First published online December 6, 2006
British Journal of Radiology (2007) 80, 355-361
© 2007 British Institute of Radiology
doi: 10.1259/bjr/31008031
Interbreath-hold reproducibility of lung tumour position and reduction of the internal target volume using a voluntary breath-hold method with spirometer during stereotactic radiotherapy for lung tumours
T Kimura, MD, PhD
1,2
Y Murakami, MD
1
M Kenjo, MD
1
Y Kaneyasu, MD, PhD
1
K Wadasaki, MD, PhD
1
K Ito, MD, PhD
1 and
M Ohkawa, MD, PhD
2
1 Department of Radiology, Hiroshima University Graduate School of Medicine, Hiroshima, 2 Department of Radiology, Kagawa University School of Medicine Kagawa, Japan
Correspondence: Tomoki Kimura, Department of Radiology, Kagawa University School of Medicine, 1750-1 Ikenobe, Miki-cho, Kida-gun, Kagawa, 761-0793, Japan. E-mail: tkkimura{at}med.kagawa-u.ac.jp
 |
Abstract
|
|---|
The purpose of this study was to evaluate the interbreath-hold reproducibility of the tumour (gross tumour volume, GTV) position and relative reduction of the internal target volume (ITV) using a voluntary breath-hold method with a spirometer in a clinical setting of stereotactic radiotherapy (SRT) for lung tumours
11 patients with 14 lung tumours were enrolled in this study. CT scans were performed once at the free breathing phase and five times at the breath holding phase before the first treatment day. Patients held their breath at the end-expiration phase under spirometer-based monitoring. All GTVs were delineated by a physician and the GTV centroid was calculated automatically. To evaluate the interbreath-hold reproducibility of the tumour position, we measured the distance of three dimensions (craniocaudal, CC; left–right, LR; anteroposterior, AP) and vectors between the GTV centroid and bony landmark. The reproducibility was defined as the average of the differences between the GTV centroid and bony landmark from the second to fifth CT scans with regard to that from the first CT scans. We also evaluated the relative reduction of ITV between the free breathing and breath-holding phase. The interbreath-hold reproducibility of the tumour position was 1.3±1.3 mm, 1.4±1.8 mm, 2.1±1.6 mm and 3.3±2.2 mm in CC, LR and AP directions and vectors, respectively. ITV at the breath-holding phase was significantly smaller than that at the free breathing phase (P<0.01). In conclusion, the voluntary breath-hold method with a spirometer is feasible, with relatively good reproducibility of the tumour position for SRT in the clinical setting.
|
Introduction
|
|---|
Stereotactic radiotherapy (SRT) in the treatment of lung tumours has been shown to have good results and low morbidity [1, 2], and has been supported by various methods using immobilization devices [3] or coordinating systems for respiratory motion. To reduce respiratory motion during SRT for lung or liver tumours and improve feasibility for elderly patients or patients with pulmonary dysfunction, we have developed a voluntary breath-hold method using spirometer-based monitoring and have verified the relatively good reproducibility of organ positioning, especially diaphragm motion, in normal volunteers [4]. However, the diaphragm position does not necessarily reflect the lung tumour position directly, especially for tumours in the lower lobe, due to the complexity of tumour motion in a clinical setting [5]. From January 2003, we have applied this method to SRT for lung tumours clinically and directly verified the interbreath-hold reproducibility of tumour (the gross tumour volume, GTV) position using CT scans before the treatment.
The purpose of this study was to evaluate the interbreath-hold reproducibility of the tumour position and relative reduction of the internal target volume (ITV) between the free-breathing phase and breath-holding phase using this method in a clinical setting of SRT for lung tumours.
|
Methods and materials
|
|---|
Patient background
In this study, 11 patients with 14 lung tumours who were treated with SRT in Hiroshima University Hospital from March 2004 to February 2005 were enrolled (Table 1
). All tumours were delivered 12 Gy per fraction at the isocenter, and the total dose was 48 Gy with four fractions, with a voluntary breath-hold method using spirometer-based monitoring. The median age was 73 years (range from 57 years to 82 years). Nine patients were male and two were female. Seven patients were diagnosed with primary lung cancers and three patients had double primary lung cancers. Four patients were diagnosed with metastatic lung cancers, of which the primary tumours were lung cancer in two patients, urethral cancer in one patient and oesophageal cancer in one patient. The median tumour size was 22 mm (ranging 7 mm to 58 mm). Tumour location was as follows: two tumours in the right upper lobe (RUL), two tumours in the right middle lobe (RML), four tumours in the right lower lobe (RLL), two tumours in the left upper lobe (LUL), and four tumours in the left lower lobe (LLL). All tumours were peripheral lung tumours and were not adjacent to other organs, such as the chest wall or heart.
The breath-hold method using spirometer-based monitoring
We described the procedure of the voluntary breath-hold method using a spirometer in a previously published paper [4], and have since improved the spirometer system. Although only the waveform was displayed and we evaluated only visually in the first type, we could evaluate quantitively in the advanced type. To monitor the respiratory phase and breath-holding phase, the patient breathed through a mouthpiece connected to a pneumotachometer (VMM-400; Interface Associates, Aliso Viejo, CA). The other end of the pneumotachometer was attached to a three-way connector and 5 l min–1 of oxygen were inhaled through one tube of the three-way connector to assist the breath hold. A nose clip was used to prevent nasal breathing and ensure that the volunteer breathed through the mouthpiece. For patient set-up, we used a vacuum cushion (Vac-Lok; MEDTEC, Orange City, IA) (Figure 1). The workstation ran a custom application developed using virtual instrument software (LabVIEW, National Instruments, Austin, TX) that integrated the spirometry signal, yielded the respiratory tidal volume of the patient and was able to display a flow–time curve which can show the state of inspiration, expiration and breath-hold (Figure 2). To familiarize patients with the procedure, we gave them training sessions prior to CT scans. In our previous study in which we used healthy volunteers, the end-expiration phase had better interbreath-hold reproducibility of organ position than the end-inspiration phase, and patients therefore held their breath in the end-expiration phase [4].
View larger version (76K):
[in this window]
[in a new window]
|
Figure 1. A volunteer patient breathed through a mouthpiece connected to the gas monitoring sensor. A nose clip was used to prevent nasal breathing and ensure that the patient breathed through the mouthpiece. For patient set-up, we used a vacuum cushion (Vac-Lok, MEDTEC, Orange City, IA).
|
|
View larger version (58K):
[in this window]
[in a new window]
|
Figure 2. The workstation ran a custom application, developed using virtual instrument software(LabVIEW, National Instruments, Austin, TX) that integrated the spirometry signal, yielded the respiratory tidal volume of the patient and was able to display a flow-time curve which can show the state of inspiration, expiration and breath-hold.
|
|
We instructed them to keep a stable tidal volume at the end-expiration phase where they were able to hold their breath comfortably under inhalation of oxygen. The reproducibility of the manoeuvre as determined by the spirometry level was carefully monitored, and they repeated this manoeuvre three to four times until they became familiarized.
CT scan procedure
CT scans (Lightspeed QX/I; GE Yokogawa Medical Systems, Tokyo, Japan) were performed at the free-breathing phase and the breath-holding (the end-expiratory) phase. CT scans at the free-breathing phase were first performed once with an extended scan (4 s per scan) to calculate ITV on the day of treatment planning. CT scans at the breath-holding phase were performed three times on the day of treatment planning and twice on the first treatment day before SRT for treatment planning and verification (total five times). The first CT scan at the breath-holding phase was used for treatment planning and the other four CT scans were used for verification and evaluation of interbreath-hold reproducibility of the tumour position. Set-up was also performed at the treatment position. Slice thickness and interval were 1.25 mm and 1.25 mm, respectively. CT volume data were transferred to a three-dimensional (3D) treatment planning system (Pinnacle3 version 6.0; ADAC, Milipitas, CA).
Target delineation
Target delineation was performed on a 3D treatment planning system. A physician delineated the target volume on the axial CT slices using lung CT window settings (window width 1600 Hounsfield units (HU) and window level –300 HU). Spiculation and pleural indentation were included within the target. The partial volume effects which were observed on CT scans at the free-breathing phase were also included within the target. (Figure 3)
View larger version (95K):
[in this window]
[in a new window]
|
Figure 3. All gross tumour volumes were delineated by a physician on each axial image in a three-dimensional (3D) treatment planning system. The CT window and level was 1600 Hounsfield Units (HU) and –300 HU during each contouring session.
|
|
Data analysis
We analyzed CT scans at the breath-holding phase which were performed five times to evaluate the interbreath-hold reproducibility of the tumour position. The first CT scan at the breath-holding phase was used for treatment planning in the clinical setting. The same physician delineated GTV on the other four CT scans at the same CT setting. The potential for intraobserver variation exists and may contribute to some variation in the geometric center of GTV (i.e. GTV centroid) so we therefore calculated the average volume and standard deviation (SD) of each patient to evaluate the intraobserver variation of GTV delineation.
To calculate the interbreath-hold reproducibility of the tumour position, we measured the distance between the GTV centroid and bony landmark in three dimensions (CC, craniocaudal; LR, left-right; AP, anteroposterior) and the vector. The GTV centroid was calculated automatically using a 3D treatment planning system. "Bony landmark" was defined as the intersection of the frontal line and lateral line with the upper border of T10 on the axial CT slice. The reproducibility was defined as the average of the differences between the GTV centroid and bony landmark from the second to fifth CT scans with regard to that from the first CT scans at the breath-holding phase. We also evaluated the difference of the interbreath-hold reproducibility of the tumour position by location (upper lobe vs lower lobe). This methodology was chosen to eliminate any external set-up errors, and thus all tumour displacement would be a measure of internal motion only.
CT scans at the free-breathing phase by extended scan length (4 s per scan) included respiratory motion of the tumour in CT images, and we therefore evaluated the relative reduction of ITV between the free-breathing and the breath-holding (the end-expiratory) phase using this method in a clinical setting of SRT for lung tumours. CT scans at the breath-holding phase were performed five times and we used the first CT scans at the breath-holding phase for treatment planning clinically, so we therefore compared ITV (equal to GTV) with CT scans at the free-breathing phase and the first CT scans at the breath-holding phase. We also evaluated the difference in the relative reduction of ITV (equal to GTV) according to tumour location (upper lobe vs lower lobe).
From these data, geometric uncertainties were determined. Geometric uncertainties in radiotherapy consist of internal organ motion and external set-up deviations. Both deviations consist of a systematic component, i.e. the same for each fraction of the treatment, as well as a random component, i.e. varying from day to day [6, 7]. In this study, we calculated geometric uncertainties of tumour motion. The systematic deviations were calculated by determining the spread (one SD) in the individual means of the differences between the first CT scan and each subsequent scan. The random deviations were calculated by the spread (one SD) of these differences around the corresponding mean in each patient and subsequently calculating the average of these SDs for the whole group [8].
Statistical significance in the differences was determined using the Student's t-test. Statistical significance was established at the level of p<0.05.
|
Results
|
|---|
Intraobserver variation of GTV delineation
Table 2 shows the average volume and SD of each patient's contoured tumour over the sequential CT scans with a voluntary breath-hold method using a spirometer.
Interbreath-hold reproducibility of tumour position
Table 3 shows the interbreath-hold reproducibility of the tumour position, which was 1.3±1.3 mm, 1.4±1.8 mm, 2.1±1.6 mm and 3.3±2.2 mm in CC, LR and AP directions and vectors, respectively.
Figures 4 and 5 show the scatterplots of the GTV centroids in the CC vs LR directions and AP vs LR directions. The distribution ranged from 0.0 mm to 5.7 mm, 0.0 mm to 8.8 mm, 0.0 mm to 8.6 mm and 0.5 mm to 10.0 mm in CC, LR and AP directions and vectors, respectively.
View larger version (24K):
[in this window]
[in a new window]
|
Figure 4. The scatter-plots of the gross target volume centroids in the craniocaudal (CC) vs left–right (LR) directions.
|
|
View larger version (24K):
[in this window]
[in a new window]
|
Figure 5. The scatter-plots of the gross target volume centroids in the anteroposterior (AP) vs left–right (LR) directions.
|
|
There was no significant difference in the interbreath-hold reproducibility in vector according to the tumour location (upper lobe vs lower lobe, p = 0.623).
Table 4 shows the results of geometric uncertainties, systematic and random deviations of tumour motion.
Relative reduction of ITV between the free-breathing and the breath-holding phase
Table 5 shows the results of ITV at the free-breathing and breath-holding phases, and the relative reduction between both phases. ITV (equal to GTV) at the breath-holding phase (using the first CT scans) was significantly smaller than that at the free breathing phase (p<0.01). There was a significant difference in the relative reduction of ITV according to tumour location (upper lobe vs lower lobe, p = 0.023).
|
Discussion
|
|---|
Several approaches have been used in the breath-hold method and relatively good reproducibility of the tumour position has been reported in each method, as follows. Cheung et al [9]. reported that the interfraction reproducibility of the tumour position with the active breath control (ABC) device, which temporary immobilizes the patient's breathing, and the average (±SD) displacement of GTV centres was 0.3 mm (±1.8 mm), 1.2 mm (±2.3 mm), and 1.1 mm (±3.5 mm) in LR, AP and CC directions, respectively. Although these results included set-up error, the results were good. Mah et al [10] reported that the interbreath-hold reproducibility of the tumour position with the deep inspiration breath-hold (DIBH) technique, which is able to displace normal lung or heart from the high-dose treatment field using a commercially based spirometer, and the inferred displacement of the centroid GTV from its position in the planning scan, as calculated from the spirometer records in over 350 breath-holds, was 0.2±1.4 mm (mean and SD). Onishi et al [11] demonstrated that the interbreath-hold reproducibility of the tumour position during 20 lung cancer patients' self-estimated breath-holding at the inspiration phase using fusion of CT and linear accelorator (FOCAL) unit without respiratory monitoring was 2.2 mm in the CC direction, 1.4 mm in the AP direction and 1.3 mm in the LR direction.
In our study, the interbreath-hold reproducibility of the tumour position was 1.3±1.3 mm, 1.4±1.8 mm, 2.1±1.6 mm and 3.3±2.2 mm in CC, LR and AP directions and vectors, respectively. These results were also good and similar to the results of other reports with breath-hold methods. Although these results included uncertainties in GTV delineation due to the potential for intraobserver variation and may contribute to some variation in the calculated GTV centroids, we believe that the intraobserver reproducibility of GTV delineation was relatively good in this study and that this method of using CT-based measurements to determine the tumour position is more accurate and reliable than using fluoroscopic images or portal images. We also evaluated the difference between the interbreath-hold reproducibility in vector according to the tumour location, by upper lobe vs lower lobe, with no significant difference noted by tumour location (p = 0.257). Seppenwoolde et al [5] demonstrated that the trajectory of the tumour during inhalation is different from the trajectory during exhalation, i.e. hysteresis, by analyzing the 3D motion of lung tumours during radiotherapy using a real-time tumour tracking system, and suggested the complexity of tumour motion, especially that of tumours in the lower lobe. Our study suggested that our method would be effective for temporary immobilization of respiratory motion even in the lower lobe.
We also evaluated the distribution of GTV centroids, which ranged from 0.5 mm to 10 mm in vector. This large distribution would be a disadvantage in our method in comparison with other breath-hold methods, and we therefore have to pay more attention to the training sessions and monitoring of respiratory motion during treatment. In contrast, the advantage of our method is feasibility in many patients and adaptability to many institutions because patients are able to hold their breath more comfortably at the end-inspiration phase under inhalation of oxygen. The other breath-hold methods such as ABC and DIBH may be especially demanding of elderly patients or those with pulmonary dysfunction, leading to concerns of lower feasibility. We consider this method to have higher feasibility for elderly patients or patients with pulmonary dysfunction.
We calculated geometric uncertainties of the tumour position from our results. There have been many reports focused on the geometric uncertainties of the set-up [12–15]; however, there are few reports focused on the geometric uncertainties of internal organs or tumour motion. Stoom et al [6] calculated the geometric uncertainties of lung cancer patients. In this study, systematic deviations were 2 mm, 3 mm and 3 mm in LR, AP and CC directions, respectively, and random deviations were 4 mm, 5 mm and 5 mm in LR, AP and CC directions, respectively. Due to breathing and cardiac motion, their results were not as accurate as our results in which systematic and random deviations in all directions were within 1.5 mm. According to the British Institute of Radiology report [16], various factors participate in geometric uncertainties and a margin is determined in radiation therapy for the lung. Systematic deviations have been divided into gaussian and linear deviations. Of the gaussian systematic deviations, there are the clinician's target delineation deviations, phantom transfer deviations and set-up deviations, and so on. Of the linear systematic deviations, the breathing positional deviations (b) can be within a large range of values. In the breath-hold method, b is zero, and therefore it is one of the advantages of our method. Random deviations have included set-up and penumbral deviations. The systematic motion deviations, which are due to representation of target position, shape and size in the CT study; random motion deviations, which are daily errors of target position; and shape and size for the group of patients, are some of the various factors of geometric uncertainties, as mentioned above. In this study, we only evaluated the result of systematic and random motion deviations and, since other factors such as set-up systematic motion deviations were not examined, a GTV-PTV (planning target volume) margin could not be calculated concretely here. However, it was considered meaningful to have obtained the reproducibility of the tumour position (x, y, z, vectors) as well as systematic and random motion deviations by our breath-hold method, and we believed the internal margin of the interbreath-hold reproducibility of lung tumour position would be within almost 5 mm in vector for the treatment planning of SRT due to the relatively good reproducibility of our method. From now on, based on the results of this study, we will examine other factors of geometric uncertainties and finally calculate a GTV-PTV margin in our method.
We also evaluated the relative reduction of ITV between the free breathing and breath-holding at the end-expiratory phase, which was significantly smaller at the breath-holding phase than at the free breathing phase. According to tumour location, upper or lower lobe, the relative reduction of ITV in the upper lobe was significantly less than the lower lobe in our study. Mah et al [10] also reported that total lung volumes increased by an average of 42%, which resulted in an average decrease in lung mass of 18% within a standard 1.5 cm PTV margin around the GTV using the ABC inspiration breath-hold. PTV is increased by the safety margin due to respiratory motion and would cause toxicities, especially radiation pneumonitis, after SRT. We believe that the reduction of ITV by controlling respiratory motion using our method will lead to a reduction in the severity of these toxicities. The results of Table 5
may exaggerate the reduction of target volume between the free breathing and breath-holding phases because we neglected the reproducibility of the tumour position at the breath-holding phase. But, if this reproducibility is improved, the result shown in Table 5 may be approached. For this reason, we have to try to improve this technology further.
|
Conclusions
|
|---|
The voluntary breath-hold method using spirometer-based monitoring was feasible, with relatively good reproducibility and geometric uncertainties of the tumour position for SRT in a clinical setting without regard to tumour location. However, we need to improve the large distribution of the tumour position by improving the training sessions for patients and monitoring of respiratory motion during treatment. This method can also significantly reduce ITV compared to the free breathing phase and can lead to a reduction of severe toxicities. We plan to collect further cases using this method in SRT for lung or liver tumours.
This work was partly presented at the 90th RSNA Annual Meeting, Chicago, Illinois, 28 November–3 December 2004.
Received for publication January 12, 2006.
Revision received July 12, 2006.
Accepted for publication August 18, 2006.
|
References
|
|---|
- Onishi H, Araki T, Shirato H, Nagata Y, Hiraoka M, Gomi K, et al. Stereotactic hypofractionated high-dose irradiation for stage I non-small cell lung carcinoma: clinical outcomes in 245 subjects in a Japanese multi-institutional study: cancer 2004;101:1623–31.[CrossRef][Medline]
- Uematsu M, Shioda A, Suda A, Fukui T, Ozcki Y, Hama Y, et al. Computed tomography-guided frameless stereotactic radiotherapy for stage I non-small-cell lung cancer: a 5-year experience. Int J Radiat Oncol Biol Phys 2001;51:666–70.[CrossRef][Medline]
- Negoro Y, Nagata Y, Aoki T, Mizowaki T, Araki N, Takayami K, et al. The effectiveness of an immobilization device in conformal radiotherapy for lung tumor: reduction of respiratory tumor movement and evaluation of the daily setup accuracy. Int J Radiat Oncol Biol Phys 2001;50:889–98.[CrossRef][Medline]
- Kimura T, Hirokawa Y, Murakami Y, Tsujimura M, Nakashima T, Ohno Y, et al. Reproducibility of organ position using voluntary breath-hold method with spirometer for extracranial stereotactic radiotherapy. Int J Radiat Oncol Biol Phys 2004;60:1307–13.[CrossRef][Medline]
- Seppenwoolde Y, Shirato H, Kitamura K, Shimizu S, Herk M, Lebesque JV, et al. Precise and real-time measurement of 3D tumour motion in lung due to breathing and heartbeat, measured during radiotherapy. Int J Radiat Oncol Biol Phys 2002;53:822–34.[CrossRef][Medline]
- Stroom JC, Boer HCJ, Huizehga H, Visset AG. Inclusion of geometrical uncertainties in radiotherapy treatment planning by means of coverage probability. Int J Radiat Oncol Biol Phys 1999;43:905–19.[CrossRef][Medline]
- Kutcher GJ, Mageras GS, Leibel SA. Control, correction, and modeling of setup errors and organ motion. Seminars in Radiation Oncology 1995;5:134–45.[CrossRef][Medline]
- Samson MJ, Koste JRS, Boer HCJ, Tankink H, Verstraate M, Essers M, et al. An analysis of anatomic landmark mobility and setup deviations in radiotherapy for lung cancer. Int J Radiat Oncol Biol Phys 1999;43:827–32.[CrossRef][Medline]
- Cheung PCF, Sixel KC, Tirona R, Ung RC. Reproducibility of lung tumor position and reduction of lung mass within the planning target volume using active breathing control (ABC). Int J Radiat Oncol Biol Phys 2003;23:1437–42.
- Mah D, Hanley J, Rosenzweig KE, Yorke E, Braban L, Ling CC, et al. Technical aspects of the deep inspiration breath-hold technique in the treatment of thoracic cancer. Int J Radiat Oncol Biol Phys 2000;48:1175–85.[CrossRef][Medline]
- Onishi H, Kuriyama K, Komiyama T, Tanaka S, Sano N, Aikawa Y, et al. A new irradiation system for lung cancer combining linear accelerator, computed tomography, patient self-breath-holding, and patient-directed beam-control without respiratory monitoring devices. Int J Radiat Oncol Biol Phys 2003;56:14–20.[CrossRef][Medline]
- Bijhold J, Lebesque JV, Hart AAM, Vijlbrief RE. Maximizing setup accuracy using portal images as applied to a conformal boost technique for prostatic cancer. Radiothear Oncol 1992;24:261–71.[CrossRef]
- Onimaru R, Shirato H, Aoyama H, Kitamura K, Seki T, Hida K, et al. Calculation of rotational setup error using the real-time tracking radiation therapy (RTRT) system and its application to the treatment of spinal schwannoma. Int J Radiat Oncol Biol Phys 2002;54:939–47.[CrossRef][Medline]
- Boer HCJ, Koste JRS, Senan S, Visset AG, Heijmen BJM. Analysis and reduction of 3D systematic and random setup errors during the simulation and treatment of lung cancer patients with CT-based external beam radiotherapy dose planning. Int J Radiat Oncol Biol Phys 2001;49:857–68.[CrossRef][Medline]
- Samuelsson A, Mercke C, Johansson KA. Systematic set-up errors for IMRT in the head and neck region: effect on dose distribution. Radiothear Oncol 2003;66:303–11.[CrossRef]
- Bidmead M, Coffey M, Crellin A, Dobbs J, Driver D, Greener T, et al. Geometric uncertainties in radiotherapy: defining the planning target volume. London, UK: British Institute of Radiology, 2003
Top
Abstract
Introduction
