4,177
Views
58
CrossRef citations to date
0
Altmetric
ORIGINAL ARTICLE

4D imaging for target definition in stereotactic radiotherapy for lung cancer

, &
Pages 966-972 | Received 10 Jul 2006, Published online: 08 Jul 2009

Abstract

Stereotactic radiotherapy of Stage I lung tumors has been reported to result in high local control rates that are far superior to those obtained with conventional radiotherapy techniques, and which approach those achieved with primary surgery. Breathing-induced motion of tumor and target tissues is an important issue in this technique and careful attention should be paid to the contouring and the generation of individualized margins. We describe our experience with the use of 4DCT scanning for this group of patients, the use of post-processing tools and the potential benefits of respiratory gating.

Local control of early stage non-small cell lung cancer (NSCLC) after conventional radiotherapy is disappointing. The delivery of doses of up to 70 Gy using 3D conformal radiotherapy results in local control rates of only about 50% Citation[1]. The delivery of higher doses using conventional conformal techniques is associated with unacceptable toxicity Citation[2]. In contrast, various studies have shown that hypofractionated stereotactic radiotherapy (SRT) increases local control rates to 90% and higher with acceptable toxicity Citation[3–6].

Motion of tumor and normal organs is an important issue in the treatment of NSCLC, since it impacts on the delineation of the target and normal tissues, the required margins and the dose distribution. As the risk of toxicity is strongly correlated with the amount of normal lung tissue irradiated, an optimal target definition is crucial. This is even more important in SRT for early stage lung cancer, in view of the high fraction doses and the steep dose gradients. In recent years, it has been demonstrated that individualized “patient-based” instead of general “population-based” margins should be used Citation[7], Citation[8]. Various approaches can be used for the determination of these individualized margins, including the use of multiple planning scans, CT scans generated at maximal inspiration and expiration Citation[9], slow CT scans Citation[10], and more recently 4D CT scans Citation[11], Citation[12]. In this paper, the use of 4-dimensional CT scanning (4DCT) or respiration-correlated CT scanning, for target definition and treatment planning in the stereotactic treatment of stage I lung tumors will be discussed.

4DCT scanning

In a 4DCT scan, spatial and temporal information on shape and mobility are acquired synchronously in a single investigation. At VU University medical center, 4DCT scans are performed on a 16-slice GE Lightspeed scanner (General Electric Co., Waukesha, WI) with a slice thickness of 2.5 mm and an index of 2.5 (contiguous reconstruction). No intravenous contrast is used. The breathing pattern is recorded using the Varian Real-Time Positioning Management System (RPM) (Varian Medical Systems, Palo Alto, CA). In brief, the respiratory signals are recorded using infrared-reflecting markers on the upper abdomen of the patient during uncoached free breathing, as described previously Citation[12]. The markers are illuminated by infrared-emitting diodes surrounding a camera, which, at a frequency of 25 frames per second, captures the motion of these markers. The respiratory signal is recorded in synchronization with the x-ray “on” signal from the CT scanner. Eight slices of 2.5 mm are obtained during quiet respiration. The generation of a single 4DCT scan during quiet respiration is relatively simple and poses no problems to patients with poor pulmonary function. The 4DCT scanning procedure of the entire thorax takes about 90 s. Although the radiation exposure from 4DCT acquisition is approximately six times the dose of a single conventional helical CT scan (range 0.02– 0.09 Gy), the generation of individualized and usually smaller target volumes derived from 4DCT scans in comparison to standard PTVs justifies this additional radiation exposure.

The 4D datasets are sorted out for 10 phase bins within the respiratory cycle, using the Advantage 4DCT application running on an Advantage Workstation 4.1 (General Electric Co., Waukesha, WI). The registration is made based on the nearest neighbor criterion, where at each slice location the image is selected so that its phase is closest to each phase bin. The volumetric datasets, which represent 10 phases of respiration (bins), are used for further analysis using the Advantage software. Axial cine mode is widely used approach for 4DCT acquisition Citation[13]. This approach has been shown to be fairly accurate, but residual motion artifacts may arise due to partial projection effects, phase tolerance values for resorting, and irregularities in the breathing pattern of patients Citation[14], Citation[15]. A recent study suggested that the 4DCT system we used fails to generate optimal data in approximately 30% of patients Citation[15]. This may be caused by the fact that processing of respiratory traces may not accurately reflect the respiratory phase, and that the RPM-software excludes tracing from periods of irregular respiration and the Advantage4D software disregards these images. We are now evaluating the delivery of SRT under respiratory monitoring to ensure that the amplitude of respiration does not exceed that during 4DCT acquisition.

Generating individualized internal target volumes (ITV)

There are several methods to use 4DCT data for generating individualized target volumes. A simple method is to import all 10 respiratory phase bins into the Brainlab stereotactic radiotherapy planning system (Brainscan version 5.2, Brainlab AG, Heimstetten, Germany). Due to the axial ciné technique of 4DCT scanning with stationary couch, the resulting respiratory bins are ideally matched and no additional co-registration is necessary. GTVs are contoured in each of the 10 CT data sets using standardized lung window level setting, and the contours of all bins are automatically projected onto the first bin of the 4DCT set. As it is recommended that no separate GTV to CTV margin is to be used for stereotactic radiotherapy for early stage lung cancer, the ITV is defined as the volume encompassing the GTVs in all bins. A margin of 3 mm, which accounts for residual errors in patient positioning using the online correction protocol from the Novalis Exactrac system (Brainlab AG, Heimstetten, Germany), is added to the ITV to derive the individualized PTV (). In a group of 10 patients, it was demonstrated that the mean GTVs derived using the 4DCT scan technique were comparable to those obtained using six consecutive rapid planning CT scans. In two patients with very mobile tumors, however, the ITVs obtained from 4D scanning were considerably larger than those obtained using the multiple planning scan method Citation[12], indicating that 4D scans captured mobility that would have been otherwise missed.

Figure 1.  Two examples of individualized PTV generation for SRT of lung cancer using 4DCT scans. GTVs are contoured in each of the 10 phase bins of the 4DCT, and the ITV is defined as the volume encompassing all GTVs. A CTV to PTV margin of 3 mm is added to the ITV to derive the individualized PTV.

Figure 1.  Two examples of individualized PTV generation for SRT of lung cancer using 4DCT scans. GTVs are contoured in each of the 10 phase bins of the 4DCT, and the ITV is defined as the volume encompassing all GTVs. A CTV to PTV margin of 3 mm is added to the ITV to derive the individualized PTV.

We recently investigated the reproducibility of 4DCT scans obtained during quiet respiration in stage I NSCLC patients Citation[16]. Analysis of these repeat 4DCT scans that were generated in a single CT session in 20 patients revealed that volumetric and spatial differences in PTVs in excess of 10% and 2 mm, respectively, were observed in only a fifth of patients. Despite the use of highly conformal stereotactic treatment planning using 8 – 12 treatment beams, the potential variations in breathing cycle between CT scanning sessions translated into significant dosimetric differences in only a single patient.

Facilitating the clinical use of 4DCT scans

The need to import multiple CT datasets, and to contour tumors and normal organs in up to 10 respiratory bins, is a major drawback to the routine clinical use of 4DCT scans. In addition, significant changes in tumor volume and/or position can occur during the course of fractionated SRT for Stage I NSCLC Citation[17], which may indicate a need for repeat CT scanning and treatment planning during treatment. The use of such adaptive SRT would further increase the workload, and reliable auto-segmentation tools remain the subject of active ongoing research and not yet clinically available.

A simple approach for reducing this workload in clinical practice would be to use two-phase planning by importing only the two extreme phases of the 4DCT scan, i.e. the end-expiration and end-inspiration bins, followed by the generation of an encompassing ITV (). However, routine use of this method has two disadvantages. In case of small highly mobile tumors, the extreme positions of the tumor may be so far apart, that intermediate tumor positions are needed for generating a reliable ITV. Secondly, the use of two-phase planning disregards the possibility of tumor hysteresis, i.e. tumor mobility following a different path during inspiration and expiration, a phenomenon that has been described previously Citation[18].

Figure 2.  Two-phase planning target volumes based on contouring of the end-inspiration (yellow) and end-expiration (pink) phase bins of the 4DCT for a tumor showing predominantly cranio-caudal mobility.

Figure 2.  Two-phase planning target volumes based on contouring of the end-inspiration (yellow) and end-expiration (pink) phase bins of the 4DCT for a tumor showing predominantly cranio-caudal mobility.

A rapid method for generating reliable of ITVs from 4DCT datasets is to use the post-processing tool of maximum intensity projection (MIP) Citation[19]. Briefly, MIP projections reflect the highest data value encountered along the viewing ray for each pixel of volumetric data, giving rise to a full intensity display of the brightest object along each ray on the projection image. As such, these projections represent composite images with phase summation of tumor positions during all phases of respiration, thereby allowing for direct generation of ITVs. The MIP-based ITVs showed an excellent agreement with ITVs derived from contouring GTVs on all selected phases of the 4DCT Citation[19]. The minor discrepancies observed in ITVs generated using the two techniques can be accounted for by contouring variations, particularly as the displacement of the center of mass of both ITVs was less than 1 mm in all cases. An example of this technique is given in . The generation and contouring of individualized ITVs on MIP images took less than 10 min per patient. The clinical introduction of the MIP technique has significantly reduced the clinical workload associated with treatment planning for peripheral lung tumors. Because the MIP scans do not reflect the appropriate lung density for treatment planning, the latter has to be performed on a co-registered planning CT, e.g., a single phase of the 4D scan. In addition, the value of MIP scans is limited in cases where the tumor is adjacent to normal structures that have an equal or greater density on CT scans. This pitfall can be avoided by also projecting the GTVs at extreme phases of the movement for tumors that are adjacent to the mediastinum, heart, or diaphragm Citation[19].

Figure 3.  Examples of MIP scans that incorporate tumor mobility from all phases of the 4DCT scan. Left panel: Projection of GTVs from 10 phase bins (yellow) onto a maximum intensity projection (MIP) CT scan, with contouring on the MIP scan shown in red. Right panel: The end-inspiratory (yellow) and end-expiratory (pink) GTVs from the 4DCT of another patient, projected onto the MIP scan.

Figure 3.  Examples of MIP scans that incorporate tumor mobility from all phases of the 4DCT scan. Left panel: Projection of GTVs from 10 phase bins (yellow) onto a maximum intensity projection (MIP) CT scan, with contouring on the MIP scan shown in red. Right panel: The end-inspiratory (yellow) and end-expiratory (pink) GTVs from the 4DCT of another patient, projected onto the MIP scan.

Although phase summation using MIP scans allows for a fast derivation of individualized ITVs, this approach does not allow for optimal use of the temporal information contained in 4DCT scans. Therefore, we developed a color intensity projection (CIP) method, which incorporates the motion information of the 10 component data sets for each slice in a single composite color image, which is based on the MIP, minimum intensity projection (MinIP) and mean intensity projection (MeanIP) post-processing tools Citation[20]. Similar to MIP datasets, the MinIP and MeanIP reflect the minimum and mean data value encountered along the viewing ray for each pixel of volumetric data of the bins of the 4DCT scan. The CIP represents a projection of the intensity changes in the component images into a single image Citation[20], and is calculated on a pixel-by-pixel basis (). All ten component images are windowed and scaled linearly and the Max IP, MeanIP and MinIP over all ten phases for each pixel are calculated. Then, the hue-saturation-brightness (HSB) for each pixel of the color intensity projection is calculated. The brightness value ranges between 0 and 1, and is a measure of the intensity of a pixel. Saturation (value 0 to 1) is a measure of a pixel's amount of color and hue is the actual color and its value ranges from 0 to 0.66 with colors covering the range of blue-green-red. Where internal mobility is present, the hue of the color in the composite image encodes the period of time a tumor or organ is present at that location. CIP images enable mobility of both tumors and normal organs to be visualized and measured within a single composite image ().

Figure 4.  Frontal reconstruction of 4DCT scans in a patient with two tumors in the right lung. Images illustrate the respective maximum intensity projection (a), minimum intensity projection (b), mean intensity projection (c), and color intensity projection (d) protocols for this 4DCT study.

Figure 4.  Frontal reconstruction of 4DCT scans in a patient with two tumors in the right lung. Images illustrate the respective maximum intensity projection (a), minimum intensity projection (b), mean intensity projection (c), and color intensity projection (d) protocols for this 4DCT study.

Figure 5.  CIP images of a patient with a prior right pneumonectomy, showing mobility of the left diaphragm and left main stem bronchus (left). A tumor in the left lower lobe shows predominantly cranio-caudal mobility (arrow) and the colored ‘column’ indicates the trajectory of mobility. The green color of the most cranial position indicates that the tumor is at its end-expiratory position for 50% of the respiratory cycle.

Figure 5.  CIP images of a patient with a prior right pneumonectomy, showing mobility of the left diaphragm and left main stem bronchus (left). A tumor in the left lower lobe shows predominantly cranio-caudal mobility (arrow) and the colored ‘column’ indicates the trajectory of mobility. The green color of the most cranial position indicates that the tumor is at its end-expiratory position for 50% of the respiratory cycle.

In the ideal case, such as a bright, isolated lung tumor surrounded by air, the hue in CIP images is a good approximation of the percentage of time during which the tumor is in a particular position. Just as for MIP scans, this approximation however may not hold true in cases where the contrast between the tumor and surrounding tissue is not high or when two tissues of similar density occupy the same pixel at different phases of a respiratory cycle. The hue of the multiple-occupied pixel will then represent the proportion of time the two tissues combined are in the pixel.

Adaptive stereotactic radiotherapy

Initially, we performed repeat planning CT scan weekly prior to each fraction in patients undergoing a three or five-fraction scheme, and in the second week when the eight-fraction scheme was used (see ). An analysis of target reproducibility on the repeat planning 4DCT scans performed for SRT planning found a substantial interfractional shift of ITVs in a number of patients during the course of SRT Citation[17]. These shifts were present from the first week of therapy and did not correlate with the observed changes in tumor mobility. In addition, in this series of 25 tumors, we demonstrated that the ITV size generally tended to decreases during treatment, but that decrease was not statistically significant Citation[17]. However, in two cases (8%) a transient increase of more than 10 cm3 in ITV size was observed during treatment. The observed inter-fractional changes, which are likely to be radiation-induced changes in the surrounding normal tissue, illustrate the importance of repeat imaging during the course of SRT for lung cancer.

Table I.  Stereotactic radiotherapy schemes for Stage I NSCLC tumors at the VUmc

Respiratory gating

Respiratory gating permits a reduction in field sizes, because irradiation can be limited to phases in which the mobile target volume is in a predetermined position. Because gating prolongs the treatment time, it is common to use a gate that allows a duty cycle of 20 – 40% of respiration. We analyzed respiratory gating using three consecutive phases at end-tidal expiration to yield a 30% treatment window, which reduced the mean 3D tumor mobility vector from 8.5 mm to 1.4 mm. However, the delivery of gated SRT can take up to 1 – 2 h and the use of larger gating windows may be more patient-friendly. When evaluating the GTVs in cine movies of 4DCTs, a larger gating window appeared possible for some patients without undue increases in PTV. Our planning study assumed that gated radiotherapy at the end-expiratory phase was optimal for this patient group. Gating at end-expiration permits longer duty cycles and is facilitated because the expiratory phase is more reproducible Citation[21].

In selecting patients for gating, the absolute volume of the PTV is also of importance. A small benefit in a large PTV might clinically be of more relevance than a large benefit in case of a small PTV. Since the incidence of toxicity with SRT is low, we defined a threshold of at least 50% PTV reduction as a potential indication for respiration-gated SRT in these patients. Although a volumetric reduction was observed for most cases, a 50% reduction in PTV was only achieved with respiratory gating in five patients (15%). The gains from gating were greatest in mobile tumors, which we arbitrarily defined as tumors with a 3D mobility vector of at least 1 cm. The latter comprised about one third of our tumors in our series, and comparable figures for mobility have been reported by others Citation[22].

As gating appears of clinical benefit in only a minority of patients, it is essential to identify potential candidates for such treatment. Early reports on the implementation of respiration-gated radiotherapy for lung tumors described the use of fluoroscopy for selecting phases for respiratory gating Citation[21], Citation[23]. However, fluoroscopy does not permit accurate 3D determination of lung tumor mobility for SRT Citation[24].

A simple method for identifying mobile tumors in 4DCT datasets was established by contouring only the two extreme respiratory bins. The measure of tumor overlap between these two phases is highly predictive for tumor mobility, and it can be used to identify patients in whom treatment planning for respiration-gated radiotherapy is warranted. Another method for visualizing mobility is the use of CIP's, which allow for the mobility of the tumor, diaphragm, abdominal wall, or external marker in selected phases for respiratory gating to become immediately apparent. As the CIP for a single slice can be calculated within seconds, interactive gating phase selection is simple to implement. Identifying patients who might benefit from gating using CIP is more efficient than approaches such as screening on a 4D viewer or viewing multiple contoured target volumes for all patients. Although mobility screening of 4DCT data sets using dedicated software (Advantage 4D) is possible, objective screening of multiple structures simultaneously is not feasible. CIP enables the concurrent evaluation of external markers and tumor position as well.

Conclusions

Respiration-induced tumor motion is an important issue in high-precision extracranial SRT, and careful attention should be paid towards contouring the target and the generation of individualized mobility margins. Use of 4DCT information can be considered mandatory as it greatly improves target definition for SRT of early stage lung cancer, and clinical implementation is facilitated by the use of post-processing tools such as maximum intensity projection and color intensity projections. Respiratory gating is likely to benefit only a minority of patients, and both careful patient selection and repeated verification of the target volume between fractions is warranted when gating is performed. The clinical results with SRT, namely high local control rates and low toxicity, indicate that randomized trials comparing SRT with primary surgery are indicated

References

  • Lagerwaard FJ, Senan S, van Meerbeeck JP, Graveland WJ: Rotterdam Oncological Thoracis Study Group. Has 3D conformal radiotherapy improved the local tumor control in stage I non-small cell lung cancer? Radiother Oncol 2002;63:151–7.
  • Bradley J, Graham MV, Winter K, et al. Toxicity and outcome results of RTOG 9311: A phase I-II dose-escalation study using three-dimensional conformal radiotherapy in patients with inoperable non-small-cell lung carcinoma. Int J Radiat Oncol Biol Phys. 2005; 61: 318–28
  • Timmerman R, Papiez L, McGarry R, et al. Extracranial stereotactic radioablation: Results of a phase I study in medically inoperable stage I non-small cell lung cancer. Chest 2003; 124: 1946–55
  • Uematsu M, Shioda A, Suda A, et al. Computed tomography guided frameless stereotactic radiotherapy for stage I nonsmall cell lung cancer: A 5-year experience. Int J Radiat Oncol Biol Phys 2001; 51: 666–70
  • Wulf J, Hadinger U, Oppitz U, et al. Stereotactic radiotherapy of targets in the lung and liver. Strahlenther Onkol 2001; 177: 645–55
  • Onishi H, Araki T, Shirato H, et al. Stereotactic hypofractionated high-dose irradiation for stage I nonsmall cell lung carcinoma: Clinical outcomes in 245 subjects in a Japanese multiinstitutional study. Cancer 2004; 101: 1623–31
  • Stevens CW, Munden RF, Forster KM, et al. Respiratory driven lung tumor motion is independent of tumor size, tumor location, and pulmonary function. Int J Radiat Oncol Biol Phys 2001; 51: 62–8
  • van Sörnsen de Koste JR, Lagerwaard FJ, Nijssen-Visser MR, et al. Tumor location cannot predict the mobility of lung tumors: A 3D analysis of data generated from multiple CT scans. Int J Radiat Oncol Biol Phys 2003; 56: 348–54
  • Aruga T, Itami J, Aruga M, et al. Target volume definition for upper abdominal irradiation using CT scans obtained during inhale and exhale phases. Int J Radiat Oncol Biol Phys 2000; 48: 465–9
  • Lagerwaard FJ, van Sörnsen 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
  • Vedam SS, Keall PJ, Kini VR, et al. Acquiring a four dimensional computed tomography dataset using an external respiratory signal. Phys Med Biol 2003; 48: 45–62
  • Underberg RWM, Lagerwaard FJ, Slotman BJ, et al. 4-dimensional CT scans for treatment planning in stereotactic radiotherapy for Stage I lung cancer. Int J Radiat Oncol Biol Phys 2004; 60: 1283–90
  • Pan T, Lee TY, Rietzel E, et al. 4D-CT imaging of a volume influenced by respiratory motion on multi-slice CT. Med Phys 2004; 31: 333–40
  • Rietzel E, Pan T, Chen GT. Four-dimensional computed tomography: Image formation and clinical protocol. Med Phys 2005; 32: 874–89
  • Rietzel E, Chen GT. Improving retrospective sorting of 4D computed tomography data. Med Phys 2006; 33: 377–9
  • van der Geld Y, Lagerwaard FJ, van Sörnsen de Koste JR, et al. Reproducibility of target volumes generated using 4-dimensional CT scans for peripheral lung cancer. Int J Radiat Oncol Biol Phys, 2006 (submitted).
  • Underberg RWM, Lagerwaard FJ, van Tinteren H, et al. Time trends in target volumes for stage in non–small-cell lung cancer after stereotactic radiotherapy. Int J Radiat Oncol Biol Phys 2006; 64: 1221–8
  • Seppenwoolde Y, Shirato H, Kitamura K, et al. Precise and real-time measurement of 3D tumor motion in lung due to breathing and heartbeat, measured during radiotherapy. Int J Radiat Oncol Biol Phys 2002; 53: 822–34
  • Underberg RWM, Lagerwaard FJ, Slotman BJ, et al. Use of maximum intensity projections (MIP) for target volume generation in 4DCT scans for lung cancer. Int J Radiat Oncol Biol Phys 2005; 63: 253–60
  • Cover KS, Lagerwaard FJ, Senan S. Color intensity projections: A rapid approach for evaluating four-dimensional CT scans in treatment planning. Int J Radiat Oncol Biol Phys 2006; 64: 954–61
  • Ford EC, Mageras GS, Yorke E, et al. Evaluation of respiratory movement during gated radiotherapy using film and electronic portal imaging. Int J Radiat Oncol Biol Phys 2002; 52: 522–31
  • Neicu T, Shirato H, Seppenwoolde Y, et al. Synchronized moving aperture radiation therapy (SMART): Average tumour trajectory for lung patients. Phys Med Biol 2003; 48: 587–98
  • Shirato H, Shimizu S, Kitamura K, et al. Four-dimensional treatment planning and fluoroscopic real-time tumor tracking radiotherapy for moving tumor. Int J Radiat Oncol Biol Phys 2000; 48: 435–42
  • van der Geld YG, Senan S, van Sornsen de Koste JR, et al. Evaluating mobility for radiotherapy planning of lung tumors: A comparison of virtual fluoroscopy and 4DCT. Lung Cancer 2006; 53: 31–7

Reprints and Corporate Permissions

Please note: Selecting permissions does not provide access to the full text of the article, please see our help page How do I view content?

To request a reprint or corporate permissions for this article, please click on the relevant link below:

Academic Permissions

Please note: Selecting permissions does not provide access to the full text of the article, please see our help page How do I view content?

Obtain permissions instantly via Rightslink by clicking on the button below:

If you are unable to obtain permissions via Rightslink, please complete and submit this Permissions form. For more information, please visit our Permissions help page.