Development of a rotational set‐up correction device for stereotactic head radiation therapy: A performance evaluation

Abstract We developed a new head supporting device to provide accurate correction of rotational setup during image‐guided radiation therapy (IGRT), evaluating its correction performance and the efficacy of dose distribution in stereotactic radiotherapy (SRT) using a helical tomotherapy (HT) system. The accuracy of rotational motion was measured using an electronic inclinometer; we compared device angles and measurement values from 0.0° to 3.0°. The correction accuracy was investigated based on the distance between rotational centers in the device and on megavoltage computed tomography (MVCT); the correction values were compared using distances in the range of 0.0–9.0 cm using a head phantom with a rotational error of 1.5°. For an SRT with a simultaneous integrated boost plan and a rotational error of 3.0° in yaw angle using a head phantom, and for a single‐isocenter SRT for multiple brain metastases in the data of three patients, dosimetric efficacy of the HT unit was evaluated for calculated dose distributions with MVCT after rotational correction. This device can correct pitch and yaw angles within 0.3° and can be corrected to within 0.5° for each rotational angle according to the result of MVCT correction regardless of the rotational center position. In the head phantom study, the device had a beneficial impact on rotational correction; D99% for the target improved by approximately 10% with rotational correction. Using patient data with the device, the mean difference based on the treatment planning data was 0.3% for D99% and −0.1% for coverage index to the target. Our rotational setup correction device has high efficacy, and can be used for IGRT.

highly conformal dose distributions with good coverage of targets and normal tissue sparing. Several studies have reported on the use of HT in the treatment of multiple brain metastases by using SRT with the simultaneous integrated boost (SIB) technique. In SRT, increases in rotational setup error in patient positioning and changes in the distance of the target to the rotational center may be associated with significant dosimetric uncertainties in multi-target, singleisocenter SRT treatments. 3,6 This necessitates patient setup position to be corrected in six directions using an in-room imaging system in conjunction with a linac-based treatment couch. 7 However, rotational setup errors in pitch and yaw angles cannot be corrected in image-guided radiation therapy (IGRT) using current HT systems.
Although rotational error information can be acquired using pretreatment megavoltage cone beam computed tomography (MVCT), patient rotational errors cannot be modified on the fly.
In this study, we developed a new head support device that allows for accurate rotational setup correction of pitch and yaw angles for IGRT for head SRT using the HT system. To evaluate the accuracy of rotational correction, we investigated the operation and correction accuracies depending on the distance between the centers of rotation of the device and the MVCT image. We assessed the dosimetric efficacy of rotational correction for head SRT using the HT unit with both a head phantom and human data.

| MATERIALS AND METHODS
The head supporting device is shown in Fig. 1(a). This device is constructed from carbon material, and can be rotated in 0.1°increments for pitch and yaw angles using two screws [Figs. 1(b) and 1(c)]. By placing the commonly used head shell system on the top of this device, we can rotate the head position during head fixation. Using an index bar, the device is fitted onto the treatment couch in the specific position shown in Fig. 2(a). The patient immobilization device is connected by an index plug shown in Fig. 2(b). Moving the position of the index plug allows adjustment of head shell position. This device may be used during CT data acquisition for treatment planning as well as during beam irradiation. The device's contour will be contained in the planning CT image data, allowing modeling of beam attenuation and changes in surface dose in any treatment planning system for dose calculation. This device passed tests of mechanical rotational motion, load carrying, and deflection at the time of construction. Therefore, we have confirmed that this device has a safety function to be used.

2.A.2 | Consistency test with MVCT correction value
To validate the consistency of rotational correction in this device, we used the MVCT registration result. We used a human head phantom (RAN110®, Alderson Research Laboratories, Long Island City, NY, USA). Rotational setup errors of 0.0°-3.0°in the pitch and yaw angles were added to the device's position. MVCT image data from the HT unit were acquired in this manner, and differences from the     four fractions for D95% to the whole brain using the planning station (TomoHD System ver. 2.1.0®, Accuray, Sunnyvale CA, USA); the calculation grid size was 2.0 × 2.0 mm 2 , the field width was 2.0 cm, the rotation pitch was 0.287 and the modulation factor was 2.5. The volume of the two CTVs were 6.6 cm 2 for CTV 1 and 3.8 cm 2 for CTV 2. In this study, the rotational center of the device and the MVCT image being identical represented the ideal situation.
Dosimetric efficacy with rotational correction was evaluated by dose volume histogram (DVH) analysis using a calculated dose distribution with MVCT after rotational correction using the device. For dose calculation, a conversion curve for MVCT values to relative electron density value was generated using an electron density phantom (RMI-467®; GAMMEX RMI GmbH, Biebertal, Germany). Figure 6 shows the workflow of dosimetric efficacy investigation. The head phantom was located with a theoretical set-up error value of 3.0°in the yaw direction. Rotational error was corrected using our device referenced to the MVCT registration results. Dose distributions with and without correction in MVCT were compared.

2.C.2 | Analysis of patient data
We retrospectively analyzed data from three patients treated by single-isocenter SRT for multiple brain metastases using the HT unit.
The SRT plans were created with 30 Gy in three fractions prescribed for D99% to the CTVs using the planning station; with the same conditions of dose calculation as those in Section 22.6.C.12.6. The     Table 3 shows the results of correction accuracy based on the distance between rotational centers. The initial rotational angle of the device and rotational correction value agreed within 0.5°, and the residual error after rotation correction was also within 0.5°.

| DISCUSSION
In this study, we assessed a new rotational setup correction device and investigated the usability of this device referring to results of MVCT image registration in the process of IGRT using the HT system. As shown in Table 1, measurements of rotational motion matched the electronic inclinometer results within ±0.3°. Therefore, our results are sufficient for use of patient setup correction in head SRT, because rotations of 0.5°exerted only a minimal effect on target coverage. 3,7,8 As shown in Tables 2 and 3 have been employed. 9,10 Combining this device with 6-DoF couch systems will allow assessment of the different degrees of rotation of articulated parts of the body, such as the head and the neck, F I G . 7. Dose distributions in whole brain + simultaneous integrated boost (SIB) stereotactic radiotherapy calculated on multislice computed tomography (reference) and megavoltage computed tomography (with/without rotational correction).
F I G . 8. Results of the dose volume histogram analysis. Dose distribution for whole brain was similar to that of the ideal multislice computed tomography plan. Conversely, the dose distribution for clinical target volume (CTV) 1 degraded with rotational error; the D99% for CTV 1 improved by approximately 10% using our device. In situations where the new 6-DoF couch system is integrated into the HT system, its rotational accuracy, correctable rotation angle, and range are unclear.
The mean transmission factor of our developed device is about 1.0% with 1 arc irradiation with a 6 MV beam. This slight attenuation can be adjusted for in the treatment planning by including the contour of this device in the planning CT data. Patients can move in their head mask during treatment. 12 As shown in Table 5, patient movement during the treatment was speculated small motion. However, this device can improve the accuracy of these inter-fractional motions without extending patient setup time, although it cannot monitor patient motion during irradiation. For accurate radiation treatment using SRT for the head region, intra-fractional motion also has to be controlled. 13 Therefore, repeat intra-fractional imaging and patient set-up correction, and real-time monitoring devices, have to be combined with our device for reduction in the errors associated with intra-fractional patient movement.

| CONCLUSION
We developed a new rotational set-up correctable device for head SRT with an operational accuracy within 0.3°. This device had a beneficial impact on the rotational setup correction and irradiated dose dis-tribution for head SRT, and can be corrected to within 0.5°for pitch and yaw angle errors according to the degree of IGRT correction.

ACKNOWLEDG MENTS
We would like to thank the members of the Departments of Radiation Oncology and Radiology in Juntendo University.

CONFLI CT OF INTEREST
No conflict of interest.