Performances of the beam monitoring system and quality assurance equipment for the HIMM of carbon‐ion therapy

Abstract Purpose The heavy‐ion medical machine (HIMM), which is the first commercial medical accelerator designed and built independently by the institute of modern physics (IMP) in Wuwei, Gansu Province, China, had officially completed clinical trials at the time of this article's writing. Three types of detector systems were developed based on the ionization‐chamber principle to monitor the beam parameters during treatment in real time, quickly verify the beam performance during a routine checkup, and ensure patient safety. Methods and materials The above‐mentioned detector systems were used for beam monitoring and quality assurance in the treatment system. The beam‐monitoring system is composed of three integral ionization chambers (ICs) and two multistrip ionization chambers (MSICs) as a redundant design. The irradiation dose, beam position, and homogeneity of a lateral profile are monitored online by the beam‐monitoring system, and safety interlocks are established to keep the test results under the predefined tolerance limitation. The quality‐assurance equipment was composed of one MSIC and one IC stack. The IC stack was used for energy verification. Results The off‐axis response of ICs is within a tolerance of 2%, and the dose interlock system (DIS) response time is less than 7 ms during the treatment process. The positioning resolution of MSICs reached 73 µm. The IC stack can verify the beam range within one spill and the measurement resolution is less than 0.2 mm. Conclusions The beam‐monitoring system (BMS) and quality‐assurance equipment (QAE) have been installed and run successfully within HIMM for two years and are associated with the HIMM treatment system to deliver the right dose to the correct position precisely. Furthermore, the daily QA task is simplified by it. Above all, the system has passed the performance test of the China Food and Drug Administration (CFDA).


| INTRODUCTION
Radiation therapy with carbon-ion beams has been rapidly spreading worldwide since it was pioneered at the Lawrence Berkeley Laboratories in 1977. 1 In contrast with conventional photon therapy, the significant advantage of carbon-ion beams is the existence of Bragg peak, which can result in a precise conformal dose deposition to the desired volume while sparing the surrounding tissues as much as possible. 2 On the basis of the technical and clinical research on carbon-ion therapy at the Heavy Ion Research Facility (HIRF) in Lanzhou, China, 3 a commercial medical accelerator with better performance than the HIRF has been designed and built independently by the Institute of Modern Physics (IMP) in Wuwei, Gansu Province, China, ie, the heavy-ion medical machine (HIMM). The entire facility mainly consists of two electron cyclotron resonance (ECR) ion sources (the second source is a backup), a cyclotron injector, a synchrotron, and three high-energy beam-transport lines that deliver the accelerated carbon ions into two treatment rooms. The beam generated by the ECR ion source is preaccelerated by the cyclotron to an energy of 7 MeV/u and then injected into the synchrotron with the charge-exchange injection scheme. The synchrotron cycle is designed to accelerate carbon ions with a kinetic energy of up to 400 MeV/u and a flux of up to 4 × 10 3 pps. 4 Once beams from ion sources have been accelerated to the required energy, the slow beam extraction starts, and particles are steered through the high-energy beam-transport line to the selected treatment room. One of the treatment rooms (Room A) is equipped with a fixed horizontal beam-transport line, and another room (Room B) has both vertical and horizontal beam-transport lines. In the vertical line, the beam is steered to the height of 15.5 m above the isocenter and then deflected to enter the treatment room vertically. The beam-transport line is operated in the modulated scanning mode 5 in treatment room A and that in treatment room B is operated in the uniform scanning mode. 6 The relevant parameters of the HIMM facility for carbon ions are listed in Table 1. 7,8 Additionally, carbonion beams have higher relative biological effects (RBEs) that can enable a superior outcome for the treatment of tumors. 2 To give full play to the advantages of a carbon-ion therapy device, precise dosing and irradiation position monitoring as well as robust safety interlock are all necessary conditions. Therefore, a beam-monitoring system (BMS) and one kind of quality-assurance equipment (QAE) were developed to confirm the beam-delivery system sending the beam to the expected target. Research began at IMP in 1993, 9 and during the last several years a variety of monitoring systems with specific technical characteristics have been developed to meet the needs of carbon-ion therapy. 10,11 Currently, three types of detector systems with better performance have been developed by the beam diagnostic group of IMP for the HIMM.
In this paper, mainly the performance of the BMS and QAE for registration of medical devices with authorities is introduced, and using the BMS and QAE to ensure beam quality and patient safety throughout irradiation is discussed.

| SYSTEM DESCRIPTION
From the point of view of medical accelerator function, the HIMM is composed of two large systems: an acceleration system and a treatment system. The BMS and QAE are essential constituents of the treatment system that measures the beam characteristics in real time, deals with abnormal test results for patient safety throughout the treatment process, and verifies the beam parameters. According to the characteristics of the HIMM and the demands of the treatment system or physicians, three BMSs, one for each high-energy beam-transport line, are used for the HIMM; meanwhile, the QAE has been designed to shorten QA times. The simplified layout is shown in Fig. 1. The elements of the beam line are the BMS and a number of optional beam-modifying elements that are chosen depending on the beam-scanning mode. 10,11 The BMS is fixed in the beamline just behind the scanning magnets, whereas the QAE at the isocenter is portable and shared among different treatment rooms.

2.A | Beam-monitoring system
The BMS is composed of three integral ionization chambers (ICs) and two multistrip ionization chambers (MSICs). For better conformal irradiation, the beam monitors located in the modulated scanning technology are used for monitoring the position of each spot and the flux delivered to each spot online; once the prescribed doses are reached, the beam is steered to the next spot by the scanning magnets. The beam current is turned on before moving the spot to the next position, until the radiation on the layer is completed, and the beam is turned off. The beam monitors placed on the uniform scan- To reduce risk that may occur during the treatment process and meet IEC requirements for the treatment device, the BMS for each treatment line is divided into two completely independent subsystems for redundancy. The simplified schematic layout of the beam  Fig. 2, in which the blue part represents the main system and the green part the redundant system.
The figure shows the communication between the BMS and the graphical user interface of the treatment system or accelerated system. The signal detected by each monitor in the BMS is sent to the treatment system; then, the treatment system analyzes the signal to select the subsequent operation.

2.A.1 | Integral ionization chambers
The ICs of the BMS are used to monitor the beam flux online, integrating the electrons produced by the entire lateral beam. 12    2 . Schematic layout of BMS; the blue part represents the main system, while the green part represents the redundant system. The signal obtained from the monitors is transmitted to the accelerator system or treatment system through front-end electronics (FEC) and a data-acquisition system (DAQ) for beam commissioning or online monitoring during the treatment process. The MSICs must also be calibrated at different beam intensities, and thus the calibrated coefficient was added into the data-acquisition program for real-time correction.
The front-end readout circuit of the MSIC is based on the gated current integrated circuit developed by IMP, 15 which converts the vertical (strip X) and horizontal (strip Y) charge signal outputs from the MSIC to a differential voltage signal. The promoted properties of the front-end electronics are given in Table 2. The MSIC is con-  The data-acquisition system is based on NI CompactRIO (CRIO).

| PERFORMANCE TESTS
The performance tests of the BMS and QAE were conducted under different irradiation conditions during the preclinical beam-commissioning phase. The test results were found to satisfy the requirements of carbon-ion radiotherapy. Details of the main tests are presented in this section.

3.A | Off-axis response of ICs
The ICs must ensure that the beam intensity measured by the effective volume responds uniformly. To avoid overdose or underdose irradiation, the spacing uniformity between the electrode plates of the ICs is required to be better than 0.
where R max and R min are the respective maximum and minimum ratios of the dose measured by the PTW-Freiburg unit to the counts measured by the ICs synchronously; R is the average value of the ratios R. The off-axis value for 17 spots was found to be within the ± 2% tolerance for all ICs. For example, the relative dose response (R) tested by IC01 with a beam energy of 261.3 MeV/u in treatment room A is shown in [Fig. 5(b)], and the maximum differences between the maximum and minimum values of the ratio R are 1.98%. Since random errors are introduced in the testing process, the off-axis value should be better than this value.  Fig. 6(a)], in which the baseline drift caused by temperature change is nearly eliminated. Figure 6(b) shows that the minimum distinguishable amplitude is approximately 2.5 mV in 50 µs and the dose error caused by the signal below 2.5 mV is less than 0.06 cGy. Therefore, the beam parameters could be monitored by the MSIC at low dose rate, which reduced the effect on the patients.

3.C | Position resolution of MSICs
The position resolution of the MSICs is limited to the accuracy of the strip anode and the performance of the front-end electronics. It Since treatment accuracy refers to the difference between beamradiation location and plan location, it is worth mentioning that an MSIC used at the isocenter can improve treatment accuracy before treatment because the position resolution of the MSIC is the main influencing factor of treatment accuracy measurement.

3.D | Homogeneity of lateral beam profile
The homogeneity of the lateral beam profile is measured for QA by the MSIC or medical film for irradiating a specific treatment field. To confirm that the MSIC is good at measuring the homogeneity, an EBT3 medical film was attached to the front window of the MSIC to test the beam profile synchronously. The measured field profiles along the X and Y axes for a square field of 110 × 110 mm 2 irradiated with 2 × 10 7 carbon ions (330 MeV/u) and a ridge filter 17 are shown in Fig. 8. In our beam-delivery system, the ridge filter is employed as the range modulator.
It can be seen from the test results that the picket of the ridge filter measured by the MSIC is well consistent with the EBT3 (resolution 72 dpi) medical film; the difference is mainly found in the picket of the ridge filter, i.e., the maximum deviation is 8.8% after data normalization. Through consulting literature and data analysis, it was determined that the difference is mainly caused by the nonlinear relationship between the beam flux and the medical film absorbed dose. 18,19 The test results prove that the commonly used medical film can be replaced by the MSIC in measuring the homogeneity of the lateral beam profile. Furthermore, since the EBT3 image must be scanned and analyzed offline, the MSIC can achieve faster measurement than medical film; this feature is more conducive to clinical daily QA and for identifying and solving problems in a timely manner.

3.F | Beam-angle measurement
The beam angle can be obtained by measuring the relative beam positions along the beamline from the MSICs in a different location.
Thus, we define the beam angle as where R is the longitudinal difference between the beam position measured by the MSICs in a different location and L the distance between the two MSICs.
The beam angle was tested by irradiating a 7 × 7 grid of spots with the same spacing from each other, and the beam position was

3.G | Beam-range verification
Since a 1-mm-thick layer of PMMA was fixed in front of every PIC as an absorbing plate to simulate different water depths, the total of 11.66 cm of equivalent water thickness determines that the IC stack cannot measure the entire depth of the dose distribution. Therefore, different thicknesses of PMMA were added in front of the IC stack to make the peak position of the Bragg curves appear in the sensitive area of the chamber. As an example, for irradiation with 120-,  Fig. 11(a)]. Since the beam-energy loss varies slightly in the flat area, the loss of partial flat-area information has an ignorable effect on the measurement of depth dose distribution in the Bragg-peak region. The relationship between beam energy and the peak position of the Bragg curves tested by the IC stack can be seen in Fig. 11(b); compared with the measured results from a water tank, it is found that the linearity is basically the same.
The accuracy of depth dose distribution in the Bragg-peak region was tested by adding PMMA with a thickness of 0.2 mm layer by layer in front of the IC stack. With the increase in the number of 0.2-mm-thick layers of PMMA, the relative position of the Bragg peak changed; the changes in position can be calculated accurately by the weighted-mean-center method 21 . Figure 12 shows the relationship between the relative position of the Bragg peak and the increased PMMA thickness. It was determined that as the PMMA thickness increases the relative position decreases linearly. Since the approximate algorithm used to calculate the Bragg-peak position vs PMMA thickness is related to the depth dose distribution, the slopes of different energies are different, but the accuracy of the specific energy that is calibrated using a water tank is not affected. The linear deviation is less than 0.04 mm with a beam energy of 260 MeV and less than 0.02 mm with a beam energy of 330 MeV, which shows the superior performance of the IC stack.

| DISCUSSION AN D CONCLUSIONS
The BMS and QAE that were specially developed for the HIMM facility were installed and run successfully within the HIMM for two years. Taking full advantage of the BMS and QAE, the accuracy of dose delivery is better than IEC requirements, and the treatment efficiency has clearly increased. In particular, the beam oscillating within a range of approximately 0.35 mm caused by a scanning-magnet power source was found from the test results of MSIC; this was solved during the accelerator-commissioning process. The performance of the BMS and QAE was tested with different irradiation conditions during the preclinical beam-commissioning phase, and some examples are reported in this paper. The test results proved that the BMS and QAE satisfy the requirements of the carbon-ion radiotherapy with a high level of reliability. The BMS and QAE associate the treatment system of the HIMM to deliver the right dose to the correct position precisely. Furthermore, it simplifies the daily verification task.

CONF LICT OF I NTEREST
The authors have no conflict of interest to disclose.