateral lung volume
Patients, tumor characteristics, and treatment information Tumor moving distance (mm)
No.
Age (y)
Sex
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
80 59 77 84 68 61 61 73 68 55 66 56 70 62 62
M M M M M M M M M M M M M F M
TNM stage T1N0M0 Metastasis T1N0M0 T2N0M0 T1N0M0 Metastasis Metastasis T2N0M0 T1N0M0 T1N0M0 T1N0M0 Metastasis T2N0M0 Metastasis T2N0M0
Tumor location Left upper lobe Right lower lobe Right lower lobe Right lower lobe Right lower lobe Left upper lobe Left lower lobe Right middle lobe Right lower lobe Left lower lobe Right upper lobe Left lower lobe Right lower lobe Left upper lobe Left upper lobe
ITV (cc)
Pathology
x
y
z
Treatment time (s)
MU
7.1 2.2 3.9 26.5 5.7 28.5 5.8 19.5 1.6 37.2 3.4 4.5 80.1 2.0 6.5
NSCLC s/o e Adeno NSCLC s/o SqCCa e e SqCCa NSCLC s/o NSCLC s/o SqCCa e SqCCa e SqCCa
0.9 0.5 0.4 1.4 1.3 1.4 1.4 2.1 1.5 1.4 1.8 0.9 0.1 0.5 0.8
1 1.5 1.8 1.2 1.1 1.6 1.2 3.9 4.4 2.5 0.5 0.9 1.4 0.9 0.5
0.9 3.3 8.4 9.3 0.9 0.9 0.4 5.6 9.0 8.9 0.8 1 1.4 0.6 0.9
285 238 265 260 310 266 268 255 254 320 353 250 352 250 335
2049 1880 2046 2052 2406 2031 2116 1989 1975 2365 2580 1965 2587 1960 2473
Abbreviations: Adeno Z adenocarcinoma; ITV Z internal target volume; MU Z motor unit; NSCLC s/o Z suspected of non-small cell lung cancer; SqCCa Z squamous cell carcinoma.
428
Takahashi et al.
receiving 20 Gy (V20) <10% and 5 Gy <25% contralateral lung volume receiving 20 Gy (V20) <0% and 5 Gy <15%; spinal cord volumes receiving 15 Gy (V15) <0%; heart volumes receiving 30 Gy <0%; liver volume receiving 30 Gy <0%; body receiving 50 Gy <0%. Dosimetric planning and plan analysis were performed in Pinnacle3. The collapsed cone convolution method in Pinnacle3 was used as the heterogeneous correction method for the lungs. All final calculations were performed with a grid size of 2.0 mm. Dose distributions were calculated using peak exhalation CT data.
Image guided RT procedure VMAT-SBRT was performed using a photon beam provided by a Synergy linear accelerator (Elekta) equipped with the kV CBCT system that included a kV x-ray tube and flat-panel detector mounted on each side of the gantry, perpendicular to the treatment beam. In advance, the isocenter information of the treatment plan, regions of interest, and CT image set for peak exhalation were sent to the workstation with application software loaded onto the x-ray volume imaging functionality PC (XVI system, version 4.2). As an image guided RT (IGRT) procedure, pre-3D CBCT images were acquired with kV imaging parameters of a beam of 120 kVp and 20 mA/20 ms at an axial field length of 20 cm with a bow-tie filter immediately before daily treatment. In this case, the typical number of frames was approximately 650 in a pre-CBCT scan. Tumor registration was performed between obtaining a planning CT image for peak exhalation and obtaining the aforementioned 3D CBCT image. In the registration procedure, the chamfer matching (bone matching) was used first, and then, the manual registration was performed using ITV and PTV. Thereafter, the patient couch was adjusted according to the registration result. For the image registration, the mean or maximum intensity projection image is more desirable as the reference image. However, the registration error caused by the use of peak exhalation CT would be negligible because the registration of tumor location was not performed by anatomical structures but instead was performed using ITV and PTV, which are independent of the respiratory phase.
Data acquisition of kV projection images during treatment To obtain volume images during treatment, kV radiography must be accompanied by a rotational treatment. This imaging was performed using an XVI system in the Motion View mode, which allows us to acquire a sequence of kV images over a period of time while the gantry can either rotate or remain static. The same parameters used for pre-3D CBCT imaging were used in the kV imaging during VMAT treatment. The typical number of frames was almost double for in-treatment 4D CBCT imaging than for pre-3D imaging because the treatment time in VMAT SBRT was 2-fold longer than the time required for pre-3D CBCT imaging.
Phase recognition and 4D CBCT reconstruction For 4D CBCT reconstruction, a respiratory signal and a binning of the respiratory phase are required. Details of the method used to acquire the respiratory signal were described previously (8). In
International Journal of Radiation Oncology Biology Physics this study, that method is briefly described as follows. After kV imaging, the respiratory signal was obtained using an imagebased phase recognition. We used a normalized crosscorrelation between adjacent kV projections for arbitrary regions based on the assumption that craniocaudal (CC) motion is dominant in frame-by-frame changes. The signal for respiration can be observed by displacement of the CC direction of the arbitrary region. We do not actually consider special structures, but we use the fact that the respiratory signal reflects the motion of the tumor or diaphragm. The original signals may have pseudo-low periodic components because of the x-ray beam traversing various thicknesses of the body and fixtures as the gantry angle changed. This component was removed by a highpass filter. The respiratory phase was then divided into 4 phases, that is, peak-inhalation, peak-exhalation, and 2 intermediate phases. In-house fast 4D CBCT reconstruction software was developed using a graphics processing unit. We used an in-house program based on the algorithms developed by Feldkamp et al (9) and Webb (10).
Evaluation The reconstructed 4D CBCT images were converted to DICOM format and exported to a Pinnacle3 treatment planning system. Planning ITV contours from the planning CT were superimposed onto the 4D CBCT images, and then, we examined the positional data of the tumors acquired from intreatment 4D CBCT studies during beam delivery. To assess setup errors and respiratory motion, the displacement of tumors between the planning ITV and the shadow of the tumor in in-treatment 4D CBCT images was measured in the CC, left-right (LR), and anterior-posterior (AP) directions (Fig. 1). This method permitted appropriate determination of the ITV-PTV margin of moving respiratory tumors. Figure 2 shows the overall procedure of treatment verification using in-treatment 4D CBCT.
Results We failed to obtain in-treatment kV projections in 5 of 60 fractions (15 patients 4 fractions). One cause for this failure was sudden beam termination, which rarely occurs in a VMAT delivery. In this case, we did not acquire the kV projections in the rest beam delivery. The other cause was that a therapist missed intreatment scan. Overall, 55 4D CBCT sets were successfully obtained during treatment for the 15 patients and analyzed with the planning CT sets. The mean time for each fraction, encompassing the treatment time for a full gantry rotation, was 284 39 seconds. Consequently, more than 1300 projection images were acquired for each fraction. These images were used in the 4D CBCT reconstruction via classification into 4 phase bins. The tumor shadow was sufficiently visible for assessing the intrafractional location in all phases of 4D CBCT. The discrepancy between the tumor location and each planning ITV was analyzed in each phase during VMAT-SBRT. The degrees of displacement were measured in the CC, LR, and AP directions. As a representative example, one of the acquired kV CBCT sets is shown in Figure 3. The distance of tumor motion during treatment is shown in Table 1. Using the SBF with abdominal pressure, the amplitude
Volume 86 Number 3 2013
Verification of PTV using 4D CBCT
429
of tumor motion caused by respiration was adequately suppressed to less than 10 mm in all directions. Table 2 presents the geometrical errors in each direction. The average displacements between ITV and actual tumor location during treatment were 0.41 0.93 mm, 0.15 0.58 mm, and 0.60 0.99 mm for the CC, LR, and AP directions, respectively. The discrepancy in each phase did not exceed 5 mm in any direction. In other words, the tumor position during VMAT was within the range of the PTV.
Discussion
Fig. 1. Example of the detection of sagittal displacement between the ITV and the actual tumor location, for which the ITV contoured on the planning CT was superimposed onto the intreatment 4D CBCT image. The displacement of the tumor location from the ITV was determined from the tumor edge.
Fig. 2.
Many previous studies revealed the fact that registration using CBCT images acquired immediately before treatment was useful for accurate setup (for SBRT see refs. 11 and 12). Furthermore, the CBCT images acquired during treatment provide the most reliable evaluation of tumor location during the actual treatment. As shown in Figure 3, an in-treatment 4D CBCT technique clearly visualized the location of the moving tumor. To the best of our knowledge, the present study is the first quantitative assessment evaluating 3D CBCT registration during VMAT-SBRT by using in-treatment 4D CBCT. This system could capture the tumor edge and provide a reasonable visualization of tumor location during treatment. Unlike pre-3D CBCT, successful tracking of tumor location using in-treatment 4D CBCT could provide the answer regardless of the delivery of appropriate irradiation. In addition, ITV and PTV settings were evaluated in VMAT-SBRT with intreatment 4D CBCT by comparing the tumor positions between the ITV (contouring based on 4D CT for treatment planning) and in-treatment 4D CBCT images simultaneously acquired during VMAT delivery. The discrepancy between the tumor edge and
Workflow of the verification of the PTV setting using in-treatment 4D CBCT.
430
International Journal of Radiation Oncology Biology Physics
Takahashi et al.
Fig. 3. Screenshots of Video E1 (available online) showing acquired in-treatment CBCT images overlaid with the planning ITV (pink curve). See supplementary Video E1 and E2 coronal/sagittal planes for a course of treatment (4 days). each planning ITV was measured, and it remained less than 5 mm in all directions. It would be essential for this result that the breathing amplitude was adequately suppressed by the SBF with an abdominal compressor. That is, the amplitudes of tumor motion in this study were less than 10 mm, and thus, the discrepancy between the actual tumor location and pre-setting for ITV was decreased compared with that of free-breathing treatment. Our result is comparable to previous data reported by Ueda et al (13), who, by using an electronic portal imaging device, demonstrated that the moving target during SBRT with a double-vacuum immobilization system was included in the PTV margin of 5 mm. Furthermore, the fact that VMAT-SBRT reduces the treatment time may permit further reductions in the PTV margin. Although the displacements were less than 5 mm in all directions, our results obtained using in-treatment 4D CBCT suggest that a relatively large difference tends to occur in the AP direction. In our additional exploration, the discrepancy observed in CBCT scans after registration was similar to that in couch shifts. Namely, the displacements in this study may be caused partly by the error of automatic registration of 3D couch shifts. Surely, the precision was relatively poor in the vertical direction. The improvement in the accuracy of couch shifts provides a potential reduction in treatment field margins in the AP direction. Daily in-treatment 4D CBCT does not prolong treatment time. However, one of the problems of using in-treatment 4D CBCT is Table 2 Displacements between the ITV and the actual tumor location during treatment for each direction Displacement (mm)
Craniocaudal Left-right Anterior-posterior
Within ITV 173 &1 21 &2 16 &3 4 &4 5 &5 1 Average (SD) 0.41 (0.93)
205 147 2 33 11 24 0 13 2 3 0 0 0.15 (0.58) 0.60 (0.99)
the additional exposure to kV x-ray irradiation. The total radiation exposure of in-treatment 4D CBCT scan was estimated to be as low as 30 mSv per day with our protocol, whereas the radiation exposure of pre-3D CBCT was approximately 15 mSv. These CBCT acquisitions delivered 4.5 cGy/fraction, which would result in an additional 18 cGy to a patient who received 50 Gy throughout the treatment period. Although in-treatment 4D CBCT imaging could provide an accurate verification for clinical treatment, it is difficult to identify baseline shifts in the tumor position, which are manifested as smaller apparent breathing motion and larger apparent tumor size. This is because the present 4D CBCT technique does not provide real-time respiratory motion but instead presents an “averaged” one. Further developments such as a “just-in-time tomography” (14) will be required to apply realistic tumor location. Other limitations of the accuracy reported in this study were that the ITV was created using only 2 phases, and the registration was performed with 3D CBCT. The motion of a tumor between maximum inspiration and expiration can be nonlinear. Therefore, the ITV in the current study might be underestimated for patients with large amplitudes of target motion. Likely, the PTV margin should be verified by as many phases as possible for in-treatment 4D CBCT. The effect caused by these limitations may be suppressed by using an abdominal compressor. Nevertheless, in our institution, improvements have been made for more accurate planning and delivery. One improvement is that the 4D CT for planning consists of 10 motion states, and GTV is delineated in each respiratory phase. These 10 GTVs are fused to form the ITV, and then, a uniform 5-mm margin is added to create the PTV. The second improvement is that similar to the IGRT, a manual 4D registration for each fraction was performed to align the ITV contours with the tumor target presented in the “pre-4D” CBCT images, and the treatment couch was shifted according to the matching result as the final target localization for VMAT delivery. A study of the registration accuracy with these improvements is in progress. In the future, the actual dose distribution will be evaluated by in-treatment CBCT using information such as the gantry angle, positional data of the multileaf collimator, and dose rate. This
Volume 86 Number 3 2013
Verification of PTV using 4D CBCT
method provides the total dose delivered to moving targets during the course of treatment. This type of review immediately after beam delivery is useful for subsequent treatment margin or dose adjustments, thus leading to adaptive radiation therapy.
5.
Conclusions
6.
With an in-treatment 4D CBCT, we confirmed that the required PTV margins were 5 mm when the registration for moving targets was performed using pre-3D CBCT. The present result may encourage the use of 3D CBCT registration even for a moving target. In addition, we conclude that in-treatment 4D CBCT is a simple and powerful tool for PTV margin determination that is easily applicable to rotational irradiation.
References
7.
8.
9. 10. 11.
1. Timmerman R, Paulus R, Galvin J, et al. Stereotactic body radiation therapy for inoperable early stage lung cancer. JAMA 2010;303: 1070-1076. 2. Fakiris AJ, McGarry RC, Yiannoutsos CT, et al. Stereotactic body radiation therapy for early-stage non-small-cell lung carcinoma: fouryear results of a prospective phase II study. Int J Radiat Oncol Biol Phys 2009;75:677-682. 3. Baumann P, Nyman J, Hoyer M, et al. Outcome in a prospective phase II trial of medically inoperable stage I non-small-cell lung cancer patients treated with stereotactic body radiotherapy. J Clin Oncol 2009;27:3290-3296. 4. McGrath SD, Matuszak MM, Yan D, et al. Volumetric modulated arc therapy for delivery of hypofractionated stereotactic lung
12.
13.
14.
431
radiotherapy: A dosimetric and treatment efficiency analysis. Radiother Oncol 2009;90:422-423. Holt A, van Vliet-Vroegindeweij C, Man A, et al. Volumetric-modulated arc therapy for stereotactic body radiotherapy of lung tumors: A comparison with intensity-modulated radiotherapy techniques. Int J Radiat Oncol Biol Phys 2011;81:1560-1567. Nakagawa K, Yamashita H, Shiraishi K, et al. Verification of intreatment tumor position using kilovoltage cone-beam computed tomography: a preliminary study. Int J Radiat Oncol Biol Phys 2007; 69:970-973. Nakagawa K, Haga A, Shiraishi K, et al. First clinical cone-beam CT imaging during volumetric modulated arc therapy. Radiother Oncol 2009;90:422-423. Kida S, Masutani Y, Yamashita H, et al. In-treatment 4D cone-beam CT reconstruction with image-based respiratory phase recognition. Radiol Phys Technol 2012;5:138-147. Feldkamp AL, Davis CL, Kress WJ. Practical cone-beam algorithm. J Opt Soc Am 1984;A16:612-619. Webb S. A modified convolution reconstruction technique for divergent beams. Phys Med Biol 1982;27:419-423. Wang Z, Wu QJ, Marks LB, et al. Cone-beam CT localization of internal target volumes for stereotactic body radiotherapy of lung lesions. Int J Radiat Oncol Biol Phys 2007;69:1618-1624. Grills IS, Hugo G, Kestin LL, et al. Image-guided radiotherapy via daily online cone-beam CT substantially reduces margin requirements for stereotactic lung radiotherapy. Int J Radiat Oncol Biol Phys 2008; 70:1045-1056. Ueda Y, Miyazaki M, Nishiyama K, et al. Craniocaudal safety margin calculation based on interfractional changes in tumor motion in lung SBRT assessed with an EPID in cine mode. Int J Radiat Oncol Biol Phys 2012;83:1064-1069. Pang G, Rowlands JA. Just-in-time tomography (JiTT): A new concept for image-guided radiation therapy. Phys. Med. Biol 2005;50: N323-N330.
Radiation Oncology biology
physics
www.redjournal.org
Physics Contribution
Verification of Planning Target Volume Settings in Volumetric Modulated Arc Therapy for Stereotactic Body Radiation Therapy by Using In-Treatment 4-Dimensional Cone Beam Computed Tomography Wataru Takahashi, MD, Hideomi Yamashita, MD, PhD, Satoshi Kida, MS, Yoshitaka Masutani, PhD, Akira Sakumi, PhD, Kuni Ohtomo, MD, PhD, Keiichi Nakagawa, MD, PhD, and Akihiro Haga, PhD Department of Radiology, The University of Tokyo Hospital, Tokyo, Japan Received Sep 18, 2012, and in revised form Jan 26, 2013. Accepted for publication Feb 15, 2013
Summary Tumor position in actual treatment can be analyzed using in-treatment cone beam computed tomography (CBCT). This 4D version is a direct method for quantitatively assessing the intrafractional location of a moving target. The present study evaluated tumor location during beam delivery and assessed the PTV margin setting for volumetric modulated arc therapy for stereotactic body radiation therapy treatment. We found that the discrepancy between the ITVand the tumor location observed by in-treatment 4D CBCT did not exceed 5 mm in any direction in all phases.
Purpose: To evaluate setup error and tumor motion during beam delivery by using 4dimensional cone beam computed tomography (4D CBCT) and to assess the adequacy of the planning target volume (PTV) margin for lung cancer patients undergoing volumetric modulated arc therapy for stereotactic body radiation therapy (VMAT-SBRT). Methods and Materials: Fifteen lung cancer patients treated by single-arc VMAT-SBRT were selected in this analysis. All patients were treated with an abdominal compressor. The gross tumor volumes were contoured on maximum inspiration and maximum expiration CT datasets from 4D CT respiratory sorting and merged into internal target volumes (ITVs). The PTV margin was isotropically taken as 5 mm. Registration was automatically performed using “pre-3D” CBCT. Treatment was performed with a D95 prescription of 50 Gy delivered in 4 fractions. The 4D tumor locations during beam delivery were determined using in-treatment 4D CBCT images acquired in each fraction. Then, the discrepancy between the actual tumor location and the ITV was evaluated in the lateral, vertical, and longitudinal directions. Results: Overall, 55 4D CBCT sets during VMAT-SBRT were successfully obtained. The amplitude of tumor motion was less than 10 mm in all directions. The average displacements between ITV and actual tumor location during treatment were 0.41 0.93 mm, 0.15 0.58 mm, and 0.60 0.99 mm for the craniocaudal, left-right, and anteroposterior directions, respectively. The discrepancy in each phase did not exceed 5 mm in any direction. Conclusions: With in-treatment 4D CBCT, we confirmed the required PTV margins when the registration for moving target was performed using pre-3D CBCT. In-treatment 4D CBCT is a direct method for quantitatively assessing the intrafractional location of a moving target. Ó 2013 Elsevier Inc.
Reprint requests to: Akihiro Haga, PhD, Department of Radiology, The University of Tokyo Hospital, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan. E-mail: [email protected]
Int J Radiation Oncol Biol Phys, Vol. 86, No. 3, pp. 426e431, 2013 0360-3016/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ijrobp.2013.02.019
This work was partially supported by JSPS KAKENHI 24689048. Conflict of interest: Keiichi Nakagawa received research funding from Elekta K.K. Supplementary material for this article can be found at www.redjournal.org.
Volume 86 Number 3 2013
Verification of PTV using 4D CBCT
validated the planning target volume (PTV) margin used in VMATSBRT when 3D CBCT registration is used for moving targets.
Introduction For early-stage lung cancer, stereotactic body radiation therapy (SBRT) is widely accepted as the standard treatment modality (1-3). Recently, it was reported that volumetric modulated arc therapy (VMAT), which is a novel rotational technique and an extension of intensity modulated radiation therapy (IMRT), is applicable for SBRT for lung tumors (4, 5). This technique achieves treatment plan qualities comparable to the noncoplanar IMRT technique and dramatically decreases the total treatment time for each fraction. The inter- and intrafractional respiratory motions of moving targets such as lung tumors and setup errors are significant concerns even for VMAT-SBRT. Although the internal target volume (ITV) setting accounts for respiratory motion, the breathing patterns, and hence tumor motion, may change between the simulation and treatment sessions. Therefore, the tumor position must be managed similarly during both simulation and treatment. Volumetric images such as a 3-dimensional cone beam computed tomography (3D CBCT) images acquired immediately before treatment could be applicable for the accurate localization of the target. However, the actual location of the tumor during treatment may differ from that of the pre-3D CBCT images. Monitoring and recording of the patient (or target) motion during treatment remain important and challenging topics for radiation therapy. The ideal is to obtain the image volume in the state of delivered beams with gantry rotation using a technique called in-treatment CBCT. In previous studies, it was reported that CBCT-based direct verification during beam delivery is feasible (6, 7). Recently, a system for performing intreatment respiration-correlated CBCT, namely 4D CBCT, was developed by using an image-based recognition technique of the respiration phase (8). These in-treatment 4D CBCT images are most reliable for evaluating displacement during treatment. With this technique, the uncertainties of patient setup and moving targets can be clearly observed. The aim of the present study was to evaluate the location of lung tumor by using in-treatment 4D CBCT. Thus, we
Table 1
427
Methods and Materials Patients After providing written informed consent, 15 patients with lung cancer undergoing VMAT-SBRT were included in this study. Fourteen of the patients were men, and their median age was 66 years old. Patient characteristics are summarized in Table 1. The ITV ranged from 1.6 to 80.1 cc (median, 5.8 cc). All patients underwent computed tomography (CT)-based SBRT planning for IMRT. Four-dimensional CT images for treatment planning were acquired with 2-mm-thick slices using an Aquilion LB model scanner (16-slice; Toshiba). The patients were in the supine position and fitted with an abdominal compressor. The Elekta stereotactic body frame (SBF) was also used to minimize breathing artifacts for both treatment planning CT and intreatment CBCT. Scans were performed using the AZ-733V system (Anzai Medical) as an external respiratory monitoring system. Each respiratory phase scan was transferred to a Pinnacle3, version 9.0, system (Phillips). The gross tumor volume (GTV) was delineated using the lung window (window, 1600 HU; level, 300 HU), on the maximum inspiration and maximum expiration CT datasets from 4D CT respiratory sorting. Then, the ITV was produced using an integration of the GTVs as defined in these 2 phases. In all cases, the PTV was defined by adding a uniform 5-mm margin to the ITV to compensate for setup errors. The tumor moving distance in Table 1 was estimated from the center-of-mass of the GTV in these 2 phases. Patients received a D95 prescription of 50 Gy for PTV in 4 fractions. The single-arc VMAT-SBRT with 6 MV was created by SmartArc (Pinnacle3; Philips). Dose constraints for normal organs at risk for complications were the ipsilateral lung volume
Patients, tumor characteristics, and treatment information Tumor moving distance (mm)
No.
Age (y)
Sex
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
80 59 77 84 68 61 61 73 68 55 66 56 70 62 62
M M M M M M M M M M M M M F M
TNM stage T1N0M0 Metastasis T1N0M0 T2N0M0 T1N0M0 Metastasis Metastasis T2N0M0 T1N0M0 T1N0M0 T1N0M0 Metastasis T2N0M0 Metastasis T2N0M0
Tumor location Left upper lobe Right lower lobe Right lower lobe Right lower lobe Right lower lobe Left upper lobe Left lower lobe Right middle lobe Right lower lobe Left lower lobe Right upper lobe Left lower lobe Right lower lobe Left upper lobe Left upper lobe
ITV (cc)
Pathology
x
y
z
Treatment time (s)
MU
7.1 2.2 3.9 26.5 5.7 28.5 5.8 19.5 1.6 37.2 3.4 4.5 80.1 2.0 6.5
NSCLC s/o e Adeno NSCLC s/o SqCCa e e SqCCa NSCLC s/o NSCLC s/o SqCCa e SqCCa e SqCCa
0.9 0.5 0.4 1.4 1.3 1.4 1.4 2.1 1.5 1.4 1.8 0.9 0.1 0.5 0.8
1 1.5 1.8 1.2 1.1 1.6 1.2 3.9 4.4 2.5 0.5 0.9 1.4 0.9 0.5
0.9 3.3 8.4 9.3 0.9 0.9 0.4 5.6 9.0 8.9 0.8 1 1.4 0.6 0.9
285 238 265 260 310 266 268 255 254 320 353 250 352 250 335
2049 1880 2046 2052 2406 2031 2116 1989 1975 2365 2580 1965 2587 1960 2473
Abbreviations: Adeno Z adenocarcinoma; ITV Z internal target volume; MU Z motor unit; NSCLC s/o Z suspected of non-small cell lung cancer; SqCCa Z squamous cell carcinoma.
428
Takahashi et al.
receiving 20 Gy (V20) <10% and 5 Gy <25% contralateral lung volume receiving 20 Gy (V20) <0% and 5 Gy <15%; spinal cord volumes receiving 15 Gy (V15) <0%; heart volumes receiving 30 Gy <0%; liver volume receiving 30 Gy <0%; body receiving 50 Gy <0%. Dosimetric planning and plan analysis were performed in Pinnacle3. The collapsed cone convolution method in Pinnacle3 was used as the heterogeneous correction method for the lungs. All final calculations were performed with a grid size of 2.0 mm. Dose distributions were calculated using peak exhalation CT data.
Image guided RT procedure VMAT-SBRT was performed using a photon beam provided by a Synergy linear accelerator (Elekta) equipped with the kV CBCT system that included a kV x-ray tube and flat-panel detector mounted on each side of the gantry, perpendicular to the treatment beam. In advance, the isocenter information of the treatment plan, regions of interest, and CT image set for peak exhalation were sent to the workstation with application software loaded onto the x-ray volume imaging functionality PC (XVI system, version 4.2). As an image guided RT (IGRT) procedure, pre-3D CBCT images were acquired with kV imaging parameters of a beam of 120 kVp and 20 mA/20 ms at an axial field length of 20 cm with a bow-tie filter immediately before daily treatment. In this case, the typical number of frames was approximately 650 in a pre-CBCT scan. Tumor registration was performed between obtaining a planning CT image for peak exhalation and obtaining the aforementioned 3D CBCT image. In the registration procedure, the chamfer matching (bone matching) was used first, and then, the manual registration was performed using ITV and PTV. Thereafter, the patient couch was adjusted according to the registration result. For the image registration, the mean or maximum intensity projection image is more desirable as the reference image. However, the registration error caused by the use of peak exhalation CT would be negligible because the registration of tumor location was not performed by anatomical structures but instead was performed using ITV and PTV, which are independent of the respiratory phase.
Data acquisition of kV projection images during treatment To obtain volume images during treatment, kV radiography must be accompanied by a rotational treatment. This imaging was performed using an XVI system in the Motion View mode, which allows us to acquire a sequence of kV images over a period of time while the gantry can either rotate or remain static. The same parameters used for pre-3D CBCT imaging were used in the kV imaging during VMAT treatment. The typical number of frames was almost double for in-treatment 4D CBCT imaging than for pre-3D imaging because the treatment time in VMAT SBRT was 2-fold longer than the time required for pre-3D CBCT imaging.
Phase recognition and 4D CBCT reconstruction For 4D CBCT reconstruction, a respiratory signal and a binning of the respiratory phase are required. Details of the method used to acquire the respiratory signal were described previously (8). In
International Journal of Radiation Oncology Biology Physics this study, that method is briefly described as follows. After kV imaging, the respiratory signal was obtained using an imagebased phase recognition. We used a normalized crosscorrelation between adjacent kV projections for arbitrary regions based on the assumption that craniocaudal (CC) motion is dominant in frame-by-frame changes. The signal for respiration can be observed by displacement of the CC direction of the arbitrary region. We do not actually consider special structures, but we use the fact that the respiratory signal reflects the motion of the tumor or diaphragm. The original signals may have pseudo-low periodic components because of the x-ray beam traversing various thicknesses of the body and fixtures as the gantry angle changed. This component was removed by a highpass filter. The respiratory phase was then divided into 4 phases, that is, peak-inhalation, peak-exhalation, and 2 intermediate phases. In-house fast 4D CBCT reconstruction software was developed using a graphics processing unit. We used an in-house program based on the algorithms developed by Feldkamp et al (9) and Webb (10).
Evaluation The reconstructed 4D CBCT images were converted to DICOM format and exported to a Pinnacle3 treatment planning system. Planning ITV contours from the planning CT were superimposed onto the 4D CBCT images, and then, we examined the positional data of the tumors acquired from intreatment 4D CBCT studies during beam delivery. To assess setup errors and respiratory motion, the displacement of tumors between the planning ITV and the shadow of the tumor in in-treatment 4D CBCT images was measured in the CC, left-right (LR), and anterior-posterior (AP) directions (Fig. 1). This method permitted appropriate determination of the ITV-PTV margin of moving respiratory tumors. Figure 2 shows the overall procedure of treatment verification using in-treatment 4D CBCT.
Results We failed to obtain in-treatment kV projections in 5 of 60 fractions (15 patients 4 fractions). One cause for this failure was sudden beam termination, which rarely occurs in a VMAT delivery. In this case, we did not acquire the kV projections in the rest beam delivery. The other cause was that a therapist missed intreatment scan. Overall, 55 4D CBCT sets were successfully obtained during treatment for the 15 patients and analyzed with the planning CT sets. The mean time for each fraction, encompassing the treatment time for a full gantry rotation, was 284 39 seconds. Consequently, more than 1300 projection images were acquired for each fraction. These images were used in the 4D CBCT reconstruction via classification into 4 phase bins. The tumor shadow was sufficiently visible for assessing the intrafractional location in all phases of 4D CBCT. The discrepancy between the tumor location and each planning ITV was analyzed in each phase during VMAT-SBRT. The degrees of displacement were measured in the CC, LR, and AP directions. As a representative example, one of the acquired kV CBCT sets is shown in Figure 3. The distance of tumor motion during treatment is shown in Table 1. Using the SBF with abdominal pressure, the amplitude
Volume 86 Number 3 2013
Verification of PTV using 4D CBCT
429
of tumor motion caused by respiration was adequately suppressed to less than 10 mm in all directions. Table 2 presents the geometrical errors in each direction. The average displacements between ITV and actual tumor location during treatment were 0.41 0.93 mm, 0.15 0.58 mm, and 0.60 0.99 mm for the CC, LR, and AP directions, respectively. The discrepancy in each phase did not exceed 5 mm in any direction. In other words, the tumor position during VMAT was within the range of the PTV.
Discussion
Fig. 1. Example of the detection of sagittal displacement between the ITV and the actual tumor location, for which the ITV contoured on the planning CT was superimposed onto the intreatment 4D CBCT image. The displacement of the tumor location from the ITV was determined from the tumor edge.
Fig. 2.
Many previous studies revealed the fact that registration using CBCT images acquired immediately before treatment was useful for accurate setup (for SBRT see refs. 11 and 12). Furthermore, the CBCT images acquired during treatment provide the most reliable evaluation of tumor location during the actual treatment. As shown in Figure 3, an in-treatment 4D CBCT technique clearly visualized the location of the moving tumor. To the best of our knowledge, the present study is the first quantitative assessment evaluating 3D CBCT registration during VMAT-SBRT by using in-treatment 4D CBCT. This system could capture the tumor edge and provide a reasonable visualization of tumor location during treatment. Unlike pre-3D CBCT, successful tracking of tumor location using in-treatment 4D CBCT could provide the answer regardless of the delivery of appropriate irradiation. In addition, ITV and PTV settings were evaluated in VMAT-SBRT with intreatment 4D CBCT by comparing the tumor positions between the ITV (contouring based on 4D CT for treatment planning) and in-treatment 4D CBCT images simultaneously acquired during VMAT delivery. The discrepancy between the tumor edge and
Workflow of the verification of the PTV setting using in-treatment 4D CBCT.
430
International Journal of Radiation Oncology Biology Physics
Takahashi et al.
Fig. 3. Screenshots of Video E1 (available online) showing acquired in-treatment CBCT images overlaid with the planning ITV (pink curve). See supplementary Video E1 and E2 coronal/sagittal planes for a course of treatment (4 days). each planning ITV was measured, and it remained less than 5 mm in all directions. It would be essential for this result that the breathing amplitude was adequately suppressed by the SBF with an abdominal compressor. That is, the amplitudes of tumor motion in this study were less than 10 mm, and thus, the discrepancy between the actual tumor location and pre-setting for ITV was decreased compared with that of free-breathing treatment. Our result is comparable to previous data reported by Ueda et al (13), who, by using an electronic portal imaging device, demonstrated that the moving target during SBRT with a double-vacuum immobilization system was included in the PTV margin of 5 mm. Furthermore, the fact that VMAT-SBRT reduces the treatment time may permit further reductions in the PTV margin. Although the displacements were less than 5 mm in all directions, our results obtained using in-treatment 4D CBCT suggest that a relatively large difference tends to occur in the AP direction. In our additional exploration, the discrepancy observed in CBCT scans after registration was similar to that in couch shifts. Namely, the displacements in this study may be caused partly by the error of automatic registration of 3D couch shifts. Surely, the precision was relatively poor in the vertical direction. The improvement in the accuracy of couch shifts provides a potential reduction in treatment field margins in the AP direction. Daily in-treatment 4D CBCT does not prolong treatment time. However, one of the problems of using in-treatment 4D CBCT is Table 2 Displacements between the ITV and the actual tumor location during treatment for each direction Displacement (mm)
Craniocaudal Left-right Anterior-posterior
Within ITV 173 &1 21 &2 16 &3 4 &4 5 &5 1 Average (SD) 0.41 (0.93)
205 147 2 33 11 24 0 13 2 3 0 0 0.15 (0.58) 0.60 (0.99)
the additional exposure to kV x-ray irradiation. The total radiation exposure of in-treatment 4D CBCT scan was estimated to be as low as 30 mSv per day with our protocol, whereas the radiation exposure of pre-3D CBCT was approximately 15 mSv. These CBCT acquisitions delivered 4.5 cGy/fraction, which would result in an additional 18 cGy to a patient who received 50 Gy throughout the treatment period. Although in-treatment 4D CBCT imaging could provide an accurate verification for clinical treatment, it is difficult to identify baseline shifts in the tumor position, which are manifested as smaller apparent breathing motion and larger apparent tumor size. This is because the present 4D CBCT technique does not provide real-time respiratory motion but instead presents an “averaged” one. Further developments such as a “just-in-time tomography” (14) will be required to apply realistic tumor location. Other limitations of the accuracy reported in this study were that the ITV was created using only 2 phases, and the registration was performed with 3D CBCT. The motion of a tumor between maximum inspiration and expiration can be nonlinear. Therefore, the ITV in the current study might be underestimated for patients with large amplitudes of target motion. Likely, the PTV margin should be verified by as many phases as possible for in-treatment 4D CBCT. The effect caused by these limitations may be suppressed by using an abdominal compressor. Nevertheless, in our institution, improvements have been made for more accurate planning and delivery. One improvement is that the 4D CT for planning consists of 10 motion states, and GTV is delineated in each respiratory phase. These 10 GTVs are fused to form the ITV, and then, a uniform 5-mm margin is added to create the PTV. The second improvement is that similar to the IGRT, a manual 4D registration for each fraction was performed to align the ITV contours with the tumor target presented in the “pre-4D” CBCT images, and the treatment couch was shifted according to the matching result as the final target localization for VMAT delivery. A study of the registration accuracy with these improvements is in progress. In the future, the actual dose distribution will be evaluated by in-treatment CBCT using information such as the gantry angle, positional data of the multileaf collimator, and dose rate. This
Volume 86 Number 3 2013
Verification of PTV using 4D CBCT
method provides the total dose delivered to moving targets during the course of treatment. This type of review immediately after beam delivery is useful for subsequent treatment margin or dose adjustments, thus leading to adaptive radiation therapy.
5.
Conclusions
6.
With an in-treatment 4D CBCT, we confirmed that the required PTV margins were 5 mm when the registration for moving targets was performed using pre-3D CBCT. The present result may encourage the use of 3D CBCT registration even for a moving target. In addition, we conclude that in-treatment 4D CBCT is a simple and powerful tool for PTV margin determination that is easily applicable to rotational irradiation.
References
7.
8.
9. 10. 11.
1. Timmerman R, Paulus R, Galvin J, et al. Stereotactic body radiation therapy for inoperable early stage lung cancer. JAMA 2010;303: 1070-1076. 2. Fakiris AJ, McGarry RC, Yiannoutsos CT, et al. Stereotactic body radiation therapy for early-stage non-small-cell lung carcinoma: fouryear results of a prospective phase II study. Int J Radiat Oncol Biol Phys 2009;75:677-682. 3. Baumann P, Nyman J, Hoyer M, et al. Outcome in a prospective phase II trial of medically inoperable stage I non-small-cell lung cancer patients treated with stereotactic body radiotherapy. J Clin Oncol 2009;27:3290-3296. 4. McGrath SD, Matuszak MM, Yan D, et al. Volumetric modulated arc therapy for delivery of hypofractionated stereotactic lung
12.
13.
14.
431
radiotherapy: A dosimetric and treatment efficiency analysis. Radiother Oncol 2009;90:422-423. Holt A, van Vliet-Vroegindeweij C, Man A, et al. Volumetric-modulated arc therapy for stereotactic body radiotherapy of lung tumors: A comparison with intensity-modulated radiotherapy techniques. Int J Radiat Oncol Biol Phys 2011;81:1560-1567. Nakagawa K, Yamashita H, Shiraishi K, et al. Verification of intreatment tumor position using kilovoltage cone-beam computed tomography: a preliminary study. Int J Radiat Oncol Biol Phys 2007; 69:970-973. Nakagawa K, Haga A, Shiraishi K, et al. First clinical cone-beam CT imaging during volumetric modulated arc therapy. Radiother Oncol 2009;90:422-423. Kida S, Masutani Y, Yamashita H, et al. In-treatment 4D cone-beam CT reconstruction with image-based respiratory phase recognition. Radiol Phys Technol 2012;5:138-147. Feldkamp AL, Davis CL, Kress WJ. Practical cone-beam algorithm. J Opt Soc Am 1984;A16:612-619. Webb S. A modified convolution reconstruction technique for divergent beams. Phys Med Biol 1982;27:419-423. Wang Z, Wu QJ, Marks LB, et al. Cone-beam CT localization of internal target volumes for stereotactic body radiotherapy of lung lesions. Int J Radiat Oncol Biol Phys 2007;69:1618-1624. Grills IS, Hugo G, Kestin LL, et al. Image-guided radiotherapy via daily online cone-beam CT substantially reduces margin requirements for stereotactic lung radiotherapy. Int J Radiat Oncol Biol Phys 2008; 70:1045-1056. Ueda Y, Miyazaki M, Nishiyama K, et al. Craniocaudal safety margin calculation based on interfractional changes in tumor motion in lung SBRT assessed with an EPID in cine mode. Int J Radiat Oncol Biol Phys 2012;83:1064-1069. Pang G, Rowlands JA. Just-in-time tomography (JiTT): A new concept for image-guided radiation therapy. Phys. Med. Biol 2005;50: N323-N330.