Set‐up error and dosimetric analysis of HexaPOD evo RT 6D couch combined with cone beam CT image‐guided intensity‐modulated radiotherapy for primary malignant tumor of the cervical spine

Abstract Purpose To investigate the set‐up error and consequent dosimetric change in HexaPOD evo RT 6D couch under image‐guided intensity‐modulated radiotherapy (IG‐IMRT) for primary malignant tumor of the cervical spine. Methods Ten cases with primary malignant tumor of the cervical spine were treated with intensity‐modulated radiotherapy (IMRT) in our hospital from August 2013 to November 2014. The X‐ray volumetric images (XVI) were scanned and obtained by cone‐beam CT (CBCT). The six directions (6D) of set‐up errors of translation and rotation were obtained by planned CT image registration. HexaPOD evo RT 6D couch made online correction of the set‐up error, and then the CBCT was conducted to obtain the residual error. Results We performed set‐up error and dosimetric analysis. First, for the set‐up error analysis, the average error in three translation directions of 6D set‐up error of the primary tumor of the cervical spine was <2 mm, whereas the single maximum error (absolute value) is 7.0 mm. Among average errors of rotation direction, Rotation X (RX) direction 0.67° ± 0.04°, Rotation Y (RY) direction 1.06° ± 0.06°, Rotation Z (RZ) direction 0.78° ± 0.05°; and the single maximum error in three rotation directions were 2.8°, 3.8°, and 2.9°, respectively. On three directions (X, Y, Z axis), the extended distance from clinical target volume (CTV) to planning target volume (PTV) was 3.45, 3.17, and 3.90 mm by calculating, respectively. Then, for the dosimetric analysis, the parameters, including plan sum PTV D98 and D95, planning gross tumor volume D98 and D95, V100% of the plan sum were significantly lower than the treatment plan. Moreover, Dmax of the spinal cord was significantly higher than the treatment plan. Conclusion 6D set‐up error correction system should be used for accurate position calibration of precise radiotherapy for patients with malignant tumor of the cervical spine.


| INTRODUCTION
The radiotherapy for the tumor of spine demands high techniques and equipment. Intensity-modulated radiotherapy (IMRT), proton radiotherapy, carbon ion or other heavy ion radiotherapy equipment; stereotactic radiosurgery (SRS), or fractionated SRS (FSRT) 1,2 can maximize tumor doses while protecting normal tissues to the utmost extent. IMRT will increase the RT dose of the tumor of spine from the traditional conventional irradiation 40~50 Gy to 60~65 Gy. 3 However, due to the physiological and anatomical characteristics of the spine which is close to the spinal cord, a strict-dose-limiting organ, high doses irradiation must have strict position verification, etc., to ensure the accuracy and safety of treatment. Linear accelerators equipped with image-guided radiotherapy (IGRT) function have been applied in radiotherapy. With the continuous progress of imaging-guided technology, the accuracy of position verification during radiotherapy is also enhancing. 4,5 There are many reports on the study of set-up error, and different image verification techniques, different parts, different ways of position fixing and even set-up errors in different unit are not the same. Rudat et al. 6 used orthogonal plain film two-dimensional verification position. After correction of translation 3D position, the radiotherapy residual error which still remained more than 5 mm in different parts including chest, abdomen, and head and neck accounted for 18%, 27%, and 10%, respectively. IMRT requires maximum error during the radiation treatment process, the set-up error of which is usually required to be <2 mm. 7 It has been reported that the translation 3D set-up error of patients with vertebral metastasis was >7 mm by CBCT measurement. After error correction, the error in X (left to right), Y (head to foot), and Z (up to down) directions were 1.7 mm, 2.1 mm, and 1.3mm, 8 respectively. The successful development and application of HexaPOD evo RT 6D treatment couch and CBCT image-guided system in clinical tumor radiotherapy is to improve the accuracy of radiotherapy by improving image-guided technology. This paper analyzes the set-up error by image-guided of cervical spine patients, as well as the dosimetric changes caused by set-up errors.

| Materials
From August 2013 to December 2014, set-up error data and verification images of IMRT for 10 patients with primary malignant tumor of the cervical spine were collected. Among 10 patients, there were five cases of chordoma and five osteosarcoma. Four cases were without operation and six cases were postoperative residual or postoperative recurrence.
The median age was 33 yrs old (aged 15-64). All patients were treated with IMRT and signed-informed consents before treatment.

| Positioning method
All patients were in supine position with C/B pillow, fixed by five-fixedsites neck and 3 shoulder thermoplastic membrane. CT (Philips, Netherlands, Brilliance 1MBigBore CT) simulation was applied. Contrast-enhanced CT scan was conducted and the scanning range was from the supraorbital margin to the thoracic spine 4 level. The range included the whole cervical spine and accessories (layer thickness 3 mm, 120 KV, 200 mAs). The scanned image was transmitted via the MOSAIQ network to Oncentra (Nucletron, External Beam v4. 3) treatment planning system for outlining target area and designing treatment plan.

| Target area outlining and treatment plan development
The target area is to be delineate the margin of the tumor on the simulated CT axial images (after the MRI image fusion of the cervical spine in the radiology department of our hospital), which is defined as gross tumor volume (GTV) (Fig. 1  whereas the rotation error was simulated and achieved by changing the couch plate angle, rack angle, and rotation treatment couch ( Fig. 6). A new total treatment plan was obtained by superimposing planning dose of which has been calculated for 22 times, that was defined as a set-up error plan sum in this study.

| Statistical processing
The data were expressed in x ± s. SPSS 16.0 software was used to analyze the data of set-up errors and residual errors; and the set-up error plan sum and treatment plan dosimetry parameters. If the analyzed groups of data conformed to the normal distribution, then the paired T test was adopted; otherwise, Wilcoxon rank-sum test was used.

| The set-up error of CBCT scanning images after matching with CT positioning images
With two times of CBCT scans, before and after correction, a total of 169 times were conducted and 338 CBCT images were collected.
After testing, the error data obtained before correction were defined as set-up error and those after correction were residual error, both of which have normal distribution. The average error of the 6D setup error of the primary tumor of the cervical spine (Table 1)

| Calculation result of CTV to PTV external expansion margin
The results are shown in Table 3. According to the set-up error, after formula calculation, the external expansion margin MPTV was 3.45, 3.17, and 3.90 mm in the direction of X, Y, and Z axes, respectively. sum. There was a difference between the plan sum and treatment plan dose distribution (Fig. 7). In the plan sum, the D98 and D95 of the dosimetry parameter PTV, as well as the D98 and D95 of PGTV was significantly lower than the original treatment plan (Tables 4 and   5).There were differences in patient's plan sum the treatment plan dose-volume histograms (DVH) (Fig. 7).The set-up error will result in an average reduction in 2.71 and 2.98 Gy of PTV D98 and D95, respectively, and a missing volume of average PTV 8.46%. In particular, the V100% of plan sum PGTV was significantly lower than that in the treatment plan, leading to a missing volume of average PGTV 17.48%, with the difference in statistical significance (F = 6.764, p = 0.018).
F I G . 5. To simulate rotation error by changing couch plate angle, rack angle, and rotation treatment couch parameters (in yellow box).
There was no statistically significant difference in dose differences between D1cc and D2cc in the spinal cord generated by errors (Tables 6 and 7       that in the treatment plan, with an average increase in 3.18 Gy dose, which increased the risk of severe radioactive myelitis. In the radiotherapy of the cervical spine, oral mucosa should be properly protected. This study found that the set-up error on the Dmax and Dmean of oral mucosa was uncertain: dose in some patients increased and some reduced. Considering the tumor's distance to the mucosa is the key factor, the impact of the set-up error on the mucosa volume dose difference requires further study.

| CONCLUSION
Translation and rotation set-up errors of cervical spine IMRT should be corrected, HexaPOD evo RT 6D can effectively correct the 6D errors. The set-up errors can lead to insufficient dose of target volume, and increase the spinal cord dose of OAR.

ACKNOWLEDGMENTS
We thank the National Natural Science Foundation of China (NSFC) to Prof Junjie Wang.

CONF LICT OF I NTEREST
The authors declare that they have no conflict of interest.

FUN DING
This study was supported by National Natural Science Foundation of China (No.61631001).

AUTHORS' CONTRIBUTI ONS
Junjie Wang designed this study, reviewed this manuscript, and provided funding support. Ping Jiang collected and analyzed the data and drafted this manuscript. Xile Zhang designed the treatment plan of radiotherapy and assisted data analysis. Shuhua Wei collected data and reviewed this manuscript. Tiandi Zhao assisted the process of treatment.

DATA AVAILABILITY STATEMENT
The authors declare that all data supporting the findings of this study are available within the article.