Abstract
Multiple commercial phantoms are now available for performing end-to-end QA testing for stereotactic procedures. This project aims at directly comparing one of the newest phantoms on the market against a more established one by performing similar tests to determine whether results are similar and they can be used interchangeably. Both phantoms were used to evaluate the coincidence of radiation and laser isocenters of a linear accelerator. End-to-end dosimetric tests were also performed using both an ion chamber and film. As part of the testing, both phantoms were also evaluated in terms of their efficiency of setup as well as the time required to switch inserts for different tests. Results showed that the laser/radiation isocenter coincidence as determined from each phantom was highly correlated. Ion chamber results were within 0.5% of the expected values. Gamma (2%, 2mm) pass rates of corresponding films were within 2% between phantoms. These results show that both phantoms are capable of producing equivalent results for the QA tests evaluated here.
Keywords: SRS, SBRT, stereotactic, brain, lung, FFF, intrafractional motion, rotations, HexaPOD
1. Introduction
Stereotactic Radiosurgery (SRS) and Stereotactic Body Radiation Therapy (SBRT) involve the delivery of large amounts of radiation using five fractions or less. These techniques demand high precision and accuracy of delivery, and they entail a large number of steps in the overall clinical delivery process e.g. imaging, planning and delivery of the radiation dose. Quality assurance must be performed on each of these steps to ensure proper functionality of all equipment involved and that the results obtained are within clinically accepted tolerances. For example, image quality and spatial accuracy of the planning CT data set should be verified to be within an acceptable level1, 2. The accuracy of registration of an MRI image set with the planning CT dataset must also be verified3-5. However, simply ensuring that each individual process step is, independently, within acceptable levels is not enough, because there exists interplay between the various steps of the process. This interaction between process steps can lead to situations wherein the individual steps may be within established tolerances, but the process as a whole may not be. The so-called end-to-end QA test (EET) fulfills this task of checking the overall accuracy of the stereotactic process. The EET is a highly recommended test for any center performing stereotactic procedures6-12. There are several commercially-available solutions that are now available that are meant to be used for such QA tests, including the EET. Because tests such as end to end QA must be performed regularly and because the test can be involved and time consuming, it is important that the phantom used facilitate an accurate and efficient workflow. Setup and take down of the phantom must be straightforward, efficient and reproducible. The purpose of this work is to perform a side-by-side comparison of a recently released (2013), commercially available, non-anthropomorphic phantom with the non-anthropomorphic phantom that has been commercially available since the early 1990s, to determine whether they yield comparable results with regard to accuracy assessment, efficiency of setup and reproducibility of results.
2. Methods
2.1. Phantom Description
The two phantoms used in this study are the Lucy 3D(R) QA Phantom from Standard Imaging Inc (Middleton, WI, USA) and the more recently released StereoPHAN(TM) from Sun Nuclear Corp (Melbourne, FL, USA). Images of the phantoms are depicted in Figure 1. Both are modular phantoms with various inserts for the user to evaluate different steps of the entire stereotactic process. The phantom’s cores and inserts are precision-milled to ensure high reproducibility of geometry. For a complete description of each phantom as well as all of the inserts available, the reader is referred to the respective manufacturer’s website13, 14.
Figure 1(a).
Lucy 3D(R) QA Phantom in Frameless Environment
2.2 CT/MR Fusion Check
Most SRS/SRT treatments are planned using CT datasets. However, due to the low soft-tissue contrast in CT images, an MR dataset is often also acquired for the purposes of delineating the target(s) and organs at risk. The CT and MR datasets are then co-registered so that the contours drawn on the MR can be transferred to the CT. The accuracy of this process must be confirmed to ensure that unacceptably large errors are not introduced into the patient treatment due to the varying discretization and spatial resolution in the two datasets, as well as the geometric distortion that is part of most MRI datasets.
Given that both phantoms provide irregularly shaped/located, target-simulating inserts of precisely known shape and volume that can be imaged both in a CT and an MR environment, the accuracy and reproducibility of the CT/MR fusion process can be verified by each phantom. To evaluate this claim, CT and MR images of both phantoms were obtained using the two primary clinical protocols used to scan SRS/SRT patients at our clinic. The CT images were obtained on a GE Lightspeed 16 slice scanner (GE Healthcare, Milwaukee, WI, USA) using 1.25 mm slices and the “large” field of view (FOV) as per our clinical protocol. MRI images were obtained on a Siemens 1.5T Avanto MRI scanner (Siemens Healthcare, Erlangen, Germany) using the FLASH protocol which we use for most of our metastatic patients, as well as a T2 protocol. Both sets of MR images were obtained in the axial plane using 1.0x1.0x1.0 mm3 voxels. .
All images were obtained back to back so that any distortion seen in one phantom’s images would be seen in the second phantom’s image set also. The CT dataset was fused to both MRI datasets in the treatment planning system (version 4.5, iPlanRTDose, Brainlab AG, Feldkirchen, Germany) by first manually aligning the two datasets, and then using the “automatic” fusion functionality within the software to perform the final alignment. The final quality of the fusion was visually assessed by the same expert user for both phantoms and assigned a qualitative score of “acceptable” if any discrepancy in alignment was less than 1 mm, “marginal” if there was an alignment discrepancy of 1-2 mm and “unacceptable” if there was any disagreement of more than 2 mm. These criteria can be related to errors on the order of 1 voxel or less, 1-2 voxels and more than 2 voxels. The test is done to verify the global fusion quality. If the fusion is determined to be marginal, the fusion may need to be focused on the area of the target or a new MRI may need to be acquired.
As an additional means of comparing the two phantoms, the target simulating objects of known volume were also contoured on each dataset (CT, FLASH MRI and T2 MRI) by the same expert user and contouring software (iPlan RTImage, Brainlab AG, Feldkirchen, Germany) so that the contoured volumes could be compared to the vendor-reported known volumes for each phantom. Results for each phantom’s assessment of contour volume accuracy were tabulated.
2.3. Phantom Placement Reproducibility
In order to evaluate the efficiency of setting up the phantom, and quantify the spread in the spatial results that stem from variations in repeated placement of the phantom, an intra- and inter-user repeatability study was done. For the intra-user study, the same procedure was repeated five times by a single expert user on a single day. For the inter-user study, the procedure was performed once by three different expert users on a single day. This placement repeatability study was done for three scenarios : (1) using a stereotactic frame with no IGRT, (2) in a frameless environment using Exactrac imaging (Brainlab AG, Feldkirchen, Germany) for position verification and (3) in a frameless environment using kV conebeam CT (kVCBCT) for position verification. These three scenarios were used to determine whether the two phantoms gave similar results when used in these three common clinical scenarios.
Figure 1(b).
Lucy 3D(R) QA Phantom in Framed Environment
Figure 1(c).
StereoPHAN(TM) Phantom in Frameless Environment
Figure 1(d).
StereoPHAN(TM) Phantom in Framed Environment
2.3.1 Reproducibility in a framed environment
Both phantoms were mounted on a Brainlab stereotactic headframe (Brainlab AG, Feldkirchen, Germany) using the manufacturer-provided interface to attach the frame to the respective phantom. For each phantom, two CT scans (1.25 mm slice thickness and large FOV to include the whole frame in the CT) were obtained: one with the film insert oriented such that the film was in the coronal plane and one such that the film was in the sagittal plane. All CT scans were imported into the iPlan treatment planning system (version 4.5, Brainlab AG, Feldkirchen, Germany) and the center of the film insert was carefully identified as Ptcenter. For cases with the film in the sagittal plane, a point 5 mm anterior and 5 mm superior of Ptcenter was defined as isosag. For scans with the film in the coronal plane, a point isocor was identified that was 5 mm superior and 5 mm to the left of Ptcenter. In each case, a plan was created to deliver a single beam that was centered on either isocor or isosag and perpendicular to the film plane. The plans were created to be delivered on a Novalis Classic (Brainlab AG, Feldkirchen, Germany) linear accelerator using a 10 mm stereotactic cone. Target Positioner (TAPO) overlays were printed for each plan to be used for positioning purposes using the in-room lasers at the treatment vault. In the treatment vault, the plans were delivered always using the same order: (1) Lucy 3D(R) QA with film in the coronal plane, (2) StereoPHAN(TM) with film in the coronal plane, (3) Lucy 3D(R) QA with film in the sagittal plane and (4) StereoPHAN(TM) with film in the sagittal plane. This protocol was followed so that the setup was started anew with every irradiation. Prior to each delivery, a fresh piece of custom laser-cut Gafchromic 3 film (IMT, Troy, NY) was put in the appropriate film holder and the phantom assembled according to the desired film geometry. Since the dose was delivered with the gantry either in the AP or lateral position, a spirit level was used to set the gantry prior to dose delivery.
The whole procedure was repeated four more times (20 total deliveries, 10 for each phantom) over one day for the intra-user study. On a separate day, the whole procedure was performed once by three different users (12 total deliveries) for an inter-user study.
Indentations from the pins in the insert on the film were used to determine and mark the center of the film. All films were analyzed using RIT software (version 6.3, Radiological Imaging Technology, Colorado Springs, CO, USA) to determine the difference in location between the film center and the center of the radiation field. Due to the intentional eccentric location of target isocenter that was defined, the center of the irradiation field was expected to be 5mm away from the center of the film in each direction. This test essentially represents one version of the Winston Lutz test15 and, since it was done with the film in both coronal and sagittal planes, the difference between the laser isocenter and radiation isocenter could be quantified in all three principal directions.
2.3.2 Image Guidance Reproducibility in a frameless environment with stereoscopic imaging
For this test, CT scans of the same phantom geometries, as described in section 2.C.1, were obtained with the frameless target positioning array. The procedure outlined above was used to create plans to deliver dose perpendicular to the film and centered about an isocenter at point isosag and isocor previously described. CT images and plans were exported to the ExacTrac system (version 5.5, Brainlab AG, Feldkirchen, Germany) for the purposes of image guidance. At the treatment vault, the phantoms were irradiated using the same order as described in section 2.C.1 using the ExacTrac system to refine the phantom position. Since the phantoms are mostly made of homogeneous material, they offer very little image contrast when viewed using the planar stereoscopic images obtained using ExacTrac. Therefore, particular attention was paid to the alignment of the small marking pins within the film insert when the image fusion was done for IGRT purposes. Once the phantom was appropriately aligned, the gantry was set using a spirit level and the plan delivered. The entire intra- and inter- user studies described above were repeated on two separate days. Films were analyzed in RIT to determine the offset between the radiation isocenter and the center of the film.
2.3.3 Reproducibility in a frameless environment with kV conebeam CT
The procedure outlined in section 2.C.2. was used to create plans in the Eclipse treatment planning system for delivery on a True Beam linear accelerator with an HD120 multi leaf collimator (Varian Medical Systems, Palo Alto, CA, USA). Once again, the points shifted from the center of the film insert (isocor and isosag) were defined as the treatment isocenter and plans were created to deliver a single MLC-defined 2x2 cm2 beam perpendicular to the film plane. At the treatment vault, the same procedure as outlined in section 2.C.2. was used except that kVCBCT was used to verify the phantom position. Once again, particular attention was paid to the markers in the film insert when evaluating and adjusting the fusion. The intra- and inter-user studies were repeated over two days and all films were analyzed in RIT to determine the offset between the radiation isocenter and the center of the film.
2.4. End to End Testing
As a final step, both phantoms were used to perform full EET in a framed and frameless environment using both ExacTrac and kV conebeam CT (kVCBCT) for image guidance. Since taking measurements using different detectors (ion chamber or film) requires different inserts to be placed in the phantom, it is important that the final dose calculation be performed on a CT dataset representative of the phantom as used. For this project, the EET was done with film in the coronal plane to measure the relative dose distribution, and an A16 micro ion chamber (Standard imaging Inc, Middleton, WI, USA) to measure the absolute dose. The phantoms were CT scanned in the exact geometry they would eventually be used such that the dose calculation could be done on the correct dataset.
For each of the phantom positioning scenarios investigated, the plans were delivered twice to each phantom with the ion chamber and twice to each phantom with a film in the coronal plane, for a total of eight deliveries. In order to limit the effect that any systematic changes in machine behavior may have on a particular phantom (such as a changing output due to a steady change in atmospheric condition for instance), the phantoms were alternated between deliveries. For analysis, the ion chamber-measured doses were compared to the mean predicted dose to the ion chamber active volume as calculated by the treatment planning system. The film dose distribution was compared to the predicted distribution using RIT and the (2%,2mm) gamma to determine how close the two distributions were to each other.
2.4.1 EET using a stereotactic frame with no IGRT
The iPlan treatment planning system was used for this test. The treatment fields used to clinically treat a patient with a brain lesion were then applied to the geometric center of each phantom and the resulting dose distribution was calculated. The treatment fields used were from a clinical 9-field IMRT plan used to deliver 25 Gy to the PTV in 5 fractions. TAPO overlays were used to position the phantoms prior to dose delivery.
2.4.2 Frameless EET using ExacTrac for IGRT
This test was also done using the iPlan treatment planning system. The same procedure outlined in section 2.D.1. was used to create plans but the CT scans obtained with the phantoms without the frame were used. In the treatment vault, each phantom was positioned coarsely using a Lok-Bar(TM) and ExacTrac images were taken to improve the position as needed prior to dose delivery. For the case with dose delivery to film, the marking pins in the film insert were used to fine-tune the registration in the ExacTrac software while the ion chamber’s central electrode was focused on for registration when using the ion chamber.
2.4.3 Frameless EET using kVCBCT for IGRT
This test was planned using the Eclipse treatment planning system. The same procedure outlined in section 2.D.2. was used to create the plans but this time, the clinical plan used consisted of two VMAT arcs meant to deliver 5 Gy per fraction to the isocenter. At the treatment vault, each phantom was positioned roughly using a Lok-Bar(TM) and a kVCBCT dataset was obtained to refine the position as required prior to dose delivery. Once again, the fusion was done by mainly concentrating on the pins in the film insert or the ion chamber itself, depending on the geometry used.
3. Results
3.1. CT/MR Fusion Test
Figure 2 depicts the use of both phantoms in registering CT and MRI datasets. Both phantoms allow the user to easily assess the fusion qualitatively. Any large geometric distortions should be immediately visible as mismatches in the three targets. For this test, the quality of the fusion was designated as “acceptable” for both phantoms for all fusions performed, meaning that any observed discrepancies were 1 mm or less.
Figure 2(a).
StereoPHAN(TM) CT fused with FLASH MRI
3.1.1 Contouring Volume Assessment
Since contouring does include a certain amount of subjectivity, the same expert user contoured the objects in the CT, FLASH MRI and T2 MRI datasets for this study. The results are given in Table 1 and show that both phantoms produced the same general trend of under/over-estimation of volumes depending on the imaging modality.
Table 1.
Evaluation of volumes contoured on images acquired using different modalities using the two phantoms.
| Lucy 3D(R) QA Phantom | ||||||||
| CT | MRI FLASH | MRI T2 | ||||||
| Actual Volume (cc) | Contoured Volume (cc) | Difference (%) | Actual Volume (cc) | Contoured Volume (cc) | Difference (%) | Actual Volume (cc) | Contoured Volume (cc) | Difference (%) |
| 1.70 | 1.75 | 3% | 1.70 | 1.81 | 6% | 1.70 | 1.94 | 14% |
| 5.25 | 5.31 | 1% | 5.25 | 5.42 | 3% | 5.25 | 5.69 | 8% |
| 12.25 | 12.11 | -1% | 12.25 | 12.76 | 4% | 12.25 | 13.63 | 11% |
| StereoPHAN(TM) | ||||||||
| CT | MRI FLASH | MRI T2 | ||||||
| Actual Volume (cc) | Contoured Volume (cc) | Difference (%) | Actual Volume (cc) | Contoured Volume (cc) | Difference (%) | Actual Volume (cc) | Contoured Volume (cc) | Difference (%) |
| 0.52 | 0.52 | 0% | 0.52 | 0.55 | 5% | 0.52 | 0.58 | 11% |
| 0.52 | 0.54 | 3% | 0.52 | 0.56 | 7% | 0.52 | 0.57 | 8% |
| 3.90 | 3.77 | -3% | 3.90 | 4.00 | 3% | 3.90 | 4.19 | 8% |
Figure 2(b).
StereoPHAN(TM) CT fused with T2 MRI
Figure 2(c).
Lucy 3D(R) QA CT fused with FLASH MRI
Figure 2(d).
Lucy 3D(R) QA CT fused with T2 MRI
3.2. Phantom Positioning Reproducibility
Figure 3 graphically shows the individual results of the intra-user and inter-user reproducibility tests run on the two phantoms in the three different configurations (framed without IGRT, frameless with stereoscopic ExacTrac imaging for IGRT and frameless with kV-CBCT for IGRT). Table 2 summarizes the results in terms of the mean offset between the center of the radiation field and the center of the phantom as well as the standard deviation in the measurements. It should be noted that the intra-user and inter-user tests were done on different days and that, ideally, the expected center of the delivered field is 5 mm offset from the center of the film in all directions.
Figure 3(a).
Results of individual tests run in a framed environment without IGRT when the film was in the coronal plane. Note that results of individual intra- and inter-user tests all cluster around the same point, very close to the expected offset of 5 mm.
Table 2.
Summary of intra-user and inter-user studies performed using both phantoms in three different configurations. Due to the experimental setup, the expected mean offset was 5.00 mm in each case. For the difference in the superior direction, the data from both coronal and sagittal films was averaged. Note that the difference between the mean offsets as measured by both phantoms in all scenarios considered never exceeded 0.15 mm.
| Anterior of Film Center | Superior of Film Center | Left of Film Center | |||||||
| Phantom | Framed/ Frameless | IGRT Type | Type of Investigation | Mean Offset (mm) | Std Dev (mm) | Mean Offset (mm) | Std Dev (mm) | Mean Offset (mm) | Std Dev (mm) |
| Lucy | Framed | None | Interuser | 4.53 | 0.25 | 4.48 | 0.22 | 5.36 | 0.26 |
| StereoPHAN | Framed | None | Interuser | 4.59 | 0.11 | 4.49 | 0.19 | 5.50 | 0.25 |
| Lucy | Framed | None | Intrauser | 4.51 | 0.12 | 4.61 | 0.11 | 5.19 | 0.12 |
| StereoPHAN | Framed | None | Intrauser | 4.49 | 0.11 | 4.53 | 0.10 | 5.12 | 0.08 |
| Lucy | Frameless | ExacTrac | Interuser | 4.55 | 0.11 | 4.52 | 0.13 | 4.66 | 0.13 |
| StereoPHAN | Frameless | ExacTrac | Interuser | 4.44 | 0.12 | 4.45 | 0.15 | 4.65 | 0.10 |
| Lucy | Frameless | ExacTrac | Intrauser | 4.52 | 0.13 | 4.57 | 0.15 | 4.72 | 0.12 |
| StereoPHAN | Frameless | ExacTrac | Intrauser | 4.49 | 0.12 | 4.52 | 0.15 | 4.72 | 0.11 |
| Lucy | Frameless | kV CBCT | Interuser | 4.47 | 0.15 | 4.69 | 0.13 | 4.78 | 0.12 |
| StereoPHAN | Frameless | kV CBCT | Interuser | 4.46 | 0.12 | 4.61 | 0.13 | 4.81 | 0.14 |
| Lucy | Frameless | kV CBCT | Intrauser | 4.45 | 0.15 | 4.66 | 0.10 | 4.85 | 0.12 |
| StereoPHAN | Frameless | kV CBCT | Intrauser | 4.42 | 0.11 | 4.59 | 0.12 | 4.82 | 0.11 |
3.3. End to End Testing
Results for the dosimetric end-to-end tests done on the two phantoms in the three configurations tested are summarized in Table 3.
Table 3.
Results of End-to-End tests performed using both phantoms in different configurations. Note that the difference in dose as measured by the ion chamber was below 1% in all scenarios evaluated while the difference in average gamma never exceeded 2%.
| Ion Chamber Results | Film Analysis Results | |||||||
| Phantom | Framed/ Frameless | IGRT Type | Plan Type | Plan Dose (Gy) | Average Measured Dose (Gy) | Difference from Planned Dose (%) | Difference between Phantoms (%) | Average Gamma (2%,2mm) |
| Lucy | Framed | None | 9 Field IMRT | 5.40 | 5.38 | -0.4% | 0.6% | 90.0 |
| StereoPHAN | Framed | None | 9 Field IMRT | 5.22 | 5.23 | 0.2% | 88.1 | |
| Lucy | Frameless | ExacTrac | 9 Field IMRT | 5.40 | 5.33 | -1.4% | 0.1% | 84.5 |
| StereoPHAN | Frameless | ExacTrac | 9 Field IMRT | 5.22 | 5.15 | -1.3% | 83.4 | |
| Lucy | Frameless | kV CBCT | 2 Arc VMAT | 5.72 | 5.78 | 1.0% | 0.4% | 98.0 |
| StereoPHAN | Frameless | kV CBCT | 2 Arc VMAT | 5.59 | 5.63 | 0.6% | 96.9 | |
Figure 3(b).
Results of individual tests run in a framed environment without IGRT when the film was in the sagittal plane.
Figure 3(c).
Results of individual tests run in a frameless environment with stereoscopic x-ray IGRT when the film was in the coronal plane.
Figure 3(d).
Results of individual tests run in a frameless environment with stereoscopic x-ray IGRT when the film was in the sagittal plane.
Figure 3(e).
Results of individual tests run in a frameless environment with kV-CBCT IGRT when the film was in the coronal plane.
Figure 3(f).
Results of individual tests run in a frameless environment with kV-CBCT IGRT when the film was in the sagittal plane.
4. Discussion
4.1. CT/MR Fusion
Both phantoms offer an MR signal generator consisting of a small tank filled with oil (Lucy 3D(R) QA Phantom) or copper sulphate solution (StereoPHAN(TM)). Even with the signal generator, neither phantom was detected in our 1.5T Avanto MRI scanner when we tried to acquire the scout scan to define the scan range. This was solved by using a saline bag within the bore.
The fluid-filled inserts for both phantoms also offer enough contrast in CT scans for contouring the objects. Given that the inserts are created with high precision, the contoured volumes are easily compared to known volumes to quantify the error introduced from each imaging modality. Table 1 confirms that both phantoms can be used to for this purpose, with the difference between contoured and known volumes being clearly correlated between imaging modalities.
Due to the limited size of the MRI inserts for both phantoms (maximum of 8.5 cm), detection of image distortions within MRI images is limited to large errors. Given that geometric distortions from non-uniform magnetic fields tend to get greater towards the periphery of the field of view16-18, we would not recommend use of either phantom for fully characterizing this effect unless the phantom is positioned at the periphery of the bore for multiple scans.
4.2. Positioning Accuracy of Phantom and End-to-End Testing
Table 2 shows that the modified Winston-Lutz test done with both phantoms give very similar results for all three configurations tested, demonstrating equivalence of the two phantoms in that respect. Given this fact, it is not surprising that both phantoms also obtain similar results when used for end-to-end testing. With regard to efficiency, when used in the same chamber configuration (perpendicular to treatment beams), both phantoms equivalently allowed for a full end-to-end to be done in roughly 90 minutes. This time includes the time to obtain a CT scan, go through the planning process and two dose deliveries, one to an ion chamber and one to a film.
4.3. General Comments about Phantom Usage
One of the largest differences between the two phantoms is the tool-less nature of the StereoPHAN(TM), with modules inserted from the back of the phantom without need to disassemble or use tools. While changing a module and roughly repositioning the phantom took about two minutes for the StereoPHAN(TM), the same activity took up to five minutes for the Lucy 3D(R) QA Phantom, which does require tools. This time estimate is for a user with extensive experience with both phantoms. Unless switching inserts and reorienting phantoms is being done multiple times, this extra time would likely be considered to be minimal by most users compared to the total time required for a full EET.
Both phantoms are mostly made of uniform plastic material and offer very little contrast when used with stereoscopic imaging such as ExacTrac. It is therefore extremely important to use reference CT scans in the exact geometry that it will be imaged at the linear accelerator in order to maximize the information that is available for registration. This is especially important with the StereoPHAN(TM) which uses rather large (4mm diameter) aluminum spheres that are a few centimeters from the actual treatment area. Marking pins from the film insert are closer and more evenly distributed around the actual target area and should be used if possible (Figure 4). When using the ion chamber, a reference CT with the chamber should be used since it shows the central electrode of the chamber that can be very accurately aligned.
Figure 4(a).
Registration with the StereoPHAN(TM) film insert in the same orientation in reference CT data and on the stereoscopic images obtained for IGRT. Note that IGRT markers (circled) and marker pins in the film cassette are all correctly aligned.
The Lucy 3D(R) QA phantom does allow the user to perform certain tests that are not possible with the StereoPHAN(TM) at the time of this writing. For instance, the Lucy 3D(R) QA Phantom allows the user to verify geometric distortion in an MRI using a specific grid-like geometric distortion insert. There is also the possibility of using multiple films in a separate film insert to obtain a coarse 3D dose distribution instead of a single planar dose distribution. We did not test these functionalities here due to the lack of similar inserts with the StereoPHAN(TM). Given the constant research and development work done with such commercial products, and the fact that StereoPHAN(TM) is a relative new-comer to the field, we recommend that any reader check with the specific manufacturer to determine whether inserts are available for other tests not specifically covered here. The final decision of which phantom to use may rest in personal preference of the end user based on what specific questions they want the phantom to help them answer, as well as financial considerations. However, in the context of end-to-end testing, the primary focus of the present work, either phantom should be expected to yield acceptable and similar results.
This was confirmed in the all of the experiments performed for this project, using three clinically relevant scenarios of (1) a frameless environment with no IGRT, (2) a framed environment using planar stereoscopic imaging for IGRT and (3) a framed environment using kV CBCT for IGRT. The reproducibility of use for both phantoms was also verified by the similar results obtained by the intra- and inter-user studies performed in all three scenarios.
Figure 4(b).
Registration with the StereoPHAN(TM) film insert in different orientations in reference CT data and on the stereoscopic images obtained for IGRT. Note that IGRT markers (circled) are correctly aligned while marker pins within the film insert end up in different positions between the planning CT and stereoscopic images and are therefore misaligned.
While these findings are not necessarily surprising given that both phantoms are commercially available and have, presumably, been extensively individually validated by the manufacturers, we believe the current work provides a key piece of information for the medical physics community, namely an independent comparison of the two phantoms under identical usage to determine whether they are equivalent. Since we provide a clear explanation of how we use the phantoms, we also believe any new user can use our description in conjunction with our results during their own commissioning.
5. Conclusion
The purpose of this manuscript was to perform a side-by-side comparison of two phantoms that are meant to be used for testing the accuracy of the stereotactic process. Both phantoms gave essentially equivalent results when used in the same configuration for the same suite of tests. We observed that both phantoms were similarly effective at characterizing key SRS quality assurance parameters for the suite of tests investigated in this work.
6. Acknowledgements
The authors wish to express their gratitude to Sun Nuclear Corporation for providing us with a loaner of the StereoPHAN(TM) phantom for testing purposes. We would also like to express our gratitude to Standard Imaging for providing us with a loaner Lok-Bar™ platform to use with the Lucy 3D(R) QA Phantom.
A subset of these results were presented at the 57th annual meeting of the AAPM held in Anaheim, CA.
Footnotes
Authors’ disclosure of potential conflicts of interest
Dr. Sarkar reports non-financial support from Sun Nuclear Corporation, personal fees from Sun Nuclear Corporation, and non-financial support from Standard Imaging, during the conduct of the study.
Drs. L. Huang, Y.J. Huang, Rassiah-Szegedi, Salter and Szegedi have nothing to disclose.
Author contributions
Conception and design: Vikren Sarkar, Bill Salter
Data collection: Vikren Sarkar, Long Huang, Jessica Huang, Martin Szegedi, Prema Rassiah
Data analysis and interpretation: Vikren Sarkar, Long Huang, Bill Salter
Manuscript writing: Vikren Sarkar, Jessica Huang, Hui Zhao, Bill Salter
Final approval of manuscript: All authors
References
- 1.Solberg T.D., Medin P.M., Mullins J., Li S., “Quality assurance of immobilization and target localization systems for frameless stereotactic cranial and extracranial hypofractionated radiotherapy,” Int J Radiat Oncol Biol Phys 71, S131-135(2008). [DOI] [PubMed] [Google Scholar]
- 2.Murphy M.J., “The importance of computed tomography slice thickness in radiographic patient positioning for radiosurgery,” Med Phys 26, 171-175(1999). [DOI] [PubMed] [Google Scholar]
- 3.Jonker B.P., “Image fusion pitfalls for cranial radiosurgery,” Surg Neurol Int 4, S123-128(2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kooy H.M., van Herk M., Barnes P.D., Alexander E., 3rd, Dunbar S.F., Tarbell N.J., Mulkern R.V., Holupka E.J., Loeffler J.S., “Image fusion for stereotactic radiotherapy and radiosurgery treatment planning,” Int J Radiat Oncol Biol Phys 28, 1229-1234(1994). [DOI] [PubMed] [Google Scholar]
- 5.Yu C., Petrovich Z., Apuzzo M.L., Luxton G., “An image fusion study of the geometric accuracy of magnetic resonance imaging with the Leksell stereotactic localization system,” J Appl Clin Med Phys 2, 42-50(2001). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Benedict S.H., Yenice K.M., Followill D., Galvin J.M., Hinson W., Kavanagh B., Keall P., Lovelock M., Meeks S., Papiez L., Purdie T., Sadagopan R., Schell M.C., Salter B., Schlesinger D.J., Shiu A.S., Solberg T., Song D.Y., Stieber V., Timmerman R., Tome W.A., Verellen D., Wang L., Yin F.F., “Stereotactic body radiation therapy: the report of AAPM Task Group 101,” Med Phys 37, 4078-4101(2010). [DOI] [PubMed] [Google Scholar]
- 7.Drzymala R.E., Klein E.E., Simpson J.R., Rich K.M., Wasserman T.H., Purdy J.A., “Assurance of high quality linac-based stereotactic radiosurgery,” Int J Radiat Oncol Biol Phys 30, 459-472(1994). [DOI] [PubMed] [Google Scholar]
- 8.Mack A., Czempiel H., Kreiner H.J., Durr G., Wowra B., “Quality assurance in stereotactic space. A system test for verifying the accuracy of aim in radiosurgery,” Med Phys 29, 561-568(2002). [DOI] [PubMed] [Google Scholar]
- 9.Ramaseshan R., Heydarian M., “Comprehensive quality assurance for stereotactic radiosurgery treatments,” Phys Med Biol 48, N199-205(2003). [DOI] [PubMed] [Google Scholar]
- 10.Shaw E., Kline R., Gillin M., Souhami L., Hirschfeld A., Dinapoli R., Martin L., “Radiation Therapy Oncology Group: radiosurgery quality assurance guidelines,” Int J Radiat Oncol Biol Phys 27, 1231-1239(1993). [DOI] [PubMed] [Google Scholar]
- 11.Tsai J.S., Buck B.A., Svensson G.K., Alexander E., 3rd, Cheng C.W., Mannarino E.G., Loeffler J.S., “Quality assurance in stereotactic radiosurgery using a standard linear accelerator,” Int J Radiat Oncol Biol Phys 21, 737-748(1991). [DOI] [PubMed] [Google Scholar]
- 12.Warrington A.P., Laing R.W., Brada M., “Quality assurance in fractionated stereotactic radiotherapy,” Radiother Oncol 30, 239-246(1994). [DOI] [PubMed] [Google Scholar]
- 13. “http://www.standardimaging.com/phantoms/lucy-3d-qa-phantom/.”
- 14. “http://www.sunnuclear.com/medphys/machineqa/stereophan/stereophan.asp.” [Google Scholar]
- 15.Lutz W., Winston K.R., Maleki N., “A system for stereotactic radiosurgery with a linear accelerator,” Int J Radiat Oncol Biol Phys 14, 373-381(1988). [DOI] [PubMed] [Google Scholar]
- 16.Wang D., Strugnell W., Cowin G., Doddrell D.M., Slaughter R., “Geometric distortion in clinical MRI systems Part I: evaluation using a 3D phantom,” Magn Reson Imaging 22, 1211-1221(2004). [DOI] [PubMed] [Google Scholar]
- 17.Wang D., Strugnell W., Cowin G., Doddrell D.M., Slaughter R., “Geometric distortion in clinical MRI systems Part II: correction using a 3D phantom,” Magn Reson Imaging 22, 1223-1232(2004). [DOI] [PubMed] [Google Scholar]
- 18.Sumanaweera T., Glover G., Song S., Adler J., Napel S., “Quantifying MRI geometric distortion in tissue,” Magn Reson Med 31, 40-47(1994). [DOI] [PubMed] [Google Scholar]