Transit Dosimetry of Stereotactic Body Radiotherapy Treatments with Electronic Portal Dosimetry Device in Patient with Spinal Implant

In recent years, the use of the Electronic Portal Imaging Device (EPID) as an in vivo dosimeter has become widespread. However, reports of EPID for stereotactic body radiotherapy (SBRT) applications is scarce. There is no data on this topic especially when there are high-density materials in the radiation field. In this study, we aimed to investigate the dose distributions of SBRT treatment plans in patients with spinal implants by transit EPID dosimetry. Implants were inserted in phantoms that mimic the vertebrae, and VMAT plans were created on the phantoms to deliver 16 Gy radiation doses to the target in 1 fraction. Transit EPID measurements were performed for each irradiation. The results were compared with the treatment planning system using the gamma analysis method. According to the gamma analysis results, while the non-implant model met the acceptance criteria with a rate of 95.4%, the implanted models did not pass the test with results between the rates of 70% to 73%. In addition, while the dose difference in the isocenter was 1.3% for the non-implanted model, this difference was observed to be between 7 and 8% in the implanted models. Our study revealed that EPID can be used as transit dosimetry for the VMAT-SBRT applications. However, unacceptable dose differences were obtained by transit EPID dosimetry in the VMAT-SBRT applications of patients with an implant. In the treatment of such patients, alternative treatment methods should be preferred in which the interaction of the implants with radiation can be prevented.


Introduction
Bones are one of the most frequent distant metastatic site in patients with solid cancers (1). Bone metastasis can cause pain, pathological fractures, hypercalcemia, mechanical instability, and spinal cord compression (2). Treatment goals in patients with bone metastases include achieving pain or symptom control, preserving and restoring function, increasing local tumor control, and stabilization if necessary.
Therapeutic options to achieve these goals include analgesics, bone modifying agents such as osteoclast inhibitors, systemic anti-cancer treatments, radiotherapy (RT), and surgery (3). More than one option is often used for optimum results. The combination of surgery and radiotherapy is frequently applied in patients with symptomatic vertebral metastasis. This treatment approach includes stabilization of the vertebral column and decompression of neurological structures via surgery. Postoperative radiotherapy aims to relieve pain, increase re-mineralization, and local tumor control. It has been suggested that radiotherapy after decompressive surgery has better symptom control, improved functional status, and quality of life compared to radiotherapy alone (4,5).
While post-operative conventional external beam radiotherapy (EBRT) is commonly used for adjuvant radiotherapy, there are several concerns regarding its practice in treating patients with vertebral metastases. Due to the limitation of precise targeting with conventional EBRT and the tolerance dose of the spinal cord, it is challenging to deliver su cient dose to the vertebra. This issue is the leading cause of inadequate pain palliation and local tumor control(6). On the other hand, stereotactic body radiotherapy (SBRT) allows ablative doses while limiting the dose received by the spinal cord and other organs at risk. However, some technical di culties arise in postoperative SBRT applications in the presence of surgical implants. Spinal implants are materials with high density and atomic number, therefore they create dosimetric uncertainties in addition to the small eld dosimetry problems in SBRT applications. Several researchers have addressed the problem of spinal implants on dose distribution (7,8). However, experiments on measuring the effects of implants were conducted with thermoluminescent (TLD), optically stimulated luminescent (OSL), or lm dosimetry. These dosimetric systems are capable to measure point doses, or 2D dose distributions. The advantages and disadvantages of these equipments have been studied widely (9)(10)(11)(12)(13).

Implantation of spinal implants
Implantation of spinal implants was performed on lumbar vertebrae models. A posterior pedicle screw(PS) xation, a posterior pedicle xation and anterior column reconstruction with titanium cage(PSC), and a lateral pedicle screw xation and anterior column reconstruction(LSC) models were employed on 3 different vertebrae models (Fig. 1). A lumbar vertebra(LV) model was also used in the experiments to compare the effect of metal implants. The implants were made of TI6AL4V titanium alloy.

Acquisition of planning images
A Toshiba® Acqulion LB (Toshiba Medical Systems Corporation, Otawara, Japan) scanner was used to acquire all CT images of the lumbar vertebrae models. The CT acquisition parameters were adjusted for the metal artifact reduction (MAR) protocol. All the models were placed in a water-lled phantom for imaging. A plastic slab and plastic rods were used to center the vertebrae model in the phantom. Rods were also placed adjacent to the models as a guide for placing the models in the same position for every CT scan.

Determination of Target and Critical Volumes, and Creation of Treatment Plans
The images obtained from the CT device were transferred to the Raystation Treatment Planning System (TPS; RaySearch Laboratories AB, Stockholm, Sweden). The target was de ned at the L3 vertebra level, where the metal implants remained within the radiation eld. The spinal cord was delineated as a critical structure. Then, VMAT plans were created. The treatment plans were generated using Raystation TPS software. Raystation uses collapsed cone(CC) algorithm for the clinical dose calculation. CT number (HU) to electron density (ED) calibration curve con gured in TPS using CTP604 module of Catphan® phantom (insert: Air, Water, Acrylic, Bone20%, PMP, LDPE, Bone50% Derlin, Te on, and Polystyrene). Scanning parameters used on the calibration CT were 120 kV, 220 mA, and slice thickness of 2.5 mm.
VMAT plans with 2 full arcs (2*360 0 ) were created using 6 MV attening lter-free (FFF) photon beams, and the dose prescription was set to 16 Gy in 1 fraction. The spinal cord dose was limited to 12 Gy. In addition, an extra treatment plan was produced for the PS model by preventing beam entry from the implants (PS_blocked). We used 2 half arcs from 90 0 to 270 0 in counterclockwise and clockwise directions for this plan.

Irradiation, Dose Measurement, and Analysis
All irradiations were performed with Elekta Versa HD™ (Elekta, Stockholm, Sweden) treatment machine. The treatment machine was calibrated to obtain 1 cGy dose at 1.5 cm depth in water equivalent solid phantoms with 1 MU before each irradiation. Setup veri cation was maintained for each model with cone-beam computed tomography (CBCT) images. After the position of the phantom was veri ed, SBRT was started and transit radiation dose information was simultaneously collected by iViewGt (Elekta, Stockholm, Sweden). The collected data was automatically exported to the iViewDose software. iViewDose compares the EPID-reconstructed doses with the TPS doses. 3D gamma evaluation with 3% global dose difference/3mm distance to the agreement and threshold:50% was performed for analysis. The passing rate criteria were described as 90%. Also, the isocenter doses were evaluated and compared for each model. The acceptance criterion was set to 3% for the dose differences.

Results
The gamma analysis and point dose measurements were given in Table 1. Table 1 The results of gamma analysis between the EPID and TPS dose distributions, and the isocenter doses for TPS and EPID for each model. According to the results, LV and PS blocked models passed the gamma analysis. The measured doses at the isocenter were higher than the TPS doses for all models. However, they were acceptable for the LV and PS blocked models.  Our 2D dose pro le analysis for the PSC model showed that there were dose peaks in the region of implants as we observed for the PS model. The same results were also obtained for the LCS models.
We limited the arc angles for the PS blocked model to prevent beam ingress from implants directly. The result of the gamma analysis showed that the gamma passing rate for this model increased from 73-93% with this method (Fig. 4).

Discussion
In this paper, we have examined the effect of metal implants on dose distribution for the VMAT-SBRT applications with in vivo EPID dosimetry. The performance of in vivo EPID dosimetry for SBRT irradiations has been previously studied (17,(22)(23)(24). It has been suggested that EPID can be used for the prostate, lung, and spine VMAT-SBRT applications and seems to be an innovative approach for in vivo dosimetry (17). According to the phantom results of McCowen et al., when gamma analysis was performed between EPID and TPS with 3%/3mm analysis parameter, the passing rate for the spine was found to be 93% (17). Despite the fact that there were TPS and reconstruction algorithm differences with McCowen's study, our ndings appear to be well substantiated by him. The gamma results were about 95% for the LV model. The AAPM TG-119 determined the lower limit of the passing rate for IMRT as 88%, our results support that EPID can be used as an in vivo dosimetry in VMAT-SBRT spine irradiations.
Our main motivation in this study was to examine the effect of metal implants on in vivo EPID results. Gamma analysis passing rates for spine models with metal implants decreased to the order of 70%. This dose mismatch between TPS and the measured dose is outside the acceptance criteria. There is a considerable amount of literature on the mismatch of TPS and measured dose in the presence of metals in the radiation eld. The difference is mainly based on two reasons. The rst is the imaging artifacts created by metal implants. The second is the inadequate modeling of radiation transport in or near higheffective atomic number materials by commercially used dose calculation algorithms. In this study, the single energy metal artifact reduction (SEMAR) technique was used during imaging to reduce artifacts caused by metal implants. Murazaki et al. a rm that gamma passing rates were improved by 6% using SEMAR for the images including metal implants (25). The algorithm utilized was collapsed cone (CC) to generate the treatment plans. They demonstrated that the CC algorithm could be used for dose calculation in cases where metal implants were present in radiation elds (26, 27). However, the difference between the measured dose and the planned dose for the CC algorithm can differ up to 25%.
The difference may vary according to the effective atomic number of the material used. In vivo EPID measurements revealed that there was a major dose difference between calculated and measured dose for implanted models. We showed that for the PS model, the failed points in the gamma analysis cuold be reduced by blocking the beam entry from the implants. For other models, it was impossible to block the metal for VMAT irradiation.
The dose distribution around the high effective atomic number materials has been studied in various studies. Monte Carlo studies have shown that there is a sudden dose jump around the high effective atomic number material. These dose peaks are due to the backscattered radiation. It is known that the magnitude of the backscatter radiation depends on the type of radiation beam, the thickness, and the density of the material. Our results are in concordance with the literature. We observed sudden dose jumps near the metal implants by in vivo EPID measurements. The effect of backscattered radiation from metal implants on the signal picked up by EPID may have caused the detection of dose variation in this region.
The dose difference of more than 7% observed in point dose measurements can have dangerous consequences for the spinal cord in SBRT applications. The measured doses were higher than the TPS dose. These results suggest that even if spinal cord doses are within tolerance according to the TPS, the spinal cord may actually be exposed to a dose above the tolerance limits due to the dose uncertainty caused by the metal implants. This effect should be considered in patients with spinal implants.
Realistic doses can be obtained by creating treatment plans in which radiation beams do not pass directly through the implants. It was previously reported that blocking the beam entry from the prosthesis improves the quality of dose distribution in patients with HIP prostheses treated with VMAT. However, in patients with spinal implants, materials with high effective atomic numbers can be located adjacent to the target structures. In such cases, it is impossible to leave the implants out of the eld for the VMAT technique. In our previous study with the Cyberknife treatment machine, we showed that in the treatment of similar cases, identifying metal implants to TPS and preventing the radiation beams to pass through it signi cantly reduces dose uncertainties(28). If we evaluate the results of both of our studies together, it can be concluded that it would be more appropriate to treat such cases with Cyberknife.

Conclusion
Our work has led us to conclude that in vivo EPID dosimetry can detect dosimetric uncertainties between TPS and treatment. The gamma analysis results can be improved by preventing beam entry from the implants. However, in many cases, it is impossible to leave the metal implants out of the radiation eld for VMAT. For these cases, we suggest alternative treatment techniques. Ethical approval: This article does not contain any studies with human participants or animals performed by any of the authors.

Con icts of Interest
The authors have no con icts of interest to disclose.   TPS and EPID doses for the PS model, the failed points in the gamma analysis between them, and the 2D dose distribution for the implant axis.

Figure 3
The dose distribution calculated from TPS, the dose distribution reconstructed from EPID measurement, and the comparison of 2D dose distribution in the region of implants between TPS and EPID for the PSC model.