British Journal of Radiology 74 (2001),73-76 © 2001 The British Institute of Radiology
Variation in size and position of the planning target volume in the transverse plane owing to respiratory movement during radiotherapy to the lung
S Essapen, MRCP, FRCR
C Knowles
and
D Tait, MD, FRCR
Department of Radiotherapy and Oncology, Royal Marsden NHS Trust, Sutton, Surrey SM2 5PT, UK
 |
Abstract
|
|---|
Movement of thoracic tumours with respiration poses a real dilemma in terms of the accuracy of delivering radical radiotherapy in patients with carcinoma of the lung. Movements in the craniocaudal direction have previously been described. This technical note describes ten patients planned for radical lung radiotherapy using CT. The study assesses the maximum impact of respiration on the planning target volume in the transverse plane by comparing the planning CT appearances during quiet respiration with those during full inspiration and full expiration. The study demonstrated the potential impact of respiratory movement on the planning target volume and, hence, implications for local tumour control.
|
Introduction
|
|---|
As more complex radiotherapy techniques are devised to improve the therapeutic ratio, the accuracy with which radiation is delivered becomes increasingly important. Quantifying this accuracy is difficult, as a number of factors are involved such as patient set-up error, organ movement and geometric penumbra of the X-ray beam. Movement of the internal tumour relative to external skin markings is another factor thatmust be considered when defining the planning target volume (PTV) in radiotherapy planning [1].
Use of CT for planning radiotherapy for carcinoma of the bronchus makes no allowance for movement of the internal anatomy of the thorax during respiration. Patients are CT scanned and simulated during quiet respiration. The PTV is then localized on sequential CT images, which record the anatomy of the thorax and tumour at a particular split second in time, with no reference to the stage of the respiratory cycle when the CT image was acquired. It is therefore possible that at a different stage in the respiratory cycle, the tumour may vary in its position within the thorax, the net effect being that the sequence of CT images may not represent thoracic anatomy or the position of the tumour relative to other thoracic structures. This phenomenon is eloquently described by Balter et al [2], and has significant implications for radiotherapy planning of lung cancers.
In radiotherapy planning for carcinoma of the bronchus, the PTV is generally defined as the gross tumour volume (GTV) plus a 2 cm margin. This margin is given to allow for the uncertainties listed above, and also includes any potential microscopic disease beyond the GTV [1]. To prevent geographical miss of the intended GTV, it is essential to have some idea of the magnitude of movement of the GTV within the chest during respiration. To accommodate any craniocaudal movement of the GTV during respiration, a 2 cm margin in the superoinferior axis is thought to be adequate to ensure that the clinical target volume (CTV) remains in the field during respiration. However, making allowances for craniocaudal movement of a thoracic tumour is not the only consideration. With respiration, there is a change in the position of the GTV in the transverse plane within the thorax, and this may be more difficult to predict.
The aim of this study was to use CT to determine the amount by which the transverse PTV would have to be increased to adequately cover disease when comparing normal respiration with the extreme cases of suspended full inspiration and full expiration.
|
Methods and materials
|
|---|
Planning CT scanning
Ten patients were scanned during quiet respiration using a Siemens Somatom Plus 40 CT scanner (Siemens, Erlangen, Germany). Parameters set were as follows: an image zoom factor of 1.0 and a normal scan time of 1 s per slice. Patient position was the same as for treatment, that is supine, with the arms supported above the head in a band. The normal concave soft scanner couch top was replaced by a hard flat couch top, which was more representative of the simulator and treatment couch tops, thus allowing reproducibility of set-up.
The area to be scanned was pre-determined from the diagnostic CT scans of the patient, and the start and end positions of the radiotherapy planning scans were determined using the diagnostic anteroposterior topograms. Sequential CT scanning was performed at 10 mm intervals giving a slice thickness for images of 8 mm during quiet respiration. It took approximately 5 min to complete a typical scan sequence.
During the radiotherapy planning CT scan, a scan level displaying the bulk of disease was selected, and the table position of this image was noted. This was taken as the index planning slice. After completion of the standard procedure, the couch was returned to the table position of the selected index planning slice. Suspended respiration scans at full inspiration and full expiration were performed at this level using the same set-up (i.e. scan time, zoom) factors as before.
Standard localization
All CT images were transferred by magnetic tape to an IGEMS Target Series 2 radiotherapy planning system (Prism Microsystems, Slough, UK). The PTV was localized on each CT slice in the region of interest, and defined as a 2 cm margin around the GTV in the transverse plane, and the same margin in the superoinferior axis. Using in-house software, a minimum planar outline was obtained, which encompassed all the individual PTVs at each CT slice. This final PTV, called the composite volume (CVL), was displayed on the index planning slice, and the CVL on this slice was covered by the 95% isodose. The radiotherapy dose was prescribed to the 95% isodose.
Adjustment of CVL to accommodate respiratory excursion
The CVL on the index planning slice was transferred directly to the suspended respiration slices using the planning computer. Each of these slices was then checked to assess whether the GTV was adequately covered by the pre-defined CVL. If the CVL was not adequate, any extra margin needed was defined, giving a new CVL. No reductions in the CVL were performed, and the orientation of any increase in the CVL was also noted.
Using in-house software, the new CVLs at suspended full inspiration (CVL2) and at full expiration (CVL3) were documented. Because the 95% isodose curve in our radiotherapy planning encompassed the proposed PTV (i.e. the CVL), the difference between CVL and CVL2 (known as ICVFI) indicated how much of the PTV was not covered by the high dose volume at full inspiration. Likewise, the difference between CVL and CVL3 (known as ICVFE) indicated how much of the PTV was not covered by the high dose volume at full expiration.
|
Results
|
|---|
Table 1
shows any increase in the CVL at full inspiration (ICVFI) and at full expiration (ICVFE). No correlation was seen between the phase of respiration or the tumour position within the thorax and its direction of movement during full inspiration or expiration.
View this table:
[in this window]
[in a new window]
|
Table 1. Composite planning volumes during quiet respiration (CVL), and at full inspiration (CVL2) and full expiration (CVL3)
|
|
Although four of the cases planned did not require a change in the PTV to accommodate full respiration, the remaining six cases had a demonstrable variation in CVL2, CVL3, or both. The difference between quiet breathing and full inspiration CVLs ranged from 0–29%, whilst the change in the CVL with full expiration ranged from 0–24%.
Figure 1a illustrates the index CT slice of one patient showing the CVL at quiet respiration. The 95% isodose would normally encompass this volume in the treatment plan. Figure 1b and Figure 1c show the increased CVL required at full inspiration (CVL2) and full expiration (CVL3), respectively, to adequately cover the GTV. The smaller original CVLs at quiet respiration are shown for comparison.
View larger version (77K):
[in this window]
[in a new window]
|
Figure 1. (a) Index CT slice in Patient 5 showing the planning composite volume at quiet respiration. (b,c) The increased composite volume required at suspended full inspiration (CVL2) (b) and suspended full expiration (CVL3) (c).
|
|
|
Discussion
|
|---|
The changes that can occur in the CVL with deep inspiration or expiration may have a significant bearing on treatment planning. Under normal circumstances, the planning CT scan is performed with the patient breathing quietly, and the images are acquired at random during the patient's respiratory cycle. Adding a 2 cm margin superior and inferior to the GTV accommodates any craniocaudal movement of the GTV during respiration. This margin is deemed sufficient to ensure adequately that the CTV remains in the field during quiet respiration, and is supported by the study by Ekberg et al [3] who used fluoroscopy in 20 patients to assess the movement of their lung tumours caused by respiration. They found that the average movement of the CTV was 3.9 mm in the craniocaudal direction with quiet respiration, with a range of 0–12 mm, and reported that they routinely use a margin of 15 mm cranially and caudally.
Changes with respiration in the position of the GTV in the transverse plane may be more difficult to predict. This study highlights the potential error of missing part of the PTV during radiotherapy should the patient's breathing pattern differ significantly from that at the time of the planning CT scan. The case illustrated by Figures 1a–c
had the greatest discrepancy between the CVL and both CVL2 and CVL3. Although not representative of the results overall, it is nevertheless of some concern that six out of the ten patients studied required an increase in their CVL to accommodate deep respiratory movements. Patients with lung cancer often have compromised respiratory function, with an unpredictable respiratory pattern. Failure to recognize this fact in radiotherapy planning will inevitably mean that the PTV may not receive the prescribed dose.
The inability to guarantee the reproducibility of a patient's breathing during a prolonged course of radical radiotherapy for lung cancer produces uncertainties in the accuracy of their treatment. There could be a significant impact on the relative position of critical structures, such as the spinal cord, to the PTV. Although the planned treatment, based on conditions of quiet respiration, may indicate the spinal cord to be outside the high dose area, deeper respiration may move the cord into the radiation field. Alternatively, such movements during respiration may just lead to greater volumes of normal lung tissue being irradiated, which may be important if the patient's lung function is already compromised. Improvements in the accuracy and delivery of lung radiotherapy relative to skin marks or immobilization shells will not reduce these potential errors.
The GTV–PTV margin perhaps has more important implications where conformal radiotherapy is used. Here, the treated volume is closely conformed to the PTV using customized blocks or a multileaf collimator. Any change in the PTV at a particular position in the patient can lead to part of the CTV moving into a region of much lower dose, which may compromise local tumour control. Conformal radiotherapy can also minimize the normal tissue being irradiated if the GTV–PTV margin is increased to accommodate significant respiratory movements.
Ohara et al [4] looked at a respiratory gating technique whereby the linear accelerator ejected X-rays intermittently in response to a predetermined gate signal. This was taken as the point during respiration when the tumour was found to be static. They found that by using this method, the treatment volume in the seven patients studied could be reduced. This method would obviously prolong the duration of a patient's daily treatment, with repercussions on any busy radiotherapy unit. It would also require a greater pre-treatment assessment of a patient's breathing cycle.
A more practical approach to reducing the chance of geographical miss of the GTV would involve individualizing radiotherapy planning. An assessment of the maximum craniocaudal and anteroposterior movement of the GTV within the thorax during quiet and deep respiration could be made on a simulator using fluoroscopy. This would indicate the minimal planning margin required in these planes. Likewise, a similar method to that described in this study would quantify any movement of the GTV in the transverse plane and allow all of this information to be used when defining the PTV.
This study demonstrated the unpredictable impact of respiratory movement on the PTV and hence the potentially detrimental effect on local control. It should raise awareness of this pitfall and source of error in terms of tumour dose and local control. Larger numbers of patients are needed to establish whether this aspect of movement should be taken into consideration on an individual basis or consumed into the initial PTV definition.
Received for publication March 3, 1999.
Revision received August 29, 2000.
Accepted for publication October 16, 2000.
|
References
|
|---|
-
International Commission on Radiation Units and Measurements, Prescribing, recording and reporting photon beam therapy, Report No. 50. Bethseda, MD: ICRU, 1993.
-
Balter JM, Ten Haken RK, Lawrence TS, Lam KL, Robertson JM. Uncertainties in CT-based radiation therapy treatment planning associated with patient breathing. Int J Radiat Oncol Biol Phys 1996;36:167–74.[Medline]
-
Ekberg L, Holmberg O, Wittgren L, Bjelkengren G, Landberg T. What margins should be added to the clinical target volume in radiotherapy treatment planning for lung cancer? Radiother Oncol 1998;48:71–7.[Medline]
-
Ohara K, Okumura T, Akisada M, et al. Irradiation synchronised with respiration gate. Int J Radiat Oncol Biol Phys 1989;17:853–7.[Medline]
Top
Abstract
Introduction
