Journal of Applied Clinical Medical Physics

Quality assurance devices for

dynamic conformal radiotherapy

 

Victy Y. W. Wong, M.Sc.
Department of Clinical Oncology
TuenMunHospital, New Territories, Hong Kong, China
Victy_Wong@hotmail.com

(Received 20 January 2004; accepted 24 February 2004)

Two quality control devices, a light-field device and a radiation-field device, have been specially designed to facilitate the clinical implementation of conformal dynamic arc treatment (CDAT) and intensity modulated radiation therapy (IMRT). With these devices, the light field as well as the radiation field, projected from the individual beam at any treatment position (i.e. arbitrary gantry angle) can be evaluated. For application, the devices are attached at the front end of the couch and placed at the isocenter of the linear accelerator treatment system (LINAC). The devices are designed to be rotated parallel to the gantry head so that the light field and the radiation field projected from a direct beam can be assessed. The aim of the study is to evaluate the geometric precision of the beam placement and dosimetric accuracy performed in CDAT and IMRT, with the aid of these devices. The devices are placed separately from the LINAC during application and provide an independent check on the quality performance of the LINAC in three dimensions. The condition of gantry sag and any mechanical displacement resulting in field shift can be observed and traced during gantry rotation. Mistakes that occurred during the isocenter calibration can lead to significant displacement in the field projection, which would not be revealed with the conventional quality control setting (i.e. gantry 0o). This fact was demonstrated with the aid of the two quality control devices in the study. The influence of gravitational acceleration in MLC leaf positioning error, which would consequently lead to inaccurate dose delivery, was investigated. The results of our study show that the existence of gravitational influence is statistically significant, although the magnitude of the dose inaccuracy has been found to be small.

Key words: quality assurance device, leaf position accuracy, dosimetric verification, IMRT, dynamic conformal arc

PACS Numbers: 87.53.Kn, 87.56.Fc, and 87.66.Cd

 

I. INTRODUCTION

Over the past decade, to perform a high degree precision treatment, field shaping techniques are widely implemented to upgrade conventional radiotherapy to a 3-dimensional conformal radiation treatment. The benefits of dose conformity to the target volume while sparing dose to normal tissues and improving the target dose uniformity have been discussed in numerous literature.(1-7) The technique is based on adjusting the beam aperture to match the shape of the target at various gantry angles. This was performed initially, with a few conformal static beams through the use of the custom-molded blocks made of lead alloy. With the development of the dynamic multi-leaf collimator (DMLC), the field shape can be dynamically conformed to the target during gantry rotation. Thus, further improvement in dose conformity can be achieved through a series of conformal dynamic arcs. Lately, the field shaping technique has been extended into the concept of intensity modulated radiation, known as intensity modulated radiation therapy (IMRT). The intensity modulated dose distribution is acquired by dividing the treatment fields into a number of segments. Each segment is automatically shaped by the multi-leaf collimator (MLC) and irradiated as a function of the fraction of the total number of monitor units (MU).  The dose rate and the leaf movement for each segment are calculated by the MLC control system. The accuracy of the MLC leaf position at isocenter is defined by a tolerance factor which normally ranges from 0.05 to 0.5 cm. To achieve a tighter tolerance, in the case of IMRT, if the leaf cannot reach to the position in the required time, the treatment time will be increased, by the electron gun delay, resulting in a reduced dose rate. However, dose variations due to beam instability may be induced by the increase in beam hold-off incidence.(8) In case of CDAT, the change of the field shape is accompanied by the rotation of the gantry, and the speed of the gantry rotation is restricted by the limitation of the dose per arc degree as specified by the system. Therefore, the speed of gantry rotation cannot be adjusted to cater to the leaf positioning accuracy; a certain amount of leaf position error has to be accepted in order to make the treatment deliverable.

Although CDAT and IMRT is characterized by the superior conformal dose distribution, compared with the conventional modalities, its routine clinical implementation is partially held back by the complexity of beam verification. Therefore, to ensure that treatment is delivered accurately, it is essential that an efficient and effective quality assurance program is applied on a routine basis.

A review of the literature indicates that dose verification shall be performed during the commissioning of the system to ensure dose accuracy.(9-16)  Prior to treatment,  a routine QA control program is also a prerequisite, to ensure that the mechanical integrity of the treatment unit is maintained in an expected standard. In general, before treatment, the shape of the individual treatment field is checked upon the light field or radiation field. For the case of IMRT, it is common to perform a radiation fluence test of individual treatment fields. This test is generally performed at the zero degree gantry angles due to the setup expediency. The result of such a test, however, does not necessarily provide the same outcome as if performed at the treatment position. The mechanical displacement of the LINAC system during gantry rotation and the effects of gravitational acceleration on leaf position accuracy are always of concern,(17) as these can be the factors which affect the dose distribution. Moreover, such setting cannot be applied to the case of CDAT, as the changed beam shapes are accomplished during gantry rotation. In this study, two devices are introduced to allow the light field as well as the radiation field to be tested at any treatment position, to confirm whether the mechanical integrity of the treatment system is fulfilled.

II. MATERIAL AND METHODS

Device for Light Field Test

The light-field device is made of Perspex and consists of two concentric dials: a scale dial and an indicator dial, a paper holder and a supporting metal rod (Fig. 1). The two dials are designed to rotate independently of each other on the same axis. The indicator dial is connected to the scale dial at one end and has the paper holder attached on the other end. An angular scale of 360 degrees in 10 degree steps and a pointer for indication of angular rotation are marked, along the rims of the scale dial and of the indicator dial, respectively. The paper holder is designed to rotate in unison with the indicator dial. The amount of rotation can be traced on the scale dial at the corresponding position, as indicated by the pointer of the indicator dial. A center cross is marked on the paper holder that illustrates the axis of rotation of the system. The paper holder consists of two pieces of Perspex with dimensions 128 ′ 113 ′ 4 mm. A light field printout generated by the planning system is held between the Perspex sheets during the light field test. In order to avoid light diffraction during the light field test, an opening in square shape with dimensions 78 ′ 78 mm has been made in the middle of the upper sheet. The entire system is supported by a metal rod which is inserted through the central axis of the scale dial and acts as a supporting pole for rotation.

 

   

FIG. 1: (A) Light-field device. (B) A schematic diagram of a light-field device.

Device For Radiation-Field Test

The radiation-field device (Fig. 2) is very similar in design to the light-field device. Instead of the paper holder, a block of Perspex is used. The Perspex block is divided into two slabs. For attenuating of 6 MV photons, the upper slab is 13.3 mm thick, which is equivalent to water with a depth of 15 mm. The lower slab has a thickness of 80 mm for eliminating the back-scattered electrons, primarily produced by the 6 MV photons. A film is placed between the upper and lower slabs and secured by screwing two Perspex slabs together. A center cross is marked on the upper slab, which indicates the rotational axis of the whole system.

 

   

FIG. 2: (A) Radiation-field device. (B) A schematic diagram of a radiation-field device.

Device Calibration

Before application, the device must be calibrated at the horizontal level to define the offset position (0°). The calibration is performed with the device attached to the treatment couch. Instead of securing the device directly onto a treatment couch, a stereotactic couch mount is applied to ease the setup procedure as the couch mount provides the facility for level adjustment. The offset position is found by rotating the paper holder/Perspex block together with the indicator dial to a horizontal level. The horizontal level is confirmed by placing water level on the surface of the paper holder/Perspex block along the transverse axis of the treatment couch. When the horizontal level is achieved, the pointer marked on the indicator dial indicates the offset position. The offset position is then denoted with the scale dial, at the position marked with "0" division. The horizontal level along the longitudinal axis of the treatment couch is also adjusted with the couch mount. After calibration, the position of the scale dial must be kept constant throughout the whole test. The center of the paper holder/Perspex block is then brought to the position of the isocenter of the treatment system by aligning it with the wall lasers and the ceiling lasers. Before the light field or radiation field test, to avoid oblique projection, the horizontal level of the holder must be checked after the printout/film has been inserted to ensure that no tilting will be caused by the thickness of the printout/film. The direct beam projection is obtained by rotating the paper holder/Perspex block and placing it at the same angle as the gantry. Since the indicator dial is rotated together with the paper holder/Perspex block, the angle of rotation is shown on the scale dial, at the position pointed by the indicator dial.

Tests for Light Field and Radiation Field

The automation precision of the LINAC system applied for the treatment of DCAT and IMRT was evaluated with the aid of the light-field device and radiation-field device. The studies were taken on a Varian CLINAC 2100 C/D equipped with an m3TM (BrainLAB, AG, Germany) micro-multileaf collimator (mMLC). The mMLC consists of 26 pairs of tungsten leaves that allow the treatment field to be shaped up to 10 ′ 10 cm2. The leaf width ranges from 3.0 mm at the center of the field to 5.5 mm at the periphery. The maximum distance over centerline is 5 cm, and the maximum separation between leaves is 10 cm. The system is capable of both dynamic MLC (DMLC) and segmental MLC (SMLC) delivery techniques; however, only the DMLC delivery technique is applied in this study. During IMRT, the control system drives the leaves at their maximum speed of 1.5 cm/s in a stepping manner while modulating the dose rate to achieve the desired leaf position. Various dose rates are applied in a range of 100 MU/min to 600 MU/min. The default tolerance factors of 0.25 cm and 0.5 cm as specified by the vendor, which control the spontaneous MLC leaf position in IMRT and CDAT, respectively, were applied in the study. According to the vendor's specification, the leaf position accuracy is also governed by the secondary feedback and the stepping motor of the system which drive the leaf to the position with an accuracy of 0.1 mm. The test results of light-field and radiation-field were compared with the dose plan generated by the planning system (BrainSCAN v.5.1, BrainLAB AG). The studies  performed were as follows:

1)   The accuracy of the field projection of a dynamic conformal arc at different gantry angles was studied with the aid of the light-field device. The field shape projected from the gantry was compared, at every 10degrees of gantry angle, with the light field printout generated by the planning system.

2)   The effect of gravitational acceleration on leaf position and the dosimetric error during IMRT  were evaluated. Two clinical plans with 13 IMRT fields were applied for this study. The maximum dimensions of the target volumes were 27 ′ 22 ′ 38 mm and 50 ′ 50 ′ 28 mm. In order to study the gravitational effect, each field was repeated to irradiate in three settings: setting 1 - gantry 0° and collimator rotation 0°; setting 2 - gantry 90° and collimator rotation 0°; and setting 3 - gantry 90° and collimator rotation 90°.  When the gantry and collimator rotation both are at 90°, the direction of the MLC leaf motion is parallel to the effect of gravity. In this case, any gravitational acceleration acting upon the MLC during gantry rotation will result in an increase in leaf position error. In the case of gravitational influence, difference in dose distribution between setting 3 and setting 1 will be expected to be significantly higher than that between setting 2 and setting 1. A film was placed between the two Perspex slabs of the radiation-field device and was irradiated according to the value of the leaf sequence factor. The amount of radiation, ranged from 100 to 148 MU was given. In order to avoid the dose response saturation, the dose response sensitivity of the Kodak Xomat-V film was tested in a dose range of 0 – 250 MU. The film was placed at the isocenter with 15 mm buildup of solid water. In the study of each field, each set of three films was developed at the same time and followed by the dose scan with the film scanner (VIDAR VXR 12 plus), to minimize the variations arisen during film processing and scanning. The dose distributions were studied with a film dosimetry system (RIT v.3.13, Radiological Imaging Technology, INC). Using the dose distributions obtained at setting 1 as reference, the difference between dose distributions obtained in setting 2 and setting 3 were compared.

3) The displacement of beam projection, resulting from the inaccurate calibration of isocenter, was studied with the aid of the radiation device. In the study, the isocenter was inaccurately calibrated by a 1.1 mm shift from the actual position. The beam fluence profile of a single IMRT field was studied at three gantry angles: 0o, 50o and 90o. The amount of field displacement was assessed by image subtraction. The fluence images obtained at the shifted isocenter were subtracted by that obtained at the correct isocenter, at gantry zero.

III. RESULTS

1) Fig. 3a shows a printout of a dynamic conformal arc, with the overlapped light fields from 40°-160° generated by the planning system. The change of the beam shape can be tested step-by-step with the light-field device. Fig. 3b shows that the light field projected at the gantry angle of 40° is perfectly matched with the printout. However, sagging of gantry was noticed by the increasing displacement of the light field, from 90° to 160° beam projection, as shown by the thickened field border.

 

   

FIG. 3: (A) A printout of overlapped light fields of a dynamic conformal arc from 40°-160° was generated by the planning system, The change of the beam shape can be monitored step-by-step with the light-field device. (B) The light field projection at gantry 40° was found perfectly corresponding to the printout. Gantry sagging was observed by the increasing displacement of the light field from 90° to 160° beam projection, as shown by the thickened field border.

2) The dose response sensitivity of the Kodak Xomat-V film was studied by plotting the optical density against doses in the range of 0–250 MU. The result confirms that the dose range (100–148 MU) applied in our study is well beyond the dose response saturation region (the plateau of the curve).

The differences between the dose distributions obtained in setting 2 (gantry = 90°, collimator = 0°) and in setting 3 (gantry = 90°, collimator = 90°) were studied using the dose distribution obtained in setting 1 (gantry = 0°, collimator = 0°) as reference. By applying setting 1 as a reference, variations in pixel-to-pixel response, processing and scanning conditions have been accounted for and eliminated. Doses were normalized to the maximum dose of individual films. The dose histogram was studied, and the number of pixels enclosed by the isodoses from 20% to 90% in 10% steps was calculated. Neither extremely low (<20%) nor high doses (>90%) were used for the study, due to the substantial errors induced by the insensitive detection in low dose region during film scanning and the limitation in spatial resolution for small area dose. The differences in pixel number between setting 3 and setting 1 and between setting 2 and setting 1, as a function of isodoses from 20% to 90%, were calculated for all fields. The mean differences in pixel number and the standard deviations of 13 fields were plotted against the isodoses, as shown in Fig. 4. The unit pixel size is 339 mm. Fig. 4 shows a substantial difference, in terms of dose coverage, in setting 3 and setting 2, using setting 1 as reference. The mean differences in pixel number between setting 3 and setting 1 and between setting 2 and setting 1, as a function of isodoses, are ranged from 482 - 2525 and 296 - 1295, respectively. This is equivalent to area coverage of 0.6 – 2.9 cm2 and 0.3 – 1.5 cm2, respectively. The differences in the number of pixels between setting 3 and setting 1 and between setting 2 and setting 1 were compared in pairs, as a function of isodose from 20% to 90% in steps of 10%, for all fields.  The result of the paired t-test suggested significant difference (P < 0.001) in dose coverage between setting 3 and setting 2. Our results therefore support that there is certain influence of gravitational acceleration in leaf position error that subsequently leads to inaccurate dosimetry.

 

   
FIG. 4:  The mean differences in pixel number between setting 3 (gantry 90°, collimator rotation  90°) and setting 1 (gantry 0°, collimator rotation 0°), and between setting 2 (gantry 90°, collimator rotation 0°) and setting 1 of 13 fields, were plotted as a function of isodoses in the range of 20% to 90%.

3)   Fig. 5 demonstrates the displacement of radiation field resulting from the inaccurate calibration of isocenter. The isocenter was miscalibrated with 1.1 mm shifted vertically away from the actual position. This is shown by an IMRT field projected at different gantry angles 0°, 50° and 90°; no displacement was observed at 0°; however, increasing displacement was noticed as the gantry moved away from 0°. The amount of displacement was assessed by image subtraction. The image which was taken at the shifted isocenter was subtracted by the image without isocenter-shift, obtained at 0°. The images were found with a displacement of 0.6 and 1.1 mm in the vertical direction as the gantry moved to 50° and 90°, respectively.

 

   

FIG. 5: Displacement of beam projection resulting from the inaccurate calibration of isocenter by a 1.1 mm shift in the vertical direction, is demonstrated with an IMRT field projected at different angles 0°, 50° and 90°. No displacement was observed at 0° as shown in (A), and increasing displacement was noticed, as gantry moves away from 0°. The amount of displacement can be observed on the subtracted radiation images, with the isocenter-shift images subtracted by an image without isocenter-shift projected at 0°. Displacements in beam projection of 0.6 and 1.1 mm in vertical direction were documented at the gantry angle of 50° and 90°, as shown in (B) and (C), respectively.

 

IV. DISCUSSION

The devices introduced in the study provide an independent check on the accuracy of the automation performance of the LINAC and the mMLC system. As the device is positioned at the isocenter and separated from the LINAC system, it offers an accurate geometrical reference without error induction from the movement of the LINAC system. The current commercially available equipments, such as the stereotactic film holder produced by BrainLAB and the portal imager, may be applied for such studies. However, these devices are attached onto the gantry and rotated about the same axis as the gantry head. It is difficult to differentiate any geometrical inaccuracy caused by gantry sag, gantry displacement and inaccurate system settings (e.g. isocenter calibrations, gantry angles calibration, etc). Moreover, due to the weight of the tools themselves, perfect circular rotation along the circumference is impossible, as displacement caused by sagging is often experienced while the film holder/portal imager starts moving away from the zero gantry angle. With the devices introduced here, the above mentioned problems have been overcome; hence, the mechanical integrity of the LINAC system can be examined in a precise manner. The accuracy of the speed response of the MLC during DCAT and IMRT is also an important factor to ensure accurate dosage. Whether this will be affected by the orientation of the gantry head due to the gravitational acceleration effect should be considered. Our study proves that there is a certain influence in leaf position accuracy due to the gravitational acceleration. Although, by restricting the tolerance level, one can minimize the effect and further improve the dosimetry accuracy; however, it should be compromised with the practical capability of the system. Therefore, it is recommended that the optimal tolerance level should be carefully investigated for each treatment system in such a way that it should be practically tolerable and conceded with the acceptable dose accuracy. Moreover, with the aid of the radiation device, the dose distribution of individual IMRT field can be obtained at any treatment position. Any inherent mechanical defects can be revealed by comparing the dose distribution with the fluence profile calculated by the planning system.

V. CONCLUSIONS

Two quality control devices have been specially designed for light field and radiation field tests for the treatments of CDAT and IMRT. The devices provide an independent check on the mechanical performance of the LINAC system, which demonstrates geometric accuracy of beam delivery. Dosimetry accuracy can also be evaluated. With the devices, mechanical deterioration of the gantry can be traced easily, in any gantry position. Moreover, mistakes made during system calibration can lead to significant errors in field placement that generally cannot be observed with the conventional quality control setting (gantry 0o), but can be revealed with the aid of these devices. IMRT and CDAT are regarded as a complex treatment that involves complicated dose calculation and complex processes in dose delivery. The devices suggested here are effective and expedient tools that can be used to ensure that the mechanical integrity of the delivery system is maintained in an expected standard.

VI. ACKNOWLEDGMENTS

The author thanks Hei Yuen Choi for the construction of the devices, Rachel Wong for the illustration, Kristal Lau for preparation of the manuscript and Martin Szegedi from BrainLAB Ltd for the technical discussion.

VII. REFERENCES

  1. D. Solberg, K. L. Boedeker, R. Fogg, M. T. Selch, A. A. F. DeSalles, Dynamic arc radiosurgery field shaping: A comparison with static field conformal and noncoplanar dynamic arcs, Int. J. Radiat. Oncol. Biol. Phys. 49, 1481-1491 (2001).
  2. R. W. Laing, R. E. Bentley, A. E. Nahum, A. P. Warrington, Stereotactic radiotherapy of irregular target: a comparison betweeen static conformal beams and non-coplanar arcs, Radiother Oncol. 28, 241-246 (1993).
  3. R. J. Hamilton, F.T. Kuchnir, P. Sweeney, S. J. Rubin, M. Dujovny, C. A. Pelizzari, G. T. Chen, Comparison of static conformal field with multiple noncoplanar arc techniques for stereotactic radiosurgery or stereotactic radiotherapy, Int. J. Radiat. Oncol. Biol. Phys. 33, 1221-1228 (1995).
  4. L. Ma, P. Xia, J. L. Verhey, A. L. Boyler, A dosimetric camparison of fan beam intensity modulated radiotherapy with Gamma knife stereotactic radiosurgery for treating intermediate intracranial lesions, Int. J. Radiat. Oncol. Biol. Phys. 45, 1325-1330 (1999).
  5. A. L. Boyer, P. Geis, W. Grant, M. Carol, Modulated beam conformal radiotherapy for head and neck tumors, Int. J. Radiat. Oncol. Biol. Phys. 39, 227-236 (1997).
  6. M. H. Philips, K. J. Stelzer, T. W. Griffin, M. R. Mayberg, H. R. Winn, Stereotactic radiosurgery. A review and comparisons of methods, J. Clin. Oncol. 12, 1085 – 1099 (1994).
  7. V. S. Khoo, M. Oldman, E. J. Adams, J. L. Bedford, S. Webb, M. Brada, Comparison of intensity modulated tomotherapy with stereotactically guided conformal radiotherapy for brain tumors, Int. J. Radiat. Oncol. Biol. Phys. 45, 415-425 (1999).
  8. T. LoSasso, C. S. Chui, C.C. Ling, Physical and dosimetry aspects of a multileaf collimation system used in dynamic mode for implementing intensity modulated radiotherapy. Med. Phys. 25, 1919 – 1927 (1998).
  9. G. J. Kutcher, L. Coia, M. Gillin, Comprehensive QA for radiation oncology:Report of AAPM radiation therapy committes task group 40, Med. Phys. 21, 581-618 (1994).
  10. J. V. Dyk, R. B. Barnett, J. E. Cygler, Commissioning and quality assurance of treatment planning computers, Int. J. Radiat. Oncol. Biol. Phys. 26, 261-273 (1993).
  11. J. S. Tsai, D. E. Wazer, M. N. Ling, Dosimetric verification of the dynamic intensity modulated radiation therapy of 92 patients, Int. J. Radiat. Oncol. Biol. Phys. 40, 1213-1230 (1998).
  12. M. J. Brugmans, A. van der Horst, J. V. Lebesque, Dosimetric verification of the 95% isodose surface for a conformal irradiation technique, Med. Phys. 25, 424-434 (1998).
  13. D. A. Low, S. Mutie, J. F. Dempsey, Quantitative dosimetric verification of an IMRT planning and delivery, Radiother. Oncol. 49, 305-316 (1998).
  14. D. A. Low, L. Russell, L. Gerber, Phantoms for IMRT dose distribution measurement and treatment verification, Int. J. Radiat. Oncol. Biol. Phys. 40, 1231-1235 (1998).
  15. G. Grebe, M. Pfaender, M. Roll, L. Luedemann, Dynamic arc radiosurgery and radiotherapy: commissioning and verification of dose distributions, Int. J. Radiat. Oncol. Biol. Phys. 49, 1451-1460 (2001).
  16. N. Agazaryan, T.D. Solbergi, J. J. DeMarco, Patient specific quality assurance for the delivery of intensity modulated radiotherapy, J. Appl. Clin. Med. Phys. 4, 40-50 (2003).
  17. C. R. Ramsey, K. M.Spencer, R. Alhakeem, A. L. Oliver, Leaf position error during conformal dynamic arc and intensity modulated arc treatment, Med. Phys. 28, 67-72 (2001).