|Year : 2019 | Volume
| Issue : 1 | Page : 204-210
Evaluation of positional accuracy of the Varian's exact-arm and retractable-arm support electronic portal imaging device using intensity-modulated radiotherapy graticule phantom
Ranjit Singh1, HS Kainth2, Sachin Dev1, Gurjot Singh2, D Mehta2, JS Shahi2, Baljinder Singh3, Teerth Raj Verma4
1 Department of Radiotherapy, Post Graduate Institute of Medical Education and Research; Department of Physics, Panjab University, Chandigarh, India
2 Department of Physics, Panjab University, Chandigarh, India
3 Department of Nuclear Medicine, Post Graduate Institute of Medical Education and Research, Chandigarh, India
4 Department of Radiotherapy, King George's Medical University, Lucknow, Uttar Pradesh, India
|Date of Web Publication||13-Mar-2019|
Dr. Teerth Raj Verma
Department of Radiotherapy, King George's Medical University, Lucknow - 226 003, Uttar Pradesh
Source of Support: None, Conflict of Interest: None
Purpose: The aim of the present study was to compare the positional accuracy of varian's exact-arm (E-arm) and retractable-arm (R-arm) supporting electronic portal imaging device (EPID) systems (amorphous silicon flat-panel detector) using the intensity-modulated radiotherapy (IMRT) graticule phantom.
Materials and Methods: The known shifts of 0.5, 1.0, and 1.5 cm were introduced to the given phantom in longitudinal, lateral, and vertical directions, respectively, with respect to treatment couch of medical linear accelerator. The experiment was repeated for different gantry angle and varying source to imager distances (SIDs). The images were acquired for each shift at varying SIDs and beam orientations for both EPID supporting systems. The corresponding shifts obtained from treatment planning system (TPS) were recorded and compared.
Results: The known (expected) and observed (recorded from TPS) shifts obtained for different beam angles (namely, 0°, 90°, 180°, and 270° for anterior, left lateral, posterior, and right-lateral portal images, respectively) in the longitudinal, lateral, and vertical direction at varying SID were compared. The maximum shift in the observed value from the expected one was 3 and 2 mm, respectively, out of the all beam configuration for R-arm and E-arm. These shifts were randomly observed for all imager position and beam orientation.
Conclusion: The IMRT graticule phantom is an effective tool to check the mechanical characteristic and consistency of different EPID supporting arms. The effect of EPID sag due to gravity (gantry and treatment couch) was not significant for detection of shift in patient's position. The E-arm support EPID has better mechanical stability and accuracy in detection of patient's position than that of R-arm.
Keywords: Electronic portal imaging device, graticule phantom, retractable-arm and exact-arm
|How to cite this article:|
Singh R, Kainth H S, Dev S, Singh G, Mehta D, Shahi J S, Singh B, Verma TR. Evaluation of positional accuracy of the Varian's exact-arm and retractable-arm support electronic portal imaging device using intensity-modulated radiotherapy graticule phantom. J Can Res Ther 2019;15:204-10
|How to cite this URL:|
Singh R, Kainth H S, Dev S, Singh G, Mehta D, Shahi J S, Singh B, Verma TR. Evaluation of positional accuracy of the Varian's exact-arm and retractable-arm support electronic portal imaging device using intensity-modulated radiotherapy graticule phantom. J Can Res Ther [serial online] 2019 [cited 2020 Jul 6];15:204-10. Available from: http://www.cancerjournal.net/text.asp?2019/15/1/204/244473
| > Introduction|| |
Nowadays, electronic portal imaging device (EPID) is an important part in medical linear accelerators and is widely used not only for the verification of patient's treatment position but also for quality assurance (QA) of the treatment machine., In patient's setup verification, images are acquired by EPID during or before the treatment and compared with images previously acquired, that is, either taken on the 1st day of radiotherapy treatment or digitally reconstructed radiograph (DRR) obtained from computed radiograph (CT) of the same patient. QA of linear accelerator involves performance, field size verification, jaws position and orthogonality check, assessment of radiation beam flatness and symmetry, performance and positional accuracy of multi-leaf collimators (MLC), and inter- and intra-leaf radiation leakage of MLC.,, The working of EPIDs is based on the conversion of signals obtained from it into radiation-absorbed dose. The modern treatment modalities, for example, intensity-modulated radiotherapy (IMRT), image-guided radiotherapy (IGRT), volumetric-modulated arc therapy (VMAT), stereotactic radiosurgery, and radiotherapy require reproducible patient setup and accurate dose delivery. The clinical benefit from these advanced radiotherapy treatment modalities could be achieved only if there is no mismatch between planned and delivered dose distribution. A small mismatch between these two decreases the tumor control probability and increases the normal tissue complications. Hence, there is a need for efficient methods for QAs which is accurate, less time-consuming, and serves as an alternative to film dosimetry. The EPID systems are connected to the medical linear accelerator with different supporting arms supplied by different vendors. EPIDs have been characterized for different manufacturers with typical tests including dose response, stability over time, field size dependence, off-axis variation, image ghosting, and the backscatter effect from arm systems. The positional accuracy and reproducibility of different EPID supporting arms are very important for better accuracy and precision in treatment delivery. Several authors have evaluated the mechanical performance of different EPID supporting arms., The aim of the present study is to evaluate the positional accuracy and reproducibility of two different EPID (amorphous silicon flat-panel detector) supporting arms, that is, retractable arm (R-arm) and exact arm (E-arm) supplied by Varian Medical System (Varian, Palo, alto, CA) using the IMRT graticule phantom.
| > Materials and Methods|| |
In the present work, the measurements were carried out on two medical linear accelerator (Linac), namely, Clinac DBX-1160 and Trilogy-5823 (Varian medical system, Palo, alto, CA, USA). The Clinax DBX-1160 has single photon energy of 6 MV and provided with inbuilt Portal vision A-Si 500 for portal imaging. The Trilogy-5823 has dual photon energy (6 MV and 15 MV) and multiple electron energies (6 MeV, 9 MeV, 12 MeV, 15 MeV, and 18 MeV). The Trilogy-5823 is provided with inbuilt portal vision A-Si 1000 for MV imaging and portal dose measurement. The supporting arms provided for EPIDs were R-arm for Clinac DBX-1160 and E-arm for Trilogy-5823.
The computer-based treatment planning system (TPS), Eclipse (V 11.0, Varian Medical System, Palo, alto, CA) used analytical anisotropic algorithms (V11.0.31) and pencil beam convolution (PBC, V 11.0.32) algorithms., The target volume and normal structure present in its close vicinity determine the grid size required during calculation of treatment plan. The minimum and maximum grid sizes possible for the present version of planning system are 1 and 5 mm, respectively. The beams from different directions were generated, and dose distribution was obtained which covers the whole target volume with the prescribed dose while minimizing dose to normal tissue. IMRT graticule phantom [Figure 1], made of tissue equivalent material (18 cm × 18 cm × 18 cm), is routinely used as a tool to carry out film dosimetry for various radiotherapy treatment modalities such as IMRT, IGRT, and VMAT. This phantom has inbuilt radiopaque graticule which is clearly visible as markers in radiograph or CT. The radiographic images of the IMRT graticule phantom were acquired from the CT simulator with slice thickness of 2.5 mm. These images were now send to the planning system through digital communication, and the same treatment plans (four-field box technique) were made for both treatment machines using AAA algorithm. The treatment setup fields for anterior, left lateral, right lateral, and posterior portal images were also introduced to the plans. The DRRs obtained for each treatment setup fields from the TPS system were considered as reference images. The present graticule phantom was placed on couch of the machine in similar position/setup as it was during its CT image acquisition. The field isocenter was made to coincide with the machine isocenter by applying appropriate shift obtained from TPS plan in a different direction, that is, longitudinal, lateral, and vertical directions.
|Figure 1: The intensity-modulated radiotherapy graticule phantom on the couch of computed radiograph machine during image acquisition. The axes of the graticule phantom were aligned with all three lasers (longitudinal, lateral, and vertical) mounted in computed radiograph room, and lead balls (1mm diameter) were placed at three accessible faces of phantom at the point of intersection of lasers|
Click here to view
This position of the phantom was taken as the reference position, and EPID images were taken for anterior, left lateral, right lateral, and posterior setups. These images were analyzed with the DRR of the respective field setups, and difference between them was recorded. In supine or prone patient's position, the anterior or posterior portal image will give variation in the lateral (x-axis) and longitudinal (y-axis) direction while the left-lateral or right-lateral portals give changes in the longitudinal (y-axis) and vertical directions (z-axis), respectively. A shift of 0.5, 1, and 1.5 cm were introduced in longitudinal (in and out), lateral (right and left), and vertical (up and down) directions, respectively, and EPID images were acquired at different beam orientations (gantry angles 0°, 90°, 180°, and 270°) for each shift at source to imager distance (SID) of 140 cm. The similar measurements were repeated for other SIDs, that is, 149.9, 159.9, and 179.9 cm. Therefore, a total of 304 portal images from R-arm EPID were acquired for a single treatment, and difference between measured and expected shift was tabulated. The expected shifts and observed shifts at different SID and different portals are given in [Table 1], [Table 2], [Table 3]. The same experiment was performed for the other treatment machine (Trilogy-5832), and again, 304 portal images were analyzed, and measured shifts are given in [Table 4], [Table 5], [Table 6]. [Figure 2] shows the offline review while matching the image acquired in anterior and left-lateral portals and their respective reference images (DRR) for a known shift of 1 cm in longitudinal direction (outward) at an imager's distance of 150 cm.
|Table 1: Expected and observed shifts obtained for retractable-arm in the longitudinal direction (inward and outward) from different portal direction at different source to imager distance|
Click here to view
|Table 2: Expected and observed lateral shifts obtained in the anterior and posterior portal of retractable-arm|
Click here to view
|Table 3: Expected and observed vertical shifts obtained for retractable-arm in the lateral-portal images at different source to imager distance|
Click here to view
|Table 4: Expected and observed shifts obtained for exact-arm in the longitudinal direction (inward and outward) for different portal direction at different source to imager distance|
Click here to view
|Table 5: Expected and observed lateral shifts obtained in the anterior and posterior portal of exact-arm|
Click here to view
|Table 6: Expected and observed vertical shifts obtained for exact-arm in the lateral-portal images at different source to imager distance|
Click here to view
|Figure 2: The image acquired by A-Si1000 mounted on linear accelerator (Trilogy-5823) at imager position P5 (149.9 cm) for a known shift of 1 cm in longitudinal direction (outward). The images obtained consist (i) anterior digitally reconstructed radiograph image, (ii) image acquired for same setup from electronic portal imaging device, and (iii) blended image obtained during manual matching of acquired image (shown in background) and digitally reconstructed radiograph image (shown in inset). Similarly, the (iv) digitally reconstructed radiograph image for left-lateral setup, (v) image acquired in left-lateral portal, and (vi) manual matching for both digitally reconstructed radiograph and image acquired at left-lateral portal|
Click here to view
| > Results|| |
The mechanical performance of the two-arm EPID supporting systems has been compared with the introduction of a known shift in all three directions (longitudinal, lateral, and vertical), and the shifts were evaluated from the TPS. [Table 1], [Table 2], [Table 3] illustrate the expected and observed shifts obtained from R-arm-supported EPID (A-Si500) at different beam angles (gantry angles 0°, 90°, 180°, and 270° for anterior, left lateral, posterior, and right-lateral portal images, respectively) in the longitudinal, lateral, and vertical direction, respectively, at variable SID. The maximum difference obtained among all shifts in the longitudinal direction (inward and outward) was found to be 3 mm [Table 1], and it was reported for all SIDs except 179.9 cm (maximum error 2 mm). The occurrence of maximum error (i.e., 3 mm) was largest for SID 149.9 cm. Although no difference was observed for inward shifts in left-lateral portal images and outward shifts in right-lateral portal images both for distance 179.9 cm. [Table 2] gives information regarding the shift in lateral direction (left and right) obtained from portal images acquired at beam orientations 0° and 180° (anterior and posterior portal images) gantry rotations. The difference between the expected and observed shift in the lateral directions was recorded and maximum value was found to be 3 mm obtained both for imager distances 140 and 159.9 cm. On the other hand, largest error obtained both for anterior and posterior portals at imager distances 149.9 and 179.9 cm was 2 mm. [Table 3] illustrates the error measured in vertical direction for manually introduced shift. The maximum differences among expected and observed shifts (vertical direction) were found to be 3 mm, and its frequency is greatest for imager distance 179.9 cm. Meanwhile, no difference was observed for images acquired from left-lateral portals at SID 149.9 cm. [Table 4], [Table 5], [Table 6] provide the information regarding expected and observed shifts obtained in the longitudinal, lateral, and vertical directions for E-arm at different beam angles, that is, gantry angles 0°, 90°, 180°, and 270° for anterior, left lateral, posterior, and right-lateral portal images, respectively, and SIDs. The maximum difference among all shifts in the longitudinal direction [Table 4], [Table 5], [Table 6] was found to be 2 mm. [Table 4] illustrates the variation between expected and observed shifts in the longitudinal direction (inward and outward). The frequency of occurrence of maximum error (i.e., 2 mm) was largest at imager distances 149.9 and 179.9 cm. [Table 5] illustrates the data of lateral shifts obtained from portal image at 0° and 180° gantry angle (anterior and posterior portal images). The maximum difference between the expected and observed shift was found to be 2 mm and was obtained for all imager distances. [Table 6] shows the error obtained in vertical direction for manually introduced shift. The maximum shift between expected and observed values was found to be 2 mm, and the frequency of occurrence is maximum for imager distance 140 cm both for lateral-portal images and least for imager distance 159.9 cm.
| > Discussion|| |
The characteristic of treatment machine varies significantly at different gantry angles, and to ensure optimal operation of EPIDs, periodical gantry angle calibration must be performed for its mechanical stability. In the present study, IMRT graticule phantom was used to compare the mechanical stability and position accuracy of EPIDs equipped with two different supporting arms, that is, R-arm and E-arm. The maximum difference between expected and observed shift was found to be 3 mm and 2 mm, respectively, for R-arm and E-arm supporting EPID system. The error obtained is independent of beam orientation and it occurs randomly for almost all gantry angles and SIDs. The imaging of patients during treatment is done at different imager distances depending on the location of tumor, coverage of larger target volumes, and limitation of imaging system. The relation between expected and observed shift was well correlated, and geometrical accuracy was maintained at all imager position for both type of EPID supporting arm (R-arm and E-arm). The weight of the given phantom is negligibly small (~ 3Kg), hence, variation due to sag in treatment couch is negligible. However, in actual patient treatment, the error due to sag in couch because of patient's weight is further reduced with increase in the frequency of routine image acquisition by EPID. The reasons for difference between image acquired by EPID and reference DRR image are (a) change in patient setup variation during inter- and intra-fraction, (b) mechanical instability of gantry, couch, or supporting arm of EPID, (c) inherent errors in imaging or offline review system, (d) personnel error during image matching, (e) gravitation sag in EPID at different orientations with respect to phantom position, (f) anatomical changes in the patient during treatment, and (g) loss of gantry, couch, or EPID calibration. McDermott et al. show that the characteristics of EPID vary among manufacturers, and changes can also occur between EPID hardware or acquisition for the same linear accelerator. The possible reasons for the variation were suggested due to differences in image acquisition panel, panel design, and changes into readout electronics.
As per AAPM TG 142 report, the variation between planned and the patient treatment position for nonstereotactic radiotherapy techniques should be <2 mm. They have reported large difference for in-plane direction (longitudinal direction) as that for cross-plane direction (lateral direction). A similar study was carried out by Rowshanfarzad et al. in which large variation due to sag in EPID was observed for in-plane direction as compared to cross-plane direction. The deviation in daily patient's setup may require verification of, for example, reproducibility of SID, sag in the gantry during rotation therapy, and the carriage mounted in the head of the Linac unit to accommodate the different accessory, for example, MLC, and blocks. We carried our similar experiment and found graticule phantom as an effective tool to determine the accuracy and mechanical stability of R-arm supporting EPID. Imaging at large SIDs is routinely performed to cover the large size tumors during treatment, and mechanical stability will be greatly affected at extended SID. This is because the EPID has to move within larger circumferences with greater velocity during movement between any two adjacent orthogonal positions and hence requires the greater necessity of maintaining mechanical stability of EPID supporting arms at large SIDs. Therefore, before patient's imaging by EPID at extended imager positions, the setups need to be verified for reproducibility of SID, sag in the gantry during rotation therapy, or carriage mounted in the head of the linear accelerator unit to accommodate the different accessory, for example, blocks, MLC, tray, and wedges. The suitability of supporting arm (R-arm) was confirmed at larger extended SIDs.
Although in modern days, Radiotherapy units are equipped with CBCT and motion management systems even then EPID is used in large number of radiotherapy centers for patient setup verification and reproducibility., A study was conducted by Zaghloul et al. to check the ability of EPID and CBCT for children patient position verification. They compared the systematic and random error and found significantly correlated results from EPID and CBCT in almost all the directions for different tumor sites. In this study, two types of EPIDS have been compared in terms of their patient position verification accuracy evaluated in different beam configurations such as gantry angle and SIDs. Grattan and McGarry studied the mechanical performance of Varian E-arm and R-arm by recording the variation in its response with gantry angle. They found the largest difference among the responses of two EPIDs at 180° gantry angle, and error was found to be greater for R-arm as compared to E-arm. Similarly, in the present study, greater error was found for R-arm supporting EPID. The low value of errors observed in E-arm demonstrates its greater stability and accuracy in the patient positioning evaluation than R-arm.
| > Conclusion|| |
The IMRT graticule phantom is a useful tool to check the mechanical stability and positional accuracy of different EPIDs supporting arms. There is close agreement between expected and measured shift at different SIDs and different orientations (gantry angles 0°, 90°, 180°, and 270°), and effect of EPID sag due to gravity was not significant for detection of shift in the patient's position. The error obtained for E-arm was found to be lower as compared to that of R-arm. Thus, E-arm EPID support is slightly better in the evaluation of daily patient's setup and positions at all gantry angles and SIDs.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| > References|| |
Herman MG, Kruse JJ, Hagness CR. Guide to clinical use of electronic portal imaging. J Appl Clin Med Phys 2000;1:38-57.
Das IJ, Cao M, Cheng CW, Misic V, Scheuring K, Schüle E, et al.
Aquality assurance phantom for electronic portal imaging devices. J Appl Clin Med Phys 2011;12:3350.
Murthy KK, Al-Rahbi Z, Sivakumar SS, Davis CA, Ravichandran R, El Ghamrawy K, et al.
Verification of setup errors in external beam radiation therapy using electronic portal imaging. J Med Phys 2008;33:49-53.
] [Full text]
Zhang G, Tang Y, Sa Y, Ma A. Daily quality assurance of linac radiation field. Zhongguo Yi Liao Qi Xie Za Zhi 2011;35:386-8.
Klein EE, Hanley J, Bayouth J, Yin FF, Simon W, Dresser S, et al.
Task Group 142 Report: Quality assurance of medical accelerators. Med Phys 2009;36:4197-212.
Balasingh ST, Singh IR, Rafic KM, Babu SE, Ravindran BP. Determination of dosimetric leaf gap using amorphous silicon electronic portal imaging device and its influence on intensity modulated radiotherapy dose delivery. J Med Phys 2015;40:129-35.
] [Full text]
Wong JW. Electronic portal imaging devices (EPID). InEncyclopedia of Radiation Oncology, Springer Berlin Heidelberg, 2013. p. 207-213.
Zelefsky MJ, Kollmeier M, Cox B, Fidaleo A, Sperling D, Pei X, et al.
Improved clinical outcomes with high-dose image guided radiotherapy compared with non-IGRT for the treatment of clinically localized prostate cancer. Int J Radiat Oncol Biol Phys 2012;84:125-9.
Michalski D, Huq MS, Hasson BF. Normal Tissue Complication Probability (NTCP). In Encyclopedia of Radiation Oncology, Springer Berlin Heidelberg 2013. p. 560.
Rowshanfarzad P, Riis HL, Zimmermann SJ, Ebert MA. A comprehensive study of the mechanical performance of gantry, EPID and the MLC assembly in Elekta linacs during gantry rotation. Br J Radiol 2015;88:20140581.
Huang YC, Yeh CY, Yeh JH, Lo CJ, Tsai PF, Hung CH, et al.
Clinical practice and evaluation of electronic portal imaging device for VMAT quality assurance. Med Dosim 2013;38:35-41.
Bragg CM, Conway J. Dosimetric verification of the anisotropic analytical algorithm for radiotherapy treatment planning. Radiother Oncol 2006;81:315-23.
Soukup M, Fippel M, Alber M. A pencil beam algorithm for intensity modulated proton therapy derived from Monte Carlo simulations. Phys Med Biol 2005;50:5089-104.
Monti AF, Berlusconi C, Gelosa S. Gantry angle dependence in IMRT pre-treatment patient-specific quality controls. Phys Med 2013;29:204-7.
Jin GH, Zhu JH, Chen LX, Deng XW, Huang BT, Yuan K, et al.
Gantry angle-dependent correction of dose detection error due to panel position displacement in IMRT dose verification using EPIDs. Phys Med 2014;30:209-14.
McDermott LN, Louwe RJ, Sonke JJ, van Herk MB, Mijnheer BJ. Dose-response and ghosting effects of an amorphous silicon electronic portal imaging device. Med Phys 2004;31:285-95.
Singh R, Verma TR, Kainth HS, Ram C, Mehta D, Rana BS, et al
. Evaluation of positional accuracy of EPID using IMRT graticule phantom in extended source to imager distance setups: formalism of QA. Int J Curr Adv Res 2017;6:2389-93.
Srinivasan K, Mohammadi M, Shepherd J. Applications of linac-mounted kilovoltage Cone-beam Computed Tomography in modern radiation therapy: A review. Pol J Radiol 2014;79:181-93.
Zhuang L, Yan D, Liang J, Ionascu D, Mangona V, Yang K, et al.
Evaluation of image guided motion management methods in lung cancer radiotherapy. Med Phys 2014;41:031911.
Zaghloul MS, Mousa AG, Eldebawy E, Attalla E, Shafik H, Ezzat S, et al.
Comparison of electronic portal imaging and cone beam computed tomography for position verification in children. Clin Oncol (R Coll Radiol) 2010;22:850-61.
Grattan MW, McGarry CK. Mechanical characterization of the varian Exact-arm and R-arm support systems for eight aS500 electronic portal imaging devices. Med Phys 2010;37:1707-13.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]