Evaluation of initial setup errors of two immobilization devices for lung stereotactic body radiation therapy (SBRT)

Abstract The aim of this study was to investigate the accuracy and efficacy of two commonly used commercial immobilization systems for stereotactic body radiation therapy (SBRT) in lung cancer. This retrospective study assessed the efficacy and setup accuracy of two immobilization systems: the Elekta Body Frame (EBF) and the Civco Body Pro‐Lok (CBP) in 80 patients evenly divided for each system. A cone beam CT (CBCT) was used before each treatment fraction for setup correction in both devices. Analyzed shifts were applied for setup correction and CBCT was repeated. If a large shift (>5 mm) occurred in any direction, an additional CBCT was employed for verification after localization. The efficacy of patient setup was analyzed for 105 sessions (48 with the EBF, 57 with the CBP). Result indicates that the CBCT was repeated at the 1st treatment session in 22.5% and 47.5% of the EBF and CBP cases, respectively. The systematic errors {left–right (LR), anterior–posterior (AP), cranio‐caudal (CC), and 3D vector shift: (LR 2 + AP 2 + CC 2)1/2 (mm)}, were {0.5 ± 3.7, 2.3 ± 2.5, 0.7 ± 3.5, 7.1 ± 3.1} mm and {0.4 ± 3.6, 0.7 ± 4.0, 0.0 ± 5.5, 9.2 ± 4.2} mm, and the random setup errors were {5.1, 3.0, 3.5, 3.9} mm and {4.6, 4.8, 5.4, 5.3} mm for the EBF and the CBP, respectively. The 3D vector shift was significantly larger for the CBP (P < 0.01). The setup time was slightly longer for the EBF (EBF: 15.1 min, CBP: 13.7 min), but the difference was not statistically significant. It is concluded that adequate accuracy in SBRT can be achieved with either system if image guidance is used. However, patient comfort could dictate the use of CBP system with slightly reduced accuracy.

introduction of kilovoltage imaging in modern linear accelerators has aided the evaluation of body fixation that rendered the stereotactic body frame less important. 9 Consequently, rigid immobilization is losing popularity due to imaging-based SBRT processes. This has led to many other devices currently available in the market. [10][11][12][13][14][15][16][17] The success of SBRT depends on the accuracy of localization and ultimately on the treatment. Generally, ≤3 mm setup uncertainty and tumor motion are acceptable in a good clinical practice, although 5 mm are acceptable in many institutions. [4][5][6][7]13 The rigid EBF device is restrictive in many respects due to its size, especially for obese and noncompliant patients, whereas a newer device, the Civco Body Pro-Lok (CBP: CIVCO Medical Solutions, Orange City, IA, USA), is an open architect and provides wider dimensions and flexibility in setting up these types of patients. In some institutions, SBRT for lung was performed with the CBP. 11,14,15 There have been great discussions regarding setup accuracy for the EBF. 6,[16][17][18][19] Wulf et al. 16 and Guckenberger et al. 17 found that positions of lung tumors have low reproducibility relative to external stereotactic coordinates and the bony anatomy. Foster et al. 18 analyzed effectiveness of the EBF in various treatment sites, such as, lung, liver, prostate, and spine. Worm et al. 19 reported inter-and intrafractional set up errors for the EBF in lung and liver. However, comparative data are not available regarding the accuracy and efficacy of EBF and CBP in IGRT-based SBRT. This study attempted to fill this gap by investigating these two commonly used commercial immobilization systems with an abdominal compression plate to evaluate their efficacy in setup accuracy for SBRT.  ing the CT simulation and treatment with SBRT. The EBF has a rigid body frame and various components, including a custom-made vacuum cushion, an abdominal compression plate, and a pneumatic bladder system that tilts the base of the frame without changing the patient position. The bladder system, consisting of a balloon under the frame, is mainly used for the correction of rotation by adjusting the amount of air in the balloon. The EBF system also features a stereotactic coordinate mapping system that allows localization of the isocenter from CT data. In contrast, the CBP is an open structure without a rigid frame and consists of a custom-made vacuum cushion under the patient's body, an abdominal compression plate or belt, knee and foot supports, and a wing board (Fig. 1). Both systems have an abdominal compression plate that is used to limit the diaphragmatic excursion visualized under fluoroscopy to ≤5 mm for CT simulation.

2.B | Treatment planning
Depending on the patient's body structure and comfort, one of the devices (EBF or CBP) was chosen at the time of simulation. However, there was an initial institutional bias based on our vast amount of experience with EBF. 1 21 Couch shifts associated with the interfractional setup errors were analyzed to assess the accuracy of the initial CBCT setups of each immobilization system along the three axes and by means of a 3D vector (3D vector: the root sum of LR 2 , AP 2 , and CC 2 ). If a large error (>5 mm) was detected in any direction (LR, AP, CC), an additional CBCT was acquired for verification of the localization after  The preparation time was defined as the interval between the patient entering the treatment room and the start of treatment, not including the time spent waiting for a physician to arrive at the treatment control room.

2.E | Statistical analysis
An unpaired independent t-test was used to compare the data for accuracy and efficacy. A value of P < 0.05 was defined as statistically significant.

3.A | Patient characteristics and acquisition of data for repeat CBCT
The means AE SD for the weight (kg), body mass index (BMI) kg m À2 , and GTV (cm 3

3.C | Accuracy of setup with CBCT
The number of cases that needed the repeat CBCT were 16  respectively. The registered shifts for EBF were significantly smaller (P < 0.01) than those for the CBP in all directions. The registered shifts using repeat CBCT were smaller than with the initial setup in any direction for either immobilization system. Table II summarizes the group means, systematic localization errors, and random errors for the second setup with repeat CBCT. The systematic and random errors both improved when compared to the initial setup errors with both immobilization systems. Because the CBP is a new immobilization system, its setup accuracy and effectiveness need to be compared to those of traditional immobilization systems, such as the EBF.

3.D | Setup efficacy
The initial setup using the EBF in the AP direction showed a large group mean error along the posterior, couch-down, direction in our study. In contrast, Guckenberger et al. 17   The initial registered shift in the CC direction was larger than that in any other direction for the CBP. In addition, the shift was significantly larger in comparison to the shift seen with the EBF.
Gutierrez et al. 11 found that the systematic error for the CBP was largest in the CC direction, which agrees with our results. This error was also larger than any reported for the EBF. The EBF uses two laser markers to determine the sternal and tibial positions of patients. Although the CBP offers a simple setup that uses only skin marks on a target position on the patient's chest to determine the patient's position along the CC direction, the EBF presumably shows greater accuracy than the CBP in the CC direction.
We also evaluated the initial setup errors for the tumor-matching localization. Soft tissue registration, wherever possible, is done routinely. Since it is image guided, physician decides about the image fusion and expected shift. Worm et al. 19 found a significant correlation between patient BMI and the mean 3D vector of the initial setup error for bone-matching localization. No researcher assessed the correlation between BMI and initial setup error for tumor matching. If the BMI is so large, it is more likely that patient will not fit in rigid frame. Hence, we hypothesize that EBF has larger errors in proportion to BMI. Our results indicated that the accuracy of setup From the first session to the fourth session, the EBF displayed a more stable setup accuracy than the CBP, which resulted from the method used for patient setup with the two immobilization systems. In this context, daily changes (e.g., a different therapist for each session) presumably affected the simple setup associated with the CBP. The setup method for the EBF, on the other hand, is generally complicated and intricate. Therefore, the EBF offers a more stable initial setup than the CBP in terms of daily changes in the accuracy of the initial setup. found the IGRT to be helpful in correcting the setup errors because its use resulted in an improvement of more than 2.0 mm in all directions for the EBF. However, there were some setup errors with additional CBCT. We found that the second registered shifts for EBF were significantly smaller than those for CBP in all directions. This may be due to the design of the systems, EBF being more rigid and CBP more freeflow. Other conditions except the body frame were almost same between both immobilization devices. Their respiratory conditions were almost the same between two immobilization devices as this was performed in a single institution and same physician. Additionally, during the simulation and treatment, the compression devices used are monitored carefully for the position and location of the compression.
This study is performed on a single platform and same machine with 4D console with same software for image matching.
We analyzed the interfractional reproducibility for the EBF and the CBP. Intrafractional reproducibility, in contrast, is a substantial concern for SBRT, given the long treatment times. Some studies have investigated the intrafractional errors for the EBF, 18,19 however, the CBP remains unstudied. One of the limitations of the present study is that no intrafractional data were included for the two devices. Further study of the intrafractional reproducibility should be conducted for the CBP.
The treatment setup time was shorter for the CBP than for the EBF because the CBP does not use a frame and involves fewer points requiring verification by the therapist. This shorter setup time constitutes a clear advantage of the CBP. When the action level was over 3 mm, the differences in incidence rate were 15.0, 20.0, and 17.5% in the LR, AP, and CC directions, respectively, and were larger when compared to the differences seen for an action level of 5 mm.
The preparation time was almost the same for both devices with our action level of 5 mm. Institutions where apply action levels of 3 mm will experience a longer preparation time for the CBP than for the EBF because more repeat CBCTs will be needed for the CBP than are needed for an action level of 5 mm.

| CONCLUSION
Our study indicates that the CBP offered shorter setup time, while the EBF required fewer shifts to compensate for interfractional setup error. Satisfactory accuracy for SBRT can be achieved with IGRT in either system, thus making SBRT more adaptable for differences in patient habits and for enhanced comfort.

ACKNOWLEDG MENTS
This article was supported by the JSPS Core-to-Core Program  | 67