- Email: [email protected]
A technique of using gated-CT images to determine internal target volume (ITV) for fractionated stereotactic lung radiotherapy
Radiotherapy and Oncology 78 (2006) 177–184 www.thegreenjournal.com
Lung radiotherapy
A technique of using gated-CT images to determine internal target volume (ITV) for fractionated stereotactic lung radiotherapy Jian-Yue Jin*, Munther Ajlouni, Qing Chen, Fang-Fang Yin1, Benjamin Movsas Department of Radiation Oncology, Henry Ford Hospital, Detroit, MI, USA
Abstract Background and purpose: To develop and evaluate a technique and procedure of using gated-CT images in combination with PET image to determine the internal target volume (ITV), which could reduce the planning target volume (PTV) with adequate target coverage. Patients and methods: A skin marker-based gating system connected to a regular single slice CT scanner was used for this study. A motion phantom with adjustable motion amplitude was used to evaluate the CT gating system. Specifically, objects of various sizes/shapes, considered as virtual tumors, were placed on the phantom to evaluate the number of phases of gated images required to determine the ITV while taking into account tumor size, shape and motion. A procedure of using gated-CT and PET images to define ITV for patients was developed and was tested in patients enrolled in an IRB approved protocol. Results: The CT gating system was capable of removing motion artifacts for target motion as large as 3-cm when it was gated at optimal phases. A phantom study showed that two gated-CT scans at the end of expiration and the end of inspiration would be sufficient to determine the ITV for tumor motion less than 1-cm, and another mid-phase scan would be required for tumors with 2-cm motion, especially for small tumors. For patients, the ITV encompassing visible tumors in all sets of gated-CT and regular spiral CT images seemed to be consistent with the target volume determined from PET images. PTV expanded from the ITV with a setup uncertainty margin had less volume than PTVs from spiral CT images with a 10-mm generalized margin or an individualized margin determined at fluoroscopy. Conclusions: A technique of determining the ITV using gated-CT images was developed and was clinically implemented successfully for fractionated stereotactic lung radiotherapy. q 2005 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology 78 (2006) 177–184. Keywords: CT images; Internal target volume (ITV); Lung radiotherapy
More than 2.5 cm tumor movement has been reported for lung cancer patients during a respiration cycle [1–5]. There is also a great deal of variability of tumor motion (from a few millimeters to several centimeters) among patients depending on tumor location and extent of emphysema. For a regular free breathing computed tomography (CT) simulation, the images could be acquired at an arbitrary phase. Therefore, using a planning target volume (PTV) with a standardized margin may either include more normal lung than necessary or result in missing the target during certain phases of the respiratory cycle. The development of 4D-CT has made it possible to define the volume of tumor excursion during a respiration cycle on the CT images relating to a patient’s fixed anatomic landmarks such as the vertebrae
1
Current address: Duke University Medical Center, Radiation Oncology, Durham, NC, USA.
body. This volume can be considered as the internal target volume (ITV) as defined by the ICRU report 62 [6,7]. It was recently reported that using multiple sets of 4D-CT images to determine the ITV could substantially reduce the PTV while safely covering the target [8]. Other techniques have also been proposed to determine the ITV. Lagerwaard et al. used multiple ‘slow’ CTs to generate the ITV [9,10]. Balter et al. combined two breath-hold scans at the end of inspiration and at the end of expiration [11]. Shih et al. have combined the two breath-hold scans with three slow spiral CT scans [12]. Caldwell et al. also proposed using the positron emission tomography (PET) image to determine the ITV [13]. The PET images are usually acquired over a time period of 30 min and hence the functional enhancement volume may be a better representation of the geometric boundaries of the ITV. However, using PET alone to define the ITV is not yet clinically acceptable because of many uncertainties associated with the PET images [14]. For
0167-8140/$ - see front matter q 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.radonc.2005.11.012
178
A technique of using gated CT images to determine ITV
example, PET volume delineation depends on the threshold value of the SUV, while the uptake of Fluorodeoxyglucose (FDG) for lung cancer varies widely from patient to patient. In addition, the resolution of PET is usually worse than 5 mm, which could induce large variation in defining the target volume. Theoretically, using 4D-CT images to determine the ITV should be the best way to approach ITV accurately. Usually, 10 sets of images are acquired in 4D-CT scan for a patient, which could be a burden for the treatment planning computer as well as to the radiation oncologists to delineate the target in each image sets. In addition, the few sets of data, the less radiation doses to patients. In this paper, we report a technique of using a regular single slice CT scanner to acquire gated CT images at different phases, and using these CT images in combination with diagnostic PET scan to determine the ITV. A phantom study is also designed to answer the question of how many sets of 4D-CT images are sufficient to determine the ITV.
Materials and methods CT-gating system A gating system (CTGating V0.5, research version, BrainLAB, Heimstetten, Germany) based on external skin markers, and connected to a regular single slice AcQSIM CT scanner (Phillips Medical Systems, Highland Heights, OH), was used for gated CT scanning. The system uses two infrared cameras attached on the ceiling of the CT room to detect the positions of 4–6 infrared markers placed on the surface of an imaging object/patient. The gating control software analyzes the positions of the markers and generates the motion cycle signal. An amplitude-based gating algorithm is used to generate the gating signals to control the CT scanner. As shown in Fig. 1, the system works
either in a ‘window’ mode or a ‘level’ mode. In the ‘window’ mode, the gating signal turns into ‘trigger-on’ status when the motion signal reaches the edge of the gating window and turns into ‘trigger-off’ status when the motion signal leaves the gating window. In the ‘level’ mode, the gating signal turns into ‘trigger-on’ status when the motion signal reaches the defined ‘level’, and automatically turns into ‘trigger-off’ status after a short time. The ‘level’ mode is used for controlling the CT scanner for ‘axial’ scanning, because the image acquisition process will be turned on when the gating signal changes to ‘trigger-on’ status, and will continue once the scanner is on, even when the gating signal changes back to the ‘trigger-off’ status, until the acquisition for one slice is completed. There are two trigger points corresponding to one gating level, one is when the motion signal is ‘rising’ and is defined as ‘rising trigger’, the other is when the motion signal is ‘falling’ and is defined as ‘falling trigger’. They can be selected separately to control the CT scanner. The AcQSIM CT scanner is a relatively slow scanner with a gantry rotation time of one second for a 3608 rotation to acquire a gated axial slice. The ‘partial’ scan angle (3198) mode was used for the gated CT in this study to reduce motion artifacts. Therefore, the imaging acquisition time for each gated CT slice (or the gating window) is approximately 0.89 s. In addition, there is a time delay of about 0.25 s for the control system to detect a trigger-on signal and to start scanning. Three-millimeter slice thickness was used for all gated CT scanning.
Phantom study A simple motion phantom as described in Ref. [15] was used in this study. It consists of a metal plate driven by a lever attached to a rotating wheel. Besides the infrared markers, objects functioning as virtual tumors are placed on a thin layer of styrofoam attached to the top surface of the
Fig. 1. Illustration of the gating signal generated by amplitude-based gating system in the window mode and level mode.
J.-Y. Jin et al. / Radiotherapy and Oncology 78 (2006) 177–184
metal plate and move with the plate in both horizontal (longitudinal) and vertical directions. The amplitude of the phantom motion can be adjusted by attaching the lever to different positions of the rotating wheel. The period of one cycle of movement is approximately 4.6 s. To evaluate the number of gated image sets required to determine the ITV, we have acquired gated CT scans at different phases with different motion amplitudes for four different objects, which would represent the size range of the lung lesions treated with fractionated stereotactic radiotherapy. Object A is a sphere of 4 cm diameter. Object B is a 2.8 cm diameter by 2.5 cm high cylinder. Object C is irregularly shaped with a maximum dimension of 2!3!5 cc. Object D is another sphere with a 1.3 cm diameter (an infrared marker). Five different gating phases were mainly used. They were: (1) peak phase (or inspiration phase for patients), (2) bottom phase (or expiration phase for patients), (3) middle phase 1, (4) middle phase 2, and (5) middle phase 3, with the middle point of each image acquisition corresponds to points A, B, C, D, E in Fig. 1, respectively. The corresponding trigger point for each gating phase was calculated by taking into account the CT time delay and the duration of scanning time for one slice. All the gated CT images, along with the regular spiral CT image, were imported to the treatment planning system (TPS) (Brainscan version 3.3, BrainLAB, Heimstetten, Germany), and fused together using the autofusion tool in the TPS for ITV analysis.
Gated CT procedure for patients A patient lies on a cradle or a vacuum bag in a supine position during simulation. The patient is first observed under a Fluoroscopy to estimate the target/diaphragm movement in both AP and lateral directions. The tumor is often difficult to see under the Fluoroscopy for some patients so the movements of diaphragm or the part of lung where the tumor is estimated to be located are considered. This procedure was initially used as a method to obtain an individualized PTV margin for each patient for the stereotactic lung radiotherapy. Before CT scanning, five infrared markers are placed on the surface of the thoracic/abdominal area. The infrared markers have two purposes: (1) They are used as the markers for the initial setup of our image guided radiotherapy (ExacTrac 3.5, BrainLAB, Heimstetten, Germany) procedure [16]. In this case, the markers have to be placed at locations with minimal movement. (2) They are used as the external motion surrogate for gated CT. In this case, some of the makers are placed at locations with maximal movement. Generally, two markers are placed on the chest surface, and three markers are placed at the upper abdominal area. The average vertical movement of the five markers is used as the external motion signal so that the couch movement in longitudinal direction during scanning would not affect the motion signal. The patient is first scanned with a regular spiral CT including the entire thorax section with the markers placed on the chest surface. The slice at the center of the target is then identified. The center of the target is then marked at patient’s skin and will be used as the isocenter for treatment. Three infrared markers are then moved to the
179
upper abdominal area. Gated CT is then performed to include the target with 3–4 cm extension in each direction. Gated CTs with different trigger points corresponding to an end of expiration phase, an end of inspiration phase and one or two middle phases are carried out for each patient. During the whole procedure, the patient is breathing normally. No patient coaching is required except an explanation of the procedure by the MD or a technologist. The whole procedure takes about one and half hours, adding about 30–40 min to the original single spiral CT procedure, which usually takes about 45–60 min. The spiral CT image set and the 3–4 gated image sets along with the isocenter information are imported into the TPS. The gated CT image sets are then registered to the spiral CT image set using an autofusion tool in the TPS system. The fusion result is verified by reviewing the match of the vertebrae body of the two images. The precision of the registration is estimated to be within 1–1.5 mm. This estimation is based on the observation that manually shifting one image set 1–1.5 mm in any directions would induce visible mismatch between two image sets. The diagnostic PET/CT images (Reveal RT, CTI, Knoxville, Tennessee, USA) are also loaded into the TPS and fused with the simulation CT images. To fuse the PET images, we fuse the diagnostic CT image with the simulation CT images first, and then fuse the PET with its diagnostic CT images. The visible tumor in each gated CT and spiral CT image sets is contoured separately as different targets. ITV is then drawn slice by slice in the spiral CT image set to conformly encompass all the contoured targets superimposed on the image. PET volumes are autocontoured with different threshold values. The percentage of the maximal image number is used as the threshold value because the original PET SUV information is lost when the PET images are imported into the TPS. Currently, PET volumes are only used to compare with the ITV to evaluate the threshold value. A 3-mm setup uncertainty margin plus a 0–2 mm residual motion margin (over all 3–5 mm margin) is usually applied to the ITV to form a PTV. The residual motion is mainly due to patient breathing irregularity. Some patients could have relatively more breathing irregularity, which would require a relative larger residual margin for PTV expansion. Intensity modulated radiotherapy (IMRT) or 3D conformal technique with 6–8 fields is used to treat the patient. The prescription dose is 12 Gy!4 fractions to the 95% dose line to cover the PTV. The organs at risk (lung, esophagus, and cord) are delineated based on the regular spiral CT image set. The dose calculation is also based on this image set.
Results Phantom data Fig. 2(a)–(d) show the sagittal images of the moving phantom with a 3 cm motion amplitude using (a) spiral CT, (b) axial CT without gating, (c) axial CT with gating at the end of expiration phase (middle point of image acquisition at point B in Fig. 1), and (d) spiral CT with no phantom motion, respectively. We note that for both the spiral CT and the axial CT without gating, the motion artifacts are very significant, and the two show a different pattern of artifacts.
180
A technique of using gated CT images to determine ITV
Fig. 2. Sagittal plane view of the CT images of the motion phantom (a) using regular spiral scan, (b) using axial scan without gating, (c) using axial scan with gating at the bottom phase, and (d) using regular spiral scan without phantom motion.
However, the gated CT image is almost identical to the images without motion. Gated CT scan taken at the end of inspiration phase showed similar image quality with no motion artifacts. This demonstrates that gated CT scan can be successfully performed in a regular single slice CT scanner with our gating system, even for a moving object with a motion-amplitude as larger as 3 cm. We have evaluated the ITV for each virtual tumor with the target motion amplitude being at (1) 1.1 cm, (2) 2.1 cm and (3) 3.0 cm in the longitudinal direction. Various ITVs were delineated by combining different gated CT image sets. We define ITV2 Z GTVP C GTVB ITV2:5 Z GTVP C GTVB C GTVS ITV3 Z GTVP C GTVB C GTVM1 ITV3:5 Z GTVP C GTVB C GTVM1 C GTVS ITV4 Z GTVP C GTVB C GTVM2 C GTVM3 ITV5 Z GTVP C GTVB C GTVM1 C GTVM2 C GTVM3 where GTVP, GTVB, GTVS, GTVM1, GTVM2 and GTVM3 are the delineated target volume in the peak phase, bottom phase, regular spiral CT, middle phase1, middle phase2 and middle phase3 CT image sets, respectively. A real ITV could be defined as the combination of GTVs from infinite number of gated CT images at different phases. We found that it could be reasonable approached by ITV5 for the conditions in our study. Table 1 shows these ITVs and their ratios to the real ITV of each virtual tumor with different object motion amplitudes. We note that when target motion equals 1.1 cm, even for the smallest target, the ITV determined by the image sets of the peak and the bottom phases reasonably represents the real ITV. When target motion equals 2.1 cm, ITV2 or even ITV2.5 are not sufficient to cover all the tumor excursions for most tumors, especially for small tumors. However, ITV3 and ITV3.5 seem to be sufficient to represent the real ITV. For target motion equals 3.0 cm, both ITV3 and ITV3.5 do not seem enough to cover all the tumor excursions, especially for small tumors. However, for large tumors, the difference
seems to be small (within 5%). We should point out that the volumes of several ITVs in the table are shown to be slightly larger than the volume of corresponding ITV5. This could be due to the uncertainty of drawing the ITV to encompass the GTVs and the limited spatial resolution (grid size) to define the targets.
Application in patients Gated-CT images have been acquired for 10 patients according to our IRB approved protocol. Table 2 lists the tumor location, GTV (spiral), ITV and tumor movement (tumor center to center) for these patients. A detailed analysis of the patient data will be reported after enough patients are accrued in the study. Here we present an example of how this technique is applied for patient treatment planning. The patient has a peripheral lesion in the left low lobe. Fig. 3(a)–(d) show the coronal and sagittal planes of the CT images of (a) gated at the expiration phase, (b) gated at the inspiration phase, (c) gated at a middle phase, and (d) a regular spiral CT, respectively, with the corresponding GTVs and ITV delineated. We note that the position difference of the inferior edge between the inspiration and expiration phases is about 1.6 cm. This is consistent with the 1.5 cm of inferior/superior (I/S) motion observed at fluoroscopy. The gated images at three different phases show the tumor in different locations related to the patient’s vertebrae. The expiration image shows almost no motion artifacts in the diaphragm, however, the motion artifacts are visible in the inspiration and the middle phases. Fig. 4 shows the coronal plane of the PET images of the patient with the contours of GTVs from the expiration phase (GTVex), the inspiration phase (GTVin) and the middle phase (GTVmid) superimposed. The combination of the three GTVs seems to be consistent with the tumor enhancement of the PET image both in shape and size. This gives us another confirmation that the ITV encompassing the three gated volumes plus the spiral CT would represent the tumor excursion. However, the diagnostic PET/CT images were taken with the patient lying in a slightly different position as compared to the simulation CT. The match of PET volume to the ITV is not perfect. In addition, the PET volume could change significantly for different threshold values. Currently
J.-Y. Jin et al. / Radiotherapy and Oncology 78 (2006) 177–184
181
Table 1 Volumes of different ITVs and their ratios to the corresponding real ITV for four different objects with tumor motions of 1.1, 2.2, and 3.0 cm. Target motion (cm)
Target volume
Object A Volume (cc)
MotionZ1.1
MotionZ2.2
MotionZ3.0
ITV2 ITV2.5 ITV3 ITV3.5 ITV4 ITV5 ITV2 ITV2.5 ITV3 ITV3.5 ITV4 ITV5 ITV2 ITV2.5 ITV3 ITV3.5 ITV4 ITV5
45.8 46.2 46.1 46.1 46.1 46 55 55.9 57.2 56.9 56.8 56.7 62.4 66 70.5 70.5 69.7 71.7
Object B Ratio 0.996 1.004 1.002 1.002 1.002 1.000 0.970 0.986 1.009 1.004 1.002 1.000 0.870 0.921 0.983 0.983 0.972 1.000
Volume (cc) 24.9 25.1 25.2 25 25.2 25.1 32.9 33.5 35.1 35 35.4 35.5 35.1 38.5 41.7 42.6 42.7 43.5
we only use PET volume as a method of verification. We have started to take PET/CT scans for some patients with simulation cradles. Because image guided radiotherapy is used in our dose delivery, the setup uncertainty is generally less that 3 mm [16]. Therefore, a 3 mm margin could be applied to the ITV to form the PTV. We used a 3 mm margin in the AP and lateral directions, and a 5 mm margin in the superior/ inferior directions to form the PTV for this patient. The 5 mm margin is to take into account possible residual motion during the gated CT imaging. As shown in Fig. 3, considerable motion artifact is still visible in the inspiration phase, suggesting that the image was not always acquired at the same inspiration phase due to breathing irregularity. Fig. 5 shows a coronal plane of a spiral CT image with PTVITV, PTV10 mm and PTVindiv drawn. The PTVITV is expanded from ITV with the margin mentioned earlier. The PTV10 mm is expanded 10 mm in all directions from the GTV of the spiral
Object C Ratio 0.992 1.000 1.004 0.996 1.004 1.000 0.927 0.944 0.989 0.986 0.997 1.000 0.807 0.885 0.959 0.979 0.982 1.000
Object D
Volume (cc)
Ratio
Volume (cc)
Ratio
25 24.7 25 24.8 25 25 31.2 31.5 32.6 32.5 32.5 32.4 34.6 36.5 38.7 39.2 39.8 40.1
1 0.988 1 0.992 1 1 0.963 0.972 1.006 1.003 1.003 1.000 0.863 0.910 0.965 0.978 0.993 1.000
3.08 3.09 3.25 3.08 3.14 3.13 3.19 3.59 4.35 4.31 4.38 4.35 3.1 4.05 5.62 5.78 6.53 6.78
0.984 0.987 1.038 0.984 1.003 1.000 0.733 0.825 1.000 0.991 1.007 1.000 0.457 0.597 0.829 0.853 0.963 1.000
CT, which is commonly used as a PTV margin at some institutions for lung patients without any motion information [8]. The PTVindiv is the individualized PTV, which is expanded from GTV of the spiral CT with the margin corresponding to the motion observed at fluoroscopy, which was commonly used at our institution for peripheral lesions prior to gated CT availability. In this case, because we observed on fluoroscopy that the target motion is about 1.5 cm in I/S direction, and less than 5 mm in the other directions, we use the 1.5 cm margin at I/S direction, and 5 mm margins for the AP and lateral direction. The PTV volumes are 24.2 CC, 41.9 CC, 34.8 CC for PTVITV, PTV10 mm and PTVindiv, respectively. In addition, we note that there is a fairly large part of PTVITV that is not included in the PTV10 mm, and a small part not included in the PTVindiv. These suggest that using PTVITV can not only reduce the irradiation volume, but more importantly, have less chance of missing the target.
Table 2 List of the tumor location, GTV (spiral), ITV and tumor movement (tumor center to center) for 10 patients enrolled in the study Patient
Location
GTV (cc)
ITV (cc)
Motion X (mm)
Motion Y (mm)
Motion Z (mm)
Motion overall
1 2 3 4 5 6 7 8 9 10
LLL RLL RLLa LLL LUL LUL Rt hilum RUL RUL RUL
52.45 11.66 4.55 5.21 4.71 43.88 6.48 18.73 32.02 5.22
89.01 26.1 5.16 10.09 7.25 50.79 9.86 30.34 51.71 7.97
3.2 K3.14 0.19 2.82 K1.34 K0.97 0.15 K1.42 K1.59 K1.73
K3.1 K3.75 0.16 3.88 K2.72 K0.28 K0.29 1.42 3.74 K1.67
15.56 13.68 1.07 15.02 3.2 K0.68 K0.85 3.88 3.9 3.19
16.2 14.5 1.1 15.8 4.4 1.2 0.9 4.4 5.6 4.0
a
For patient 3, the tumor is attached on the chest wall.
182
A technique of using gated CT images to determine ITV
Fig. 3. Coronal and sagittal plane view of the CT images of a patient at different gating phases with corresponding GTV and ITV delineated. (a) Gated at the expiration phase, (b) gated at the inspiration phase, (c) gated at a middle phase, and (d) a regular spiral scan.
J.-Y. Jin et al. / Radiotherapy and Oncology 78 (2006) 177–184
183
reasonably good to determine the ITV. This suggests that using two breath-hold CT sets with a free breathing spiral CT could be an alternative to determine the ITV. However, it should be kept in mind that the tumor positions of breathhold and free-breath at the end of inspiration or expiration phases could be different. In addition, the quality control of breathing hold solely based on patients’ cooperation should also be considered. The target motion in this phantom study is a simple back and forth motion that the motion trajectory from the bottom to the peak is the same as that of peak to bottom position. It has been reported that some tumors could move in a hysteretic trajectory during a respiratory cycle [3,4]. In that case, another middle phase at a different tumor motion direction should be used to determine the ITV.
Conclusions Fig. 4. Coronal plane view of the PET images of a patient with the contours of GTV of CT images at different phases superimposed.
Discussion We have developed a gated CT technique in regular single slice CT scanner. The phantom study demonstrated that using three gated CT image sets, one in inspiration phase, one in expiration phase, and one in a middle phase, is reasonably safe to determine the ITV for a target with movement less than 2 cm. The PTV from this ITV can have better coverage of the target and reduce the target volume. The phantom study results also demonstrated that for a target with motion less than 1.1 cm, using only the image sets of the inspiration and the expiration phases could be
We have developed a gated CT technique in regular single slice CT scanner. The phantom study demonstrated that using three gated CT image sets, one in inspiration phase, one in expiration phase, and one in a middle phase, is reasonably safe to determine the ITV for a target with movement less than 2 cm. The PTV from this ITV can have better coverage of the target and reduce the target volume.
* Corresponding author. Jian-Yue Jin Tel.: C1 313 916 0234; fax: C1 313 916 3235. E-mail address: [email protected] : Received 18 May 2005; received in revised form 28 November 2005; accepted 30 November 2005; available online 10 January 2006
References
Fig. 5. The contours of PTVITV, PTV10 mm and PTVindiv of the patient in a regular spiral CT image’s coronal view.
[1] Ross C, Hussey D, Pennington E. Analysis of movement of intrathoracic neoplasms using ultrafast computed tomography. Int J Radiat Oncol Biol Phys 1990;18:671–7. [2] Seppenwoolde Y, Shirato H, Kitamura K, et al. Precise and realtime measurement of 3D tumor motion in lung due to breathing and heartbeat, measured during radiotherapy. Int J Radiat Oncol Biol Phys 2002;53:822–34. [3] Shirato H, Seppenwoolde Y, Kitamura K, Onimura R, Shimizu S. Intrafractional tumor motion: lung and liver. Semin Radiat Oncol 2004;14:10–18. [4] Mageras GS, Pevsner A, Yorke ED, et al. Measurement of lung tumor motion using respiration-correlated CT. Int J Radiat Oncol Biol Phys 2004;60:933–41. [5] Plathow C, Fink C, Ley S, et al. Measurement of tumor diameter-dependent mobility of lung tumors by dynamic MRI. Radiother Oncol 2004;73:349–54. [6] International Commission on Radiation Units and Measurements. Prescribing, recording, and reporting photon beam therapy. Supplement to report 50. Report 62. Washington, DC: ICRU; 1999. [7] Purdy JA. Current ICRU definitions of volumes: limitations and future directions. Semin Radiat Oncol 2004;14:27–40. [8] Underberg RW, Lagerwaard FJ, Cuijpers JP, Slotman BJ, van Sornsen de Koste JR, Senan S. Four-dimensional CT scans for treatment planning in stereotactic radiotherapy for stage I lung cancer. Int J Radiat Oncol Biol Phys 2004;60:1283–90.
184
A technique of using gated CT images to determine ITV
[9] Lagerwaard FJ, Van Sornsen de Koste JR, Nijssen-Visser MR, et al. Multiple ‘slow’ CT scans for incorporating lung tumor mobility in radiotherapy planning. Int J Radiat Oncol Biol Phys 2001;51:932–7. [10] de Koste JR, Lagerwaard FJ, de Boer HC, Nijssen-Visser MR, Senan S. Are multiple CT scans required for planning curative radiotherapy in lung tumors of the lower lobe? Int J Radiat Oncol Biol Phys 2003;55:1394–9. [11] Balter JM, Ten Haken RK, Lawrence TS, et al. Uncertainties in CT-based radiation therapy treatment planning associated with patient breathing. Int J Radiat Oncol Biol Phys 1996;36:167–74. [12] Shih HA, Jiang SB, Aljarrah KM, Doppke KP, Choi NC. Internal target volume determined with expansion margins beyond composite gross tumor volume in three-dimensional conformal radiotherapy for lung cancer. Int J Radiat Oncol Biol Phys 2004;60:613–22.
[13] Caldwell CB, Mah K, Skinner M, Danjoux CE. Can PET provide the 3D extent of tumor motion for individualized internal target volumes? A phantom study of the limitations of CT and the promise of PET. Int J Radiat Oncol Biol Phys 2003;55: 1381–93. [14] Lavrenkov K, Partridge M, Cook G, Brada M. Positron emission tomography for target volume definition in the treatment of non-small cell lung cancer. Radiother Oncol 2005;77:1–4. [15] Jin JY, Yin FF. Time delay measurement for linac based treatment delivery in synchronized respiratory gating radiotherapy. Med Phys 2005;32:1293–6. [16] Yin FF, Ryu S, Ajlouni M, et al. A technique of intensitymodulated radiosurgery (IMRS) for spinal tumors. Med Phys 2002;29:2815–22.
Lung radiotherapy
A technique of using gated-CT images to determine internal target volume (ITV) for fractionated stereotactic lung radiotherapy Jian-Yue Jin*, Munther Ajlouni, Qing Chen, Fang-Fang Yin1, Benjamin Movsas Department of Radiation Oncology, Henry Ford Hospital, Detroit, MI, USA
Abstract Background and purpose: To develop and evaluate a technique and procedure of using gated-CT images in combination with PET image to determine the internal target volume (ITV), which could reduce the planning target volume (PTV) with adequate target coverage. Patients and methods: A skin marker-based gating system connected to a regular single slice CT scanner was used for this study. A motion phantom with adjustable motion amplitude was used to evaluate the CT gating system. Specifically, objects of various sizes/shapes, considered as virtual tumors, were placed on the phantom to evaluate the number of phases of gated images required to determine the ITV while taking into account tumor size, shape and motion. A procedure of using gated-CT and PET images to define ITV for patients was developed and was tested in patients enrolled in an IRB approved protocol. Results: The CT gating system was capable of removing motion artifacts for target motion as large as 3-cm when it was gated at optimal phases. A phantom study showed that two gated-CT scans at the end of expiration and the end of inspiration would be sufficient to determine the ITV for tumor motion less than 1-cm, and another mid-phase scan would be required for tumors with 2-cm motion, especially for small tumors. For patients, the ITV encompassing visible tumors in all sets of gated-CT and regular spiral CT images seemed to be consistent with the target volume determined from PET images. PTV expanded from the ITV with a setup uncertainty margin had less volume than PTVs from spiral CT images with a 10-mm generalized margin or an individualized margin determined at fluoroscopy. Conclusions: A technique of determining the ITV using gated-CT images was developed and was clinically implemented successfully for fractionated stereotactic lung radiotherapy. q 2005 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology 78 (2006) 177–184. Keywords: CT images; Internal target volume (ITV); Lung radiotherapy
More than 2.5 cm tumor movement has been reported for lung cancer patients during a respiration cycle [1–5]. There is also a great deal of variability of tumor motion (from a few millimeters to several centimeters) among patients depending on tumor location and extent of emphysema. For a regular free breathing computed tomography (CT) simulation, the images could be acquired at an arbitrary phase. Therefore, using a planning target volume (PTV) with a standardized margin may either include more normal lung than necessary or result in missing the target during certain phases of the respiratory cycle. The development of 4D-CT has made it possible to define the volume of tumor excursion during a respiration cycle on the CT images relating to a patient’s fixed anatomic landmarks such as the vertebrae
1
Current address: Duke University Medical Center, Radiation Oncology, Durham, NC, USA.
body. This volume can be considered as the internal target volume (ITV) as defined by the ICRU report 62 [6,7]. It was recently reported that using multiple sets of 4D-CT images to determine the ITV could substantially reduce the PTV while safely covering the target [8]. Other techniques have also been proposed to determine the ITV. Lagerwaard et al. used multiple ‘slow’ CTs to generate the ITV [9,10]. Balter et al. combined two breath-hold scans at the end of inspiration and at the end of expiration [11]. Shih et al. have combined the two breath-hold scans with three slow spiral CT scans [12]. Caldwell et al. also proposed using the positron emission tomography (PET) image to determine the ITV [13]. The PET images are usually acquired over a time period of 30 min and hence the functional enhancement volume may be a better representation of the geometric boundaries of the ITV. However, using PET alone to define the ITV is not yet clinically acceptable because of many uncertainties associated with the PET images [14]. For
0167-8140/$ - see front matter q 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.radonc.2005.11.012
178
A technique of using gated CT images to determine ITV
example, PET volume delineation depends on the threshold value of the SUV, while the uptake of Fluorodeoxyglucose (FDG) for lung cancer varies widely from patient to patient. In addition, the resolution of PET is usually worse than 5 mm, which could induce large variation in defining the target volume. Theoretically, using 4D-CT images to determine the ITV should be the best way to approach ITV accurately. Usually, 10 sets of images are acquired in 4D-CT scan for a patient, which could be a burden for the treatment planning computer as well as to the radiation oncologists to delineate the target in each image sets. In addition, the few sets of data, the less radiation doses to patients. In this paper, we report a technique of using a regular single slice CT scanner to acquire gated CT images at different phases, and using these CT images in combination with diagnostic PET scan to determine the ITV. A phantom study is also designed to answer the question of how many sets of 4D-CT images are sufficient to determine the ITV.
Materials and methods CT-gating system A gating system (CTGating V0.5, research version, BrainLAB, Heimstetten, Germany) based on external skin markers, and connected to a regular single slice AcQSIM CT scanner (Phillips Medical Systems, Highland Heights, OH), was used for gated CT scanning. The system uses two infrared cameras attached on the ceiling of the CT room to detect the positions of 4–6 infrared markers placed on the surface of an imaging object/patient. The gating control software analyzes the positions of the markers and generates the motion cycle signal. An amplitude-based gating algorithm is used to generate the gating signals to control the CT scanner. As shown in Fig. 1, the system works
either in a ‘window’ mode or a ‘level’ mode. In the ‘window’ mode, the gating signal turns into ‘trigger-on’ status when the motion signal reaches the edge of the gating window and turns into ‘trigger-off’ status when the motion signal leaves the gating window. In the ‘level’ mode, the gating signal turns into ‘trigger-on’ status when the motion signal reaches the defined ‘level’, and automatically turns into ‘trigger-off’ status after a short time. The ‘level’ mode is used for controlling the CT scanner for ‘axial’ scanning, because the image acquisition process will be turned on when the gating signal changes to ‘trigger-on’ status, and will continue once the scanner is on, even when the gating signal changes back to the ‘trigger-off’ status, until the acquisition for one slice is completed. There are two trigger points corresponding to one gating level, one is when the motion signal is ‘rising’ and is defined as ‘rising trigger’, the other is when the motion signal is ‘falling’ and is defined as ‘falling trigger’. They can be selected separately to control the CT scanner. The AcQSIM CT scanner is a relatively slow scanner with a gantry rotation time of one second for a 3608 rotation to acquire a gated axial slice. The ‘partial’ scan angle (3198) mode was used for the gated CT in this study to reduce motion artifacts. Therefore, the imaging acquisition time for each gated CT slice (or the gating window) is approximately 0.89 s. In addition, there is a time delay of about 0.25 s for the control system to detect a trigger-on signal and to start scanning. Three-millimeter slice thickness was used for all gated CT scanning.
Phantom study A simple motion phantom as described in Ref. [15] was used in this study. It consists of a metal plate driven by a lever attached to a rotating wheel. Besides the infrared markers, objects functioning as virtual tumors are placed on a thin layer of styrofoam attached to the top surface of the
Fig. 1. Illustration of the gating signal generated by amplitude-based gating system in the window mode and level mode.
J.-Y. Jin et al. / Radiotherapy and Oncology 78 (2006) 177–184
metal plate and move with the plate in both horizontal (longitudinal) and vertical directions. The amplitude of the phantom motion can be adjusted by attaching the lever to different positions of the rotating wheel. The period of one cycle of movement is approximately 4.6 s. To evaluate the number of gated image sets required to determine the ITV, we have acquired gated CT scans at different phases with different motion amplitudes for four different objects, which would represent the size range of the lung lesions treated with fractionated stereotactic radiotherapy. Object A is a sphere of 4 cm diameter. Object B is a 2.8 cm diameter by 2.5 cm high cylinder. Object C is irregularly shaped with a maximum dimension of 2!3!5 cc. Object D is another sphere with a 1.3 cm diameter (an infrared marker). Five different gating phases were mainly used. They were: (1) peak phase (or inspiration phase for patients), (2) bottom phase (or expiration phase for patients), (3) middle phase 1, (4) middle phase 2, and (5) middle phase 3, with the middle point of each image acquisition corresponds to points A, B, C, D, E in Fig. 1, respectively. The corresponding trigger point for each gating phase was calculated by taking into account the CT time delay and the duration of scanning time for one slice. All the gated CT images, along with the regular spiral CT image, were imported to the treatment planning system (TPS) (Brainscan version 3.3, BrainLAB, Heimstetten, Germany), and fused together using the autofusion tool in the TPS for ITV analysis.
Gated CT procedure for patients A patient lies on a cradle or a vacuum bag in a supine position during simulation. The patient is first observed under a Fluoroscopy to estimate the target/diaphragm movement in both AP and lateral directions. The tumor is often difficult to see under the Fluoroscopy for some patients so the movements of diaphragm or the part of lung where the tumor is estimated to be located are considered. This procedure was initially used as a method to obtain an individualized PTV margin for each patient for the stereotactic lung radiotherapy. Before CT scanning, five infrared markers are placed on the surface of the thoracic/abdominal area. The infrared markers have two purposes: (1) They are used as the markers for the initial setup of our image guided radiotherapy (ExacTrac 3.5, BrainLAB, Heimstetten, Germany) procedure [16]. In this case, the markers have to be placed at locations with minimal movement. (2) They are used as the external motion surrogate for gated CT. In this case, some of the makers are placed at locations with maximal movement. Generally, two markers are placed on the chest surface, and three markers are placed at the upper abdominal area. The average vertical movement of the five markers is used as the external motion signal so that the couch movement in longitudinal direction during scanning would not affect the motion signal. The patient is first scanned with a regular spiral CT including the entire thorax section with the markers placed on the chest surface. The slice at the center of the target is then identified. The center of the target is then marked at patient’s skin and will be used as the isocenter for treatment. Three infrared markers are then moved to the
179
upper abdominal area. Gated CT is then performed to include the target with 3–4 cm extension in each direction. Gated CTs with different trigger points corresponding to an end of expiration phase, an end of inspiration phase and one or two middle phases are carried out for each patient. During the whole procedure, the patient is breathing normally. No patient coaching is required except an explanation of the procedure by the MD or a technologist. The whole procedure takes about one and half hours, adding about 30–40 min to the original single spiral CT procedure, which usually takes about 45–60 min. The spiral CT image set and the 3–4 gated image sets along with the isocenter information are imported into the TPS. The gated CT image sets are then registered to the spiral CT image set using an autofusion tool in the TPS system. The fusion result is verified by reviewing the match of the vertebrae body of the two images. The precision of the registration is estimated to be within 1–1.5 mm. This estimation is based on the observation that manually shifting one image set 1–1.5 mm in any directions would induce visible mismatch between two image sets. The diagnostic PET/CT images (Reveal RT, CTI, Knoxville, Tennessee, USA) are also loaded into the TPS and fused with the simulation CT images. To fuse the PET images, we fuse the diagnostic CT image with the simulation CT images first, and then fuse the PET with its diagnostic CT images. The visible tumor in each gated CT and spiral CT image sets is contoured separately as different targets. ITV is then drawn slice by slice in the spiral CT image set to conformly encompass all the contoured targets superimposed on the image. PET volumes are autocontoured with different threshold values. The percentage of the maximal image number is used as the threshold value because the original PET SUV information is lost when the PET images are imported into the TPS. Currently, PET volumes are only used to compare with the ITV to evaluate the threshold value. A 3-mm setup uncertainty margin plus a 0–2 mm residual motion margin (over all 3–5 mm margin) is usually applied to the ITV to form a PTV. The residual motion is mainly due to patient breathing irregularity. Some patients could have relatively more breathing irregularity, which would require a relative larger residual margin for PTV expansion. Intensity modulated radiotherapy (IMRT) or 3D conformal technique with 6–8 fields is used to treat the patient. The prescription dose is 12 Gy!4 fractions to the 95% dose line to cover the PTV. The organs at risk (lung, esophagus, and cord) are delineated based on the regular spiral CT image set. The dose calculation is also based on this image set.
Results Phantom data Fig. 2(a)–(d) show the sagittal images of the moving phantom with a 3 cm motion amplitude using (a) spiral CT, (b) axial CT without gating, (c) axial CT with gating at the end of expiration phase (middle point of image acquisition at point B in Fig. 1), and (d) spiral CT with no phantom motion, respectively. We note that for both the spiral CT and the axial CT without gating, the motion artifacts are very significant, and the two show a different pattern of artifacts.
180
A technique of using gated CT images to determine ITV
Fig. 2. Sagittal plane view of the CT images of the motion phantom (a) using regular spiral scan, (b) using axial scan without gating, (c) using axial scan with gating at the bottom phase, and (d) using regular spiral scan without phantom motion.
However, the gated CT image is almost identical to the images without motion. Gated CT scan taken at the end of inspiration phase showed similar image quality with no motion artifacts. This demonstrates that gated CT scan can be successfully performed in a regular single slice CT scanner with our gating system, even for a moving object with a motion-amplitude as larger as 3 cm. We have evaluated the ITV for each virtual tumor with the target motion amplitude being at (1) 1.1 cm, (2) 2.1 cm and (3) 3.0 cm in the longitudinal direction. Various ITVs were delineated by combining different gated CT image sets. We define ITV2 Z GTVP C GTVB ITV2:5 Z GTVP C GTVB C GTVS ITV3 Z GTVP C GTVB C GTVM1 ITV3:5 Z GTVP C GTVB C GTVM1 C GTVS ITV4 Z GTVP C GTVB C GTVM2 C GTVM3 ITV5 Z GTVP C GTVB C GTVM1 C GTVM2 C GTVM3 where GTVP, GTVB, GTVS, GTVM1, GTVM2 and GTVM3 are the delineated target volume in the peak phase, bottom phase, regular spiral CT, middle phase1, middle phase2 and middle phase3 CT image sets, respectively. A real ITV could be defined as the combination of GTVs from infinite number of gated CT images at different phases. We found that it could be reasonable approached by ITV5 for the conditions in our study. Table 1 shows these ITVs and their ratios to the real ITV of each virtual tumor with different object motion amplitudes. We note that when target motion equals 1.1 cm, even for the smallest target, the ITV determined by the image sets of the peak and the bottom phases reasonably represents the real ITV. When target motion equals 2.1 cm, ITV2 or even ITV2.5 are not sufficient to cover all the tumor excursions for most tumors, especially for small tumors. However, ITV3 and ITV3.5 seem to be sufficient to represent the real ITV. For target motion equals 3.0 cm, both ITV3 and ITV3.5 do not seem enough to cover all the tumor excursions, especially for small tumors. However, for large tumors, the difference
seems to be small (within 5%). We should point out that the volumes of several ITVs in the table are shown to be slightly larger than the volume of corresponding ITV5. This could be due to the uncertainty of drawing the ITV to encompass the GTVs and the limited spatial resolution (grid size) to define the targets.
Application in patients Gated-CT images have been acquired for 10 patients according to our IRB approved protocol. Table 2 lists the tumor location, GTV (spiral), ITV and tumor movement (tumor center to center) for these patients. A detailed analysis of the patient data will be reported after enough patients are accrued in the study. Here we present an example of how this technique is applied for patient treatment planning. The patient has a peripheral lesion in the left low lobe. Fig. 3(a)–(d) show the coronal and sagittal planes of the CT images of (a) gated at the expiration phase, (b) gated at the inspiration phase, (c) gated at a middle phase, and (d) a regular spiral CT, respectively, with the corresponding GTVs and ITV delineated. We note that the position difference of the inferior edge between the inspiration and expiration phases is about 1.6 cm. This is consistent with the 1.5 cm of inferior/superior (I/S) motion observed at fluoroscopy. The gated images at three different phases show the tumor in different locations related to the patient’s vertebrae. The expiration image shows almost no motion artifacts in the diaphragm, however, the motion artifacts are visible in the inspiration and the middle phases. Fig. 4 shows the coronal plane of the PET images of the patient with the contours of GTVs from the expiration phase (GTVex), the inspiration phase (GTVin) and the middle phase (GTVmid) superimposed. The combination of the three GTVs seems to be consistent with the tumor enhancement of the PET image both in shape and size. This gives us another confirmation that the ITV encompassing the three gated volumes plus the spiral CT would represent the tumor excursion. However, the diagnostic PET/CT images were taken with the patient lying in a slightly different position as compared to the simulation CT. The match of PET volume to the ITV is not perfect. In addition, the PET volume could change significantly for different threshold values. Currently
J.-Y. Jin et al. / Radiotherapy and Oncology 78 (2006) 177–184
181
Table 1 Volumes of different ITVs and their ratios to the corresponding real ITV for four different objects with tumor motions of 1.1, 2.2, and 3.0 cm. Target motion (cm)
Target volume
Object A Volume (cc)
MotionZ1.1
MotionZ2.2
MotionZ3.0
ITV2 ITV2.5 ITV3 ITV3.5 ITV4 ITV5 ITV2 ITV2.5 ITV3 ITV3.5 ITV4 ITV5 ITV2 ITV2.5 ITV3 ITV3.5 ITV4 ITV5
45.8 46.2 46.1 46.1 46.1 46 55 55.9 57.2 56.9 56.8 56.7 62.4 66 70.5 70.5 69.7 71.7
Object B Ratio 0.996 1.004 1.002 1.002 1.002 1.000 0.970 0.986 1.009 1.004 1.002 1.000 0.870 0.921 0.983 0.983 0.972 1.000
Volume (cc) 24.9 25.1 25.2 25 25.2 25.1 32.9 33.5 35.1 35 35.4 35.5 35.1 38.5 41.7 42.6 42.7 43.5
we only use PET volume as a method of verification. We have started to take PET/CT scans for some patients with simulation cradles. Because image guided radiotherapy is used in our dose delivery, the setup uncertainty is generally less that 3 mm [16]. Therefore, a 3 mm margin could be applied to the ITV to form the PTV. We used a 3 mm margin in the AP and lateral directions, and a 5 mm margin in the superior/ inferior directions to form the PTV for this patient. The 5 mm margin is to take into account possible residual motion during the gated CT imaging. As shown in Fig. 3, considerable motion artifact is still visible in the inspiration phase, suggesting that the image was not always acquired at the same inspiration phase due to breathing irregularity. Fig. 5 shows a coronal plane of a spiral CT image with PTVITV, PTV10 mm and PTVindiv drawn. The PTVITV is expanded from ITV with the margin mentioned earlier. The PTV10 mm is expanded 10 mm in all directions from the GTV of the spiral
Object C Ratio 0.992 1.000 1.004 0.996 1.004 1.000 0.927 0.944 0.989 0.986 0.997 1.000 0.807 0.885 0.959 0.979 0.982 1.000
Object D
Volume (cc)
Ratio
Volume (cc)
Ratio
25 24.7 25 24.8 25 25 31.2 31.5 32.6 32.5 32.5 32.4 34.6 36.5 38.7 39.2 39.8 40.1
1 0.988 1 0.992 1 1 0.963 0.972 1.006 1.003 1.003 1.000 0.863 0.910 0.965 0.978 0.993 1.000
3.08 3.09 3.25 3.08 3.14 3.13 3.19 3.59 4.35 4.31 4.38 4.35 3.1 4.05 5.62 5.78 6.53 6.78
0.984 0.987 1.038 0.984 1.003 1.000 0.733 0.825 1.000 0.991 1.007 1.000 0.457 0.597 0.829 0.853 0.963 1.000
CT, which is commonly used as a PTV margin at some institutions for lung patients without any motion information [8]. The PTVindiv is the individualized PTV, which is expanded from GTV of the spiral CT with the margin corresponding to the motion observed at fluoroscopy, which was commonly used at our institution for peripheral lesions prior to gated CT availability. In this case, because we observed on fluoroscopy that the target motion is about 1.5 cm in I/S direction, and less than 5 mm in the other directions, we use the 1.5 cm margin at I/S direction, and 5 mm margins for the AP and lateral direction. The PTV volumes are 24.2 CC, 41.9 CC, 34.8 CC for PTVITV, PTV10 mm and PTVindiv, respectively. In addition, we note that there is a fairly large part of PTVITV that is not included in the PTV10 mm, and a small part not included in the PTVindiv. These suggest that using PTVITV can not only reduce the irradiation volume, but more importantly, have less chance of missing the target.
Table 2 List of the tumor location, GTV (spiral), ITV and tumor movement (tumor center to center) for 10 patients enrolled in the study Patient
Location
GTV (cc)
ITV (cc)
Motion X (mm)
Motion Y (mm)
Motion Z (mm)
Motion overall
1 2 3 4 5 6 7 8 9 10
LLL RLL RLLa LLL LUL LUL Rt hilum RUL RUL RUL
52.45 11.66 4.55 5.21 4.71 43.88 6.48 18.73 32.02 5.22
89.01 26.1 5.16 10.09 7.25 50.79 9.86 30.34 51.71 7.97
3.2 K3.14 0.19 2.82 K1.34 K0.97 0.15 K1.42 K1.59 K1.73
K3.1 K3.75 0.16 3.88 K2.72 K0.28 K0.29 1.42 3.74 K1.67
15.56 13.68 1.07 15.02 3.2 K0.68 K0.85 3.88 3.9 3.19
16.2 14.5 1.1 15.8 4.4 1.2 0.9 4.4 5.6 4.0
a
For patient 3, the tumor is attached on the chest wall.
182
A technique of using gated CT images to determine ITV
Fig. 3. Coronal and sagittal plane view of the CT images of a patient at different gating phases with corresponding GTV and ITV delineated. (a) Gated at the expiration phase, (b) gated at the inspiration phase, (c) gated at a middle phase, and (d) a regular spiral scan.
J.-Y. Jin et al. / Radiotherapy and Oncology 78 (2006) 177–184
183
reasonably good to determine the ITV. This suggests that using two breath-hold CT sets with a free breathing spiral CT could be an alternative to determine the ITV. However, it should be kept in mind that the tumor positions of breathhold and free-breath at the end of inspiration or expiration phases could be different. In addition, the quality control of breathing hold solely based on patients’ cooperation should also be considered. The target motion in this phantom study is a simple back and forth motion that the motion trajectory from the bottom to the peak is the same as that of peak to bottom position. It has been reported that some tumors could move in a hysteretic trajectory during a respiratory cycle [3,4]. In that case, another middle phase at a different tumor motion direction should be used to determine the ITV.
Conclusions Fig. 4. Coronal plane view of the PET images of a patient with the contours of GTV of CT images at different phases superimposed.
Discussion We have developed a gated CT technique in regular single slice CT scanner. The phantom study demonstrated that using three gated CT image sets, one in inspiration phase, one in expiration phase, and one in a middle phase, is reasonably safe to determine the ITV for a target with movement less than 2 cm. The PTV from this ITV can have better coverage of the target and reduce the target volume. The phantom study results also demonstrated that for a target with motion less than 1.1 cm, using only the image sets of the inspiration and the expiration phases could be
We have developed a gated CT technique in regular single slice CT scanner. The phantom study demonstrated that using three gated CT image sets, one in inspiration phase, one in expiration phase, and one in a middle phase, is reasonably safe to determine the ITV for a target with movement less than 2 cm. The PTV from this ITV can have better coverage of the target and reduce the target volume.
* Corresponding author. Jian-Yue Jin Tel.: C1 313 916 0234; fax: C1 313 916 3235. E-mail address: [email protected] : Received 18 May 2005; received in revised form 28 November 2005; accepted 30 November 2005; available online 10 January 2006
References
Fig. 5. The contours of PTVITV, PTV10 mm and PTVindiv of the patient in a regular spiral CT image’s coronal view.
[1] Ross C, Hussey D, Pennington E. Analysis of movement of intrathoracic neoplasms using ultrafast computed tomography. Int J Radiat Oncol Biol Phys 1990;18:671–7. [2] Seppenwoolde Y, Shirato H, Kitamura K, et al. Precise and realtime measurement of 3D tumor motion in lung due to breathing and heartbeat, measured during radiotherapy. Int J Radiat Oncol Biol Phys 2002;53:822–34. [3] Shirato H, Seppenwoolde Y, Kitamura K, Onimura R, Shimizu S. Intrafractional tumor motion: lung and liver. Semin Radiat Oncol 2004;14:10–18. [4] Mageras GS, Pevsner A, Yorke ED, et al. Measurement of lung tumor motion using respiration-correlated CT. Int J Radiat Oncol Biol Phys 2004;60:933–41. [5] Plathow C, Fink C, Ley S, et al. Measurement of tumor diameter-dependent mobility of lung tumors by dynamic MRI. Radiother Oncol 2004;73:349–54. [6] International Commission on Radiation Units and Measurements. Prescribing, recording, and reporting photon beam therapy. Supplement to report 50. Report 62. Washington, DC: ICRU; 1999. [7] Purdy JA. Current ICRU definitions of volumes: limitations and future directions. Semin Radiat Oncol 2004;14:27–40. [8] Underberg RW, Lagerwaard FJ, Cuijpers JP, Slotman BJ, van Sornsen de Koste JR, Senan S. Four-dimensional CT scans for treatment planning in stereotactic radiotherapy for stage I lung cancer. Int J Radiat Oncol Biol Phys 2004;60:1283–90.
184
A technique of using gated CT images to determine ITV
[9] Lagerwaard FJ, Van Sornsen de Koste JR, Nijssen-Visser MR, et al. Multiple ‘slow’ CT scans for incorporating lung tumor mobility in radiotherapy planning. Int J Radiat Oncol Biol Phys 2001;51:932–7. [10] de Koste JR, Lagerwaard FJ, de Boer HC, Nijssen-Visser MR, Senan S. Are multiple CT scans required for planning curative radiotherapy in lung tumors of the lower lobe? Int J Radiat Oncol Biol Phys 2003;55:1394–9. [11] Balter JM, Ten Haken RK, Lawrence TS, et al. Uncertainties in CT-based radiation therapy treatment planning associated with patient breathing. Int J Radiat Oncol Biol Phys 1996;36:167–74. [12] Shih HA, Jiang SB, Aljarrah KM, Doppke KP, Choi NC. Internal target volume determined with expansion margins beyond composite gross tumor volume in three-dimensional conformal radiotherapy for lung cancer. Int J Radiat Oncol Biol Phys 2004;60:613–22.
[13] Caldwell CB, Mah K, Skinner M, Danjoux CE. Can PET provide the 3D extent of tumor motion for individualized internal target volumes? A phantom study of the limitations of CT and the promise of PET. Int J Radiat Oncol Biol Phys 2003;55: 1381–93. [14] Lavrenkov K, Partridge M, Cook G, Brada M. Positron emission tomography for target volume definition in the treatment of non-small cell lung cancer. Radiother Oncol 2005;77:1–4. [15] Jin JY, Yin FF. Time delay measurement for linac based treatment delivery in synchronized respiratory gating radiotherapy. Med Phys 2005;32:1293–6. [16] Yin FF, Ryu S, Ajlouni M, et al. A technique of intensitymodulated radiosurgery (IMRS) for spinal tumors. Med Phys 2002;29:2815–22.