Quantitative analysis of treatments using real‐time image gated spot‐scanning with synchrotron‐based proton beam therapy system log data

Abstract A synchrotron‐based real‐time image gated spot‐scanning proton beam therapy (RGPT) system with inserted fiducial markers can irradiate a moving tumor with high accuracy. As gated treatments increase the beam delivery time, this study aimed to investigate the frequency of intra‐field adjustments corresponding to the baseline shift or drift and the beam delivery efficiency of a synchrotron‐based RGPT system. Data from 118 patients corresponding to 127 treatment plans and 2810 sessions between October 2016 and March 2019 were collected. We quantitatively analyzed the proton beam delivery time, the difference between the ideal beam delivery time based on a simulated synchrotron magnetic excitation pattern and the actual treatment beam delivery time, frequency corresponding to the baseline shift or drift, and the gating efficiency of the synchrotron‐based RGPT system according to the proton beam delivery machine log data. The mean actual beam delivery time was 7.1 min, and the simulated beam delivery time in an ideal environment with the same treatment plan was 2.9 min. The average difference between the actual and simulated beam delivery time per session was 4.3 min. The average frequency of intra‐field adjustments corresponding to baseline shift or drift and beam delivery efficiency were 21.7% and 61.8%, respectively. Based on our clinical experience with a synchrotron‐based RGPT system, we determined the frequency corresponding to baseline shift or drift and the beam delivery efficiency using the beam delivery machine log data. To maintain treatment accuracy within ± 2.0 mm, intra‐field adjustments corresponding to baseline shift or drift were required in approximately 20% of cases. Further improvements in beam delivery efficiency may be realized by shortening the beam delivery time.


| INTRODUCTION
In proton beam therapy, it is known that interplay effects may arise, and the baseline of an internal tumor location may change during beam delivery. 1 Compared to the conventional passive scattering method, scanning methods have greater sensitivity to interplay effects such as baseline shift or drift due to breathing, heartbeats, or intestinal activity, which can lead to an inhomogeneous target dose distribution (e.g., hot or cold spots) in the absence of motion mitigation techniques. [2][3][4][5][6][7] Thus, interplay effects must be considered when treating a moving tumor with scanning proton beam delivery.
Various procedures have been developed to mitigate the effects of interplay on target dose distribution. Clinical approaches include respiratory gating using surrogate markers 2,6,8 and tumor tracking using implanted fiducial markers. 7,[9][10][11][12] Respiratory gating with surface markers uses the relationship between internal motion and surface markers or respiratory holding and surface motion. 2,6,8 Although monitoring the body surface is a valid surrogate for target motion in tumors close to the surface, such as those in the breast, these methods are vulnerable to intra-fractional changes in the relationship between the internal tumor motion and external surface. [13][14][15] Tumor tracking with implanted fiducial markers uses the relationship between the internal marker coordinates and planned marker coordinates calculated from treatment planning with computed tomography (CT). 7,[9][10][11][12] Since 2014, we have been clinically operating a synchrotronbased real-time image gated spot-scanning proton beam therapy (RGPT) system, 7,10,11 which is based on the X-ray real-time tumortracking radiation therapy (RTRT) system developed by Shirato et al. in 1999. 16 The synchrotron-based RGPT system uses two orthogonal sets of X-ray fluoroscopes, and the target is irradiated only when the measured marker position is within AE2.0 mm of the planned marker position. 12,17 The synchrotron-based RGPT system can reduce the irradiation volume by 50%-75%, which represents a significant reduction in the irradiation of normal tissue around the target. 7 Although the synchrotron-based RGPT system is effective for To evaluate the planning robustness related to patient positioning, we simulated the dose distribution by shifting the isocenter for six axes (left-right (L-R), anterior-posterior (A-P), and superior-inferior (S-I)) on the planning CT. 26,27 At our institution, the shift tolerance for each direction was set to a maximum of 8 mm for the liver and a maximum of 5 mm for other sites. 20 During the actual treatment, physicians determined the final patient-specific tolerance range based on the robust evaluation report and actual tumor motion observed in the fluoroscopy image. Table 1 lists the characteristics of the treatment plans for each category in this study, including the prescribed dose, fraction, number of plans, number of plans using a short-range applicator, number of sessions, number of fields per session, optimization method, and the CTV. We evaluated the use factor of the beam angle as the ratio of the usage number of the beam angle to the total number of fields. 18 Figure 1 shows the use factors of the gantry angles used in this study.

2.C | Synchrotron-based real-time image gated spot-scanning proton beam therapy (RGPT) system
We used the synchrotron-based spot-scanning proton beam system PROBEAT-RT. The synchrotron beam has a maximum range of 30 g/ cm 2 and an irradiation field size of 30 × 40 cm. In the synchrotronbased RGPT system, fluoroscopy images obtained from two orthogonal sets of the X-ray tube and flat panel detectors placed at AE45°r elative to the proton beam direction are used to observe static and dynamic tumor locations. Details of the proposed synchrotron-based RGPT system are provided in a previous report. 10,20 We used the gating function to manage the internal motion. The synchrotron-based RGPT system was used if the tumor was within a movable organ such as the lung, liver, pancreas, or prostate. We We checked the location of the marker with two orthogonal sets of X-ray fluoroscopes during radiotherapy by using real-time pattern recognition technology for automatic recognition of the projected figure of the gold marker in fluoroscopic images (Fig. 2). The pulse rate for fluoroscopy was 30 or 15 Hz for the liver, pancreas, and lung patients, and 1 Hz for prostate patients. 20 The ranges of the Xray tube voltage and current imaging parameters were 30-125 kV and 20-125 mA, respectively. The imaging parameters depend not only on the patient characteristics of the disease site but also on the beam angle. We set the imaging parameters for each patient and gantry angle as low as possible to reduce the exposure dose from the fluoroscopy imaging.
In the synchrotron-based RGPT system, the proton beam is gated when the marker enters a preassigned gating window. The gating window tolerance for the actual treatment of each patient was set to AE2.0 mm based on our previous study. 7,9,17,28 Figure 3 shows an example of how the gate signal was recorded for liver patients.

2.D | Synchrotron magnetic excitation cycles
Fundamentally, the synchrotron magnetic excitation cycles comprise injection, acceleration, flat top, and deceleration phases, as shown in When the synchrotron is operating, multiple gating beam delivery improves the gate irradiation efficiency and reduces the proton beam delivery time. Suzuki et al, 18,19 where R is the usage of the gating function. 20 The transmission and reception of the signal between the beam delivery machine and the control device are recorded chronologically in treatment log data. The details of the treatment process calculation are described in our previous report. 20 To compare the beam delivery time when the gate irradiation was not performed as part of the same treatment planning, we evaluated the difference between T BSR X, V, R ð Þ and the simulated beam delivery time under the ideal environment (T BSR,sim X, V, R ð Þ ). Figure 5 illustrates the difference between the synchrotron magnetic excitation cycles for an ideal environment and actual treatment. Here, the ideal environment is a situation where the gate on/off signal is always output as shown in Fig. 5a. The value of T BSR X, V, R ð Þ was quantitatively analyzed using proton beam delivery machine log data, and T BSR,sim X, V, R ð Þwas calculated from the in-house simulation system, which used the treatment planning data. In the in-house simulation system, it is possible to simulate the estimated T BSR,sim X, V, R ð Þ using the synchrotron magnetic excitation cycles and the number of spots and layers. The details of the calculation parameters in the synchrotron magnetic excitation cycle, such as injection time, maximum flat top period, maximum spill length, acceleration and deceleration times, and scan speed, are described in a previous report. 9

2.F | Evaluation of frequency corresponding to baseline shift or drift and beam delivery efficiency
To treat a moving tumor, it is important to understand the frequency corresponding to the baseline shift or drift and the beam delivery efficiency. We defined the frequency of intra-field adjustment corresponding to the baseline shift or drift as the ratio of the total number of sessions to the number of sessions of intervention or adjustment to address baseline shift or drift, such as couch moving.
In this study, we used the proton beam delivery machine log data to analyze the patient treatment process flow.
A flowchart of the RGPT machine log system in one treatment session with X treatment fields without the gating function (R = 0) F I G . 3. Example gate signals for a liver patient. The gate signal is only turned on when the difference between the actual and planned marker positions is in the gate window (AE2.0 mm). and with the gating function (R = 1) is shown in Fig. 6. In the RGPT machine log system, if we identified baseline shift or drift during the beam delivery, a characteristic signal (redo bone matching) corresponding to the baseline shift or drift was recorded between the beam-on signal for the first spot and the beam-off signal after the beam delivery to the last spot. Thus, we analyzed the frequency of intra-field adjustments corresponding to the baseline shift or drift for each category. We also defined the beam delivery efficiency of the RGPT system as the ratio of the total number of the gate on/off signals to the number of gates on signals during the beam delivery, as recorded in the log data. We evaluated the size of intra-field adjustment as the amount of couch movement using the record of the Patient Positioning Image and Analysis System (PIAS) (Hitachi, Ltd., Tokyo, Japan).

3.A | Evaluation of beam delivery time
We analyzed 127 sets of treatment log data from the proton beam delivery machine. In total, 2810 sessions were delivered with gated irradiation. As listed in Table 2, the mean T BSR X, V, R ð Þ was 7.1 min, the mean T BSR,sim X, V, R ð Þ was 2.9 min, and the difference between

3.B | Evaluation of frequency corresponding to baseline shift or drift and beam delivery efficiency
During the treatment with gating (see Table 3) the average frequency of intra-field adjustment corresponding to the baseline shift or drift was 21.7%, while the average beam delivery efficiency was 61.8%. Figure 7 shows the relationship between the CTV and beam delivery efficiency.
The average size of intra-field adjustment corresponding to the baseline shift or drift was 0.30 cm. The average change in the F I G . 4. Schematic illustration that shows the difference between the synchrotron magnetic excitation cycle of (a) the previous operation with respiration and (b) operation with multiple gating beam delivery. In the flat top phase of the synchrotron magnetic excitation cycle, the red line shows the period between the start of spot irradiation based on the gate on/off signal and the deceleration phase. In the waiting function, the green arrow indicates the elapsed time between the gate signal being turned off and the next gate signal being turned on.

| DISCUSSION
Currently, real-time image acquisition using biaxial fluoroscopy devices is clinically available not only from Hitachi's RGPT technology but also through other vendors. For example, the treatment is provided to lung and liver cancer patients using markerless tumor tracking with a carbon-ion pencil beam scanning system. 30 The number of patients that can be treated each day in particle therapy facilities is determined by various factors, including the patient setup phase and the beam delivery phase. [18][19][20] One way to increase the patient throughput is to decrease the beam delivery time, especially for gating irradiation. It is important to understand the processes underlying the gating irradiation. The findings of this study will be useful for particle beam therapy facilities where there is a need to treat more patients in a limited time, to predict the patient throughput.
Careful observation of the location of fiducial markers has shown that intra-field adjustments of the patient couch with the RTRT system are useful for maintaining treatment accuracy within AE2.0 mm despite the baseline shift or drift. 28 In the results for prostate cancer F I G . 6. Flowchart of the RGPT machine log system in one treatment session with X treatment fields. K is the index number of the treatment field and R represents the usage of the gating function.  (Table 3). duty cycle was 2-5 times longer than nongated proton beam delivery. 6 In our results, the actual beam delivery time with gated irradiation was 2.5 times higher than the simulated data for the same treatment plan in an ideal environment. Our study shows that the synchrotron-based RGPT system can realize a similar beam delivery time as respiratory-gated proton beam delivery. Our results also appear to indicate that the beam delivery efficiency does not depend on the CTV (Fig. 6).
This study had some limitations. One limitation was the difference in proton pencil beam scanning methods. The beam scanning process in the lateral plane is typically performed in different ways. [31][32][33][34][35] In contrast, this study on the synchrotron-based RGPT system was focused on the spot-scanning proton beam delivery with an inserted fiducial marker and gating. Thus, we could not confirm that the results would be the same for all beam delivery methods.
Another limitation was the beam delivery time, which may be too long in the era of proton therapy as an external beam radiation therapy. Because T BSR X,V, R ð Þdepends on the number of energy layers and spots, 18 39 or is applied to layer stacking beam irradiation in carbonion therapy. 40 Matsuura et al developed and evaluated a short-range applicator with an MRF for treating superficial moving tumors. 41 Clinically, we have used this applicator with and without gated irradiation.
The structure of the short-range applicator in the RGPT system limited the maximum irradiation field. The proton beam without the short-range applicator has a maximum irradiation field size of 30 × 40 cm at the isocenter. In contrast, the proton beam with the short-range applicator has a maximum irradiation field size of 14 × 14 cm at the isocenter because it must be situated a certain distance away so as not to block the X-ray FOV. Therefore, it cannot be applied to all tumor sites. In this study, this short-range applicator was used for gated irradiation in only 15 out of 74 treatment plans.
The other method for shortening T BSR X, V, R ð Þ is to reduce the number of layers and spots during the treatment plan. Van  in an energy layer, which may not be efficient. Multi-energy extraction can deliver multiple discrete energies within a single spill. [43][44][45] As mentioned above, it takes approximately 2s to change the next spill, 36 and there are many pauses during the beam delivery that depend on the gate signal, as shown in Fig. 5. There are no reports in the literature on combining the synchrotron-based RGPT system and multi-energy extraction. Thus, further research is required.

| CONCLUSION
In this retrospective study, we quantitatively analyzed the proton beam delivery machine log data. Based on our clinical experience with a synchrotron-based RGPT system, we determined the beam

CONF LICTS OF INTEREST
Dr. Shirato reports grants from Hitachi, Ltd. and Shimadzu Corporation during the study and has licensed patents titled "Moving body pursuit irradiating device and positioning method using this device" (US6307914B1) and "Charged particle beam system" (US9757590).
Dr. Umegaki reports grants from Hitachi, Ltd. during the study and has licensed patents titled "Charged particle beam system" (US9757590) and "Radiotherapy control apparatus and radiotherapy control" (US9616249). Dr. Shimizu reports grants from Hitachi, Ltd.
during the study and has licensed patents titled "Charged particle beam system" (US9757590) and "Radiotherapy control apparatus and radiotherapy control program" (US9616249). The other authors have no relevant conflicts of interest to disclose.