Characterization of a shielded beam current transformer for ultra‐high dose rate (FLASH) electron beam monitoring and dose reporting

Abstract

Background

Real‐time beam monitoring and accurate dose reporting is challenging in ultra‐high dose rate (UHDR) electron beams. Although beam current transformers (BCTs) can effectively track parameters such as pulse width (PW) and repetition frequency for UHDR electron beams, recent work has highlighted their sensitivity to electric fields induced by transient charge buildup in irradiated media under UHDR conditions.

Purpose

This study evaluates the performance of a novel electrostatically shielded BCT for real‐time, high‐accuracy dose monitoring in UHDR electron beams.

Methods

Irradiations were conducted using the Mobetron linear accelerator configured for UHDR electron beams with energies of 6 and 9 MeV. A shielded BCT was implemented to monitor beam delivery, with dose calibration established using alanine dosimeters in solid water phantoms. Dose stability was assessed over short (7‐day) and long (16‐week) periods. The BCT's response to variations in PW, pulse number, and pulse repetition frequency was also evaluated to determine its robustness across beam configurations.

Results

The BCT showed high reproducibility and accuracy, with standard deviations of the difference between BCT‐predicted and alanine‐measured doses within 0.21% over short‐term measurements and 0.57% over long‐term measurements, even when subject to large (10%) machine output adjustments. When varying beam parameters, the BCT maintained accurate dose prediction within 1.0% and 1.4% of alanine measurements for 6 and 9 MeV, respectively, with high linearity (${\mathrm{R}}^2\ge$ 0.9997) across total doses.

Conclusion

Shielded BCTs provide a stable and accurate solution for real‐time dose monitoring in FLASH radiotherapy, demonstrating robustness against output fluctuations and beam parameter variations. While further calibration standardization is required, this study supports the feasibility of using shielded BCTs for reliable UHDR dose monitoring, facilitating safe and precise implementation of FLASH radiotherapy in preclinical and clinical settings.

1. INTRODUCTION

FLASH radiotherapy represents a novel paradigm in radiotherapy, utilizing ultra‐high dose rates (UHDR) typically exceeding 40 Gy/s to achieve effective tumor control while minimizing damage to healthy tissues. This sparing effect has been validated across diverse preclinical models, including rodents and larger mammals, and is hypothesized to involve mechanisms such as differential oxygen consumption, enhanced DNA repair, and altered immune responses in normal tissues. ¹ , ² , ³ Due to their wide availability, cost‐effectiveness, and reliable capacity to produce UHDR beams, electron accelerators have become the primary platform for treatment delivery in FLASH studies, utilizing dedicated UHDR linacs, adapted clinical linacs, and intraoperative systems to achieve the required dose rates. ⁴ , ⁵

The practical application of FLASH‐RT in a clinical setting requires accurate real‐time dose monitoring to ensure both precision and safety in high‐dose‐rate deliveries. Precise dose reporting is equally important in preclinical studies, where repeatability and precision are essential to assess and understand the FLASH sparing effect. However, conventional transmission ionization chambers, typically used as beam monitors, encounter significant limitations in UHDR electron beams due to ion recombination and saturation effects. ⁶ , ⁷ To circumvent these issues, Beam Current Transformers (BCTs) have been proposed and integrated in some UHDR electron linacs for real‐time, high‐fidelity monitoring of essential beam parameters such as number of pulses (NP), pulse width (PW), and pulse repetition frequency (PRF), while remaining unaffected by saturation. ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² Furthermore, because a BCT accurately measures the pulse charge—the number of electrons in a pulse that contribute to the dose—and since this charge is directly proportional to the delivered dose, it can, in principle, serve as a dose monitor. However, BCT implementation for dose monitoring remains challenging due to a lack of standardized calibration protocols and reported signal variations across different irradiation setups and beam parameters. ⁸ , ¹⁰ , ¹³ , ¹⁴ , ¹⁵

Our group recently identified transient charge buildup in nonconductive materials irradiated by UHDR electron beams (e.g., solid water, plastic phantoms or collimators, 3D‐printed devices, etc.) as the primary driver of these signal variations. ¹⁶ We demonstrated that the electric fields induced by this transient charge buildup significantly perturb the BCT signal, compromising the reliability of dose monitoring. Additionally, variations in parameters such as pulse length, NP, field size, source‐to‐surface distance (SSD), and phantom material exacerbate these signal discrepancies, further limiting the utility and reliability of BCTs to track beam delivery under UHDR conditions. To address these challenges, electrostatic shielding of the BCT has emerged as a promising corrective measure. ¹⁶ Results indicated that such shielding can mitigate the effects of charge‐induced electric fields, stabilizing BCT readings under varying irradiation conditions. While electrostatic shielding appears to be essential to stabilizing BCT performance, further investigations are required to demonstrate the dose accuracy of a shielded BCT under UHDR conditions.

The aim of this work is to characterize the performance of a shielded BCT configuration for real‐time monitoring in UHDR electron beam applications, focusing on stability and accuracy across varied beam parameters. Furthermore, it proposes and evaluates a calibration protocol tailored for the specific requirements of FLASH‐RT, with the ultimate aim of contributing to standardized protocols for clinical and research applications.

2. MATERIALS AND METHODS

2.1. Irradiation device

Irradiations were performed using the Mobetron (IntraOp, California), a mobile electron linear accelerator designed for intraoperative and dermatologic applications. Equipped with a research‐specific console, this system is capable of producing UHDR electron beams at nominal energies of 6 and 9 MeV and instantaneous dose rate above ${10}^6$ Gy/s. The UHDR console enables precise modulation of beam parameters, including PW, PRF, and the total NP, providing flexibility in UHDR beam configuration. ⁵

2.2. Shielded beam current transformer (BCT)

The shielded BCT implementation design used in this work is based on the one introduced by Pageot et al. ¹⁶ and shown in Figure 1. The BCT model used was an ACCT‐S‐082‐H from Bergoz (France), ¹⁷ mounted below the linac head using a custom 3D‐printed applicator made of polylactic acid (PLA). This design allowed for a stable yet removable mounting configuration, essential given the Mobetron's clinical use for intraoperative treatments at our institution (in conventional mode). Indeed, permanent modification of the linac head to integrate internal BCTs, as used in recent versions of the modified Mobetron, ¹⁰ , ¹⁸ was not possible for our machine as it is regularly used clinically in the intraoperative setting.

FIGURE 1.

Schematic of the shielded BCT setup for UHDR beam monitoring.

The applicator incorporated a 0.025 mm thick aluminium foil grounded to earth just downstream of the BCT that acted as an electrostatic shield, preventing the electric field generated by the transient charge buildup in the collimator and phantom to reach the BCT's aperture. The distal end of the applicator was designed to accommodate 4 cm thick Delrin collimators provided by IntraOp, with exchangeable apertures ranging from 2.5 to 6 cm in diameter (although only the 6 cm aperture was used in this study) yielding a SSD of 33.3 cm. Voltage outputs from the BCT were processed by its dedicated analog electronics module, which implements multistage feedback and amplification using low‐noise operational amplifiers. The resulting voltage signal was recorded with a CAEN DT5751 digital oscilloscope. Pulse data were automatically captured by the CAENscope software using a fixed 0.05 V trigger threshold. Each pulse was recorded over a 12.75 $\mu \mathrm{s}$ window, spanning 2.75 $\mu \mathrm{s}$ before the trigger to 10 $\mu \mathrm{s}$ after. The total signal per pulse was derived by integrating the BCT signal across this period after baseline subtraction of the first 2 $\mu \mathrm{s}$ using custom MATLAB scripts. A sensitivity of 10 V/A and standard uncertainty of 0.25% were considered for the BCT signal in this work, based on the manufacturer's specifications. It is also noted that only the UHDR beams were investigated in this work as the beam current in the conventional beam is too low to be detected with our BCT model.

2.3. Experimental measurements and data analysis

2.3.1. BCT calibration

An initial calibration of the BCT was performed to establish a correspondance between the total charge measured by the BCT to the absorbed dose to a point downstream of the BCT, analog to the calibration of transmission chambers in a conventional linac. Alanine dosimeters were chosen as reference detectors for the BCT calibration due to their high stability, tissue‐equivalent response, and accuracy in measuring absorbed dose across a range of dose rates, including UHDRs. ¹⁹ While these features make alanine dosimetry attractive for dose measurements of UHDR beams, a more important benefit is their direct traceability to primary standards for absorbed dose that is independent of any other calibrated detector, such as ionization chambers.

Alanine dosimeters work by forming stable free radicals when irradiated, with the absorbed dose being proportional to the concentration of these free radicals. The free radical concentration can be measured using an electron paramagnetic resonance (EPR) spectrometer. The selection of alanine for calibration of the BCT in this work was supported by its widespread use in industrial and medical applications, such as food irradiation, pharmaceutical sterilization, and radiotherapy, owing to its lack of dose‐rate dependency over an extensive range of rates and its linear response across doses from 2 Gy to 5 kGy. ²⁰ , ²¹ Additionally, alanine has demonstrated excellent performance at UHDR levels ²² , ²³ , ²⁴ with clinical uncertainty of less than 1%, validated against primary standards such as water calorimeters. ¹¹

The alanine dosimeters used for this work consisted of seven 60 mg alanine pellets (L‐Alanine with 9.1% paraffin wax binder, Harwell Dosimeters; Oxfordshire, UK) in specially crafted Polyether ether ketone (PEEK) holders. A photograph of a sealed holder, along with an open holder loaded with alanine pellets is shown in Figure 2. These holders were designed to match the external dimensions of a PTW Advanced Markus chamber, allowing consistent and reproducible placement within a solid water phantom. The holders are sealed with an O‐ring, minimizing environmental effects during transport and reducing the potential for contamination of the alanine pellets. The calibration and readout of the alanine pellets was performed by a primary standards laboratory, as detailed in Section 2.3.4. While we acknowledge that such calibration procedures might not be readily achievable for all users, our primary objective was to leverage the well‐established advantages of alanine dosimetry for validating the performance of the shielded BCT. Future work is needed to standardize BCT calibration protocols, potentially incorporating alternative dosimeters based on recommendations from relevant task groups.

FIGURE 2.

Alanine dosimeters used for dosimetric calibration.

The initial BCT calibration was done by performing three consecutive irradiations of alanine pellets in their PEEK holder with concurrent BCT beam monitoring for both beam energies. The reference conditions for the BCT calibration are presented in Table 1 and were chosen to achieve a total dose of approximately 25 Gy, at a depth of low dose gradient using the same beam parameters for both 6 and 9 MeV. For both energies, the dose per pulse and mean dose rate were thus approximately 2.5 Gy and 150 Gy/s respectively. The experimental setup is shown in Figure 3.

TABLE 1.

Reference conditions used to calibrate our BCT.

Parameter	Value
Phantom material	Solid water
Depth (mm)	6
Number of pulses	10
Pulse width ($\mu \mathrm{s}$)	2.67
Pulse repetition frequency (Hz)	60
Nominal Energies (MeV)	6, 9

FIGURE 3.

Experimental setup for the BCT calibration and evaluation of its short‐ and long‐term stability. The PEEK holder, containing the alanine pellets, is placed within a machined solid water slab.

From these measurements, a calibration coefficient $X_{calib}$ was obtained for each energy by dividing the mean dose measured by the alanine pellets, ${\overline{D}}_{alanine}$, by the total charge measured by the BCT, $Q_{BCT}$:

$$X_{calib}=\frac{{\overline{D}}_{alanine}}{Q_{BCT}}.$$

The final calibration coefficient for each energy was obtained by taking the average of three calibration measurements. Once calibrated, the BCT could then be used to report the dose delivered in UHDR conditions by multiplying the calibration coefficient to the total charge measured, or monitor the dose rate by multiplying the calibration coefficient to the current reported by the BCT. The short‐ and long‐term stability of this calibration as well as its robustness against variable beam parameters was evaluated as detailed below.

2.3.2. Short‐ and long‐term stability

The stability of the BCT's calibration was assessed by reproducing the reference conditions and comparing the dose reported by the calibrated BCT to the one measured with alanine. This evaluation was performed daily for the first 7 days (short term stability), followed by weekly measurements for sixteen weeks (long‐term stability). An initial warm‐up of the machine using 500 MU delivered with the conventional dose rate 9 MeV beam (note that the Mobetron is equipped with two monitor chambers used in conventional mode to monitor beam output) was performed before each daily and weekly measurement.

2.3.3. Influence of pulse width, number of pulses, and pulse repetition frequency

The BCT's response to varied beam configurations was systematically evaluated by adjusting independently the PW, NP and PRF. The PW was varied between 2.4 and 3.5 $\mu \mathrm{s}$, NP between 10 and 18 pulses and the PRF between 10 and 90 Hz. These parameters were selected to maintain a total dose above 20 Gy (necessary to ensure accurate alanine results), while varying only one parameter at a time from the calibration conditions. This approach allowed for a comprehensive characterization of the BCT's sensitivity to each parameter, enabling validation of its accuracy under diverse operational settings relevant to clinical and preclinical FLASH applications.

2.3.4. Alanine calibration, readout, and uncertainty analysis

The alanine calibration and readout were performed at the National Research Council (NRC) of Canada. Holders were mailed to the Centre Hospitalier de l'Université de Montréal (CHUM) for UHDR irradiations in batches of 6 to 24 dosimeters, always including unirradiated and irradiated control samples as detailed below.

Alanine dosimeters were calibrated at NRC in the same PEEK holders used for the BCT calibration. Calibration irradiations of alanine dosimeters were performed in an 8 MeV electron beam from the NRC Elekta Precise conventional clinical accelerator with a field size of $10\times 10{\mathrm{cm}}^2$ shaped with an applicator. By calibrating alanine in the same holders in beams with similar energies as those for BCT calibration, the impact of the holder on the results can be considered negligible.

The dose per monitor unit (MU) was determined with a laboratory maintained secondary standard PTW 30013 ionization chamber calibrated directly against NRC primary standards for absorbed dose in electron beams ²⁵ , ²⁶ Two other secondary standard chambers, a PTW Roos and an NE2571, were used to verify the results.

Alanine dosimeters and ionization chambers were positioned at $d_{\mathrm{ref}}$ in a water phantom with a horizontal beam geometry. ²⁷ Water‐equivalent scaling was used to determine the shift of the effective point of measurement of the alanine dosimeter, accounting for the density of the PEEK entrance window and alanine pellets. Cylindrical ionization chambers were positioned with their central axes at $d_{\mathrm{ref}}$. The PTW Roos chamber was positioned with a shift of the effective point of measurement of 0.126 cm, which is the standard shift for this chamber used at NRC and is close to the value determined with water‐equivalent scaling of the front window.

Preirradiation of ionization chambers was performed with a dose of 10 Gy. To determine dose per MU, a set of at least five 200 MU (approximately 2 Gy) irradiations were delivered to the chambers while collecting charge. Chamber readings were corrected using the factors described by the electron beam Addendum to TG‐51. ²⁸ After determining dose per MU with the secondary standard chamber, alanine dosimeters were irradiated to doses of 20, 30, 40, 60, 80 and 100 Gy. A repeat irradiation was performed with a dose of 20 Gy.

A set of five control alanine dosimeters were also irradiated to 25 Gy to track spectrometer reproducibility and any impact of travel over the course of the study. Three of the control dosimeters remained at NRC and two were sent to CHUM and returned to NRC with every shipment of alanine dosimeters. In addition, unirradiated controls were sent and returned with the alanine dosimeters for background subtraction. All controls were transported in sealed holders.

Readout of alanine dosimeters was performed following the same protocol for alanine dosimeter calibration as well as for BCT calibration. The readout of alanine pellets was performed with the NRC Bruker EPR spectrometer as described by Surensoy. ²⁶ Pellets are read individually and normalized to the signal from a permanently fixed ruby obtained immediately after readout of the alanine pellet, which is used to account for environmental impacts on the spectrometer. The alanine signal is normalized by the mass of the pellet. The temperature during irradiation has an impact on free radical production, so temperature is corrected to a reference temperature of 21. ²⁹

Absorbed dose can be determined using alanine dosimeters with an uncertainty better than 0.8% (k = 1). ²⁶ , ²⁹ This estimate includes components from the primary standard dose measurement and the calibration of alanine pellets, which dominate the uncertainty budget. The overall uncertainty estimate of 0.8% is relevant for the initial calibration of the BCT. However, in this study we were mainly interested in the stability of the BCT calibration with time, which is monitored using alanine dosimeters. The primary standard dose measurement and calibration measurement therefore do not need to be considered, and what is primarily of interest is the type A uncertainty of the readout of alanine pellets as well as the spectrometer reproducibility over the time period of interest.

The type A uncertainty of the alanine readout consists of components related to fitting the alanine spectrum, fitting the ruby spectrum, the mass and temperature corrections to the signals, and the variability of the results among the 7 pellets from a given holder. The dominant component is the variability amount the 7 pellets, which is calculated as the standard deviation of the mean. This type A uncertainty on the alanine signal was equal to 0.3% on average for all measurements performed in this work, with no significant differences between 6 and 9 MeV.

The component related to the spectrometer reproducibility can be evaluated using the control dosimeters as well as quality assurance measurements made on the spectrometer over the time period of the study. An analysis of the results obtained with the control dosimeters and the QA measurements allows an uncertainty estimate of 0.3% related to spectrometer reproducibility.

3. RESULTS

3.1. Short‐and long‐term dose tracking

The short‐ and long‐term dose tracking of the BCT calibration are presented in Figure 4a,b, respectively. The BCT calibration was performed 1 day prior to the start of the short‐term measurements (i.e., on Day 0), ensuring that the initial differences observed on Day 1 reflect the system's immediate performance rather than a calibration error. Error bars indicate the standard measurement uncertainty based on the uncertainty provided by the manufacturer for the BCT signal (0.25%) and the uncertainty of the alanine readout, representing the Type A uncertainty of the alanine readout for each irradiation (ranging between 0.18% and 0.47%, with an average of 0.3%) and the EPR spectrometer reproducibility over the course of the study (0.3%). The mean and standard deviation of the relative differences between the dose measured by the alanine and the one predicted by the calibrated BCT are reported in Table 2.

FIGURE 4.

(a) Short‐ and (b) long‐term tracking of the BCT calibration.

TABLE 2.

Short‐ and long‐ term stability of the shielded BCT calibration.

	Short term stability		Long term stability
Beam energy	Mean (%)	STD (%)	Mean (%)	STD (%)
6 MeV	−0.20	0.21	0.17	0.42
9 MeV	−0.33	0.12	0.02	0.57

Within the short term analysis, the output of the machine (i.e., the dose‐per‐pulse) was within 2% and 1% for the 6 and 9 MeV beams, respectively, according to the charge measured by the BCT. Despite these machine fluctuations, the dose predicted by the calibrated BCT was within 0.5% of the alanine measurements for both energies, indicating stability and reliability of the BCT dose calibration over several days. The difference between alanine and BCT was in fact outside of their combined uncertainty for only 1 out of the 14 measurements done during the short‐term tracking. This is lower than what is expected for a measurement uncertainty of k = 1 (around 30% of the measurements should be outside of the combined uncertainty), suggesting that some uncertainties might have been overestimated in this work.

For the long term tracking, a few elements need to be considered. In Week 10, a maintenance intervention on the linac by the manufacturer resulted in an output change of about 10% for both energies. Although adjustments were made to restore the original beam energy characteristics (mainly the depth of 50% and 80% isodoses), the dose per pulse differed from the pre‐intervention values. Importantly, the BCT calibration was not updated following the intervention. Nevertheless, the dose reported by the BCT stayed within 1% of the dose read by the alanine afterwards, which convincingly demonstrates that the BCT can accurately capture abrupt output changes–such as the unscheduled 10% shift induced by the intervention. The BCT's ability to retain dose accuracy post‐intervention suggests that it may offer reliable long‐term performance without frequent recalibration, an important characteristic for future clinical applications where machine maintenance or configuration changes are routine.

The other element worth attention is the measured outlier on week 11 at 9 MeV. For this irradiation, the third pulse recorded by the BCT showed an odd shape, as shown in Figure 5, suggesting a transient dysfunction of the machine. This hypothesis is supported by the fact that the dose predicted by the BCT (including the incomplete third pulse) for this acquisition was within 1.2% of the one read by the alanine. While the exact cause for this misbehavior of the machine is unknown, this result highlights the potential of BCT beam monitoring to capture machine errors and adequately report the delivered dose in such conditions.

FIGURE 5.

BCT signal for the 9 MeV irradiation on Week 11.

3.2. Effect of pulse width, number of pulses and pulse repetition frequency

The difference between the dose predicted by the calibrated BCT and that measured by alanine as a function of PW, NP, and PRF is presented in Figure 6. As expected, discrepancies between measured doses and predictions generally increase as beam parameters diverge further from the calibration conditions.

FIGURE 6.

Relative difference in dose predicted by the calibrated BCT and that measured by the alanine as a function of variable (a) pulse width, (b) number of pulses, and (c) pulse repetition frequency. The reference condition for each beam parameter is identified with a vertical dashed red line.

Previous studies have shown that all three beam parameters considered here affect the effective energy of the beam. ⁵ , ¹⁰ , ³⁰ Since the BCT is insensitive to changes in beam energy, shifts in electron spectrum could impact the dose delivered to the reference point, even if the total beam fluence measured by the BCT remains constant. For our machine, and within the same timeframe as these experiments, beam energy was observed to increase with PW, with a greater rate of increase for the 9 MeV beam compared to the 6 MeV beam. ³¹ This behavior aligns with the results shown in Figure 6a, where the BCT underestimates the dose measured by alanine for longer PWs, particularly for the 9 MeV beam.

Figure 7 presents the linearity between the dose predicted by the BCT and that measured by alanine over a wide range of total doses achieved by independently varying PW and total NP. Data points reported in this figure are the same as the ones in Figure 6a,b, but expressed in terms of total dose delivered rather than pulse parameter values. Coefficients of determination (adjusted‐R ) were 0.9999 and 0.9997 for 6 and 9 MeV, respectively, again supporting the reliability of UHDR electron dose reporting using a shielded BCT. The linearity of the BCT signal as a function of the dose was further tested for doses below 20 Gy using a plastic scintillator, as reported in Appendix A.

FIGURE 7.

Dose predicted by the calibrated BCT versus measured by the alanine as a function of the total dose delivered for UHDR beams of (a) 6 MeV and (b) 9 MeV. Error bars not shown as they are substantially smaller than the symbols.

4. DISCUSSION

In this study, we comprehensively evaluated the performance of an electrostatically shielded BCT in tracking and reporting dose delivery in UHDR electron beams, focusing on both short‐ and long‐term stability as well as its response under varying beam parameters. By comparing the dose predicted by the calibrated BCT to that measured by alanine, our results demonstrate reliability of the calibration over time as well as its robustness against beam fluctuations and operational changes.

The short‐ and long‐term stability of the BCT calibration observed in this study demonstrates noticeably lower variability compared to previous reports using non‐shielded BCTs. Using the same linac and a similar external BCT implementation (albeit non‐shielded) as ours, Oesterle et al. ⁸ reported a standard deviation in the charge‐dose relationship of approximately 2% over 5 days and 3% over 3 months. Similarly, Jorge et al. ⁹ reported short‐ and long‐term stability around 1.5% using a non‐shielded BCT and the Oriatron eRT6 linac. In contrast, this study achieved standard deviations in the charge‐dose relationship around 0.2% for short‐term and 0.5% for long‐term stability. While some of this reduction in variability could potentially be attributed to using alanine as the reference detector rather than an Advanced Markus Chamber, it is likely that the large majority of the stability improvement resulted from the electrostatic shield, which minimized BCT signal perturbations caused by transient charge build‐up in the irradiated media. The fact that Jorge et al. reported a much lower variability in the charge‐dose relationship than Oesterle et al. using twice the SSD further supports this hypothesis, as the perturbation of BCT signal due to the transient charge buildup in medium was shown to decrease with larger SSD. ¹⁶

The stability of the shielded BCT calibration was also evaluated by varying beam parameters. Since variation of the pulse parameters are known to affect the beam energy, BCT calibration should ideally use beam parameters that closely match expected clinical or preclinical conditions. Despite this, the BCT's dose prediction accuracy reported in this work remained within 1.0% and 1.4% of alanine measurements for 6 and 9 MeV beams, respectively, when varying independently the PW, PRF and the total NP. This high level of agreement across a range of beam configurations suggests that, while matching calibration conditions to operational settings is ideal, the shielded BCT demonstrates sufficient robustness to maintain dose tracking accuracy within reasonable parameter variations. This characteristic supports its suitability for FLASH‐RT applications, where flexible beam configurations may be necessary for individualized treatment planning.

The reliability of the BCT across a wide range of delivered doses was demonstrated by a strong linearity ($R^2\ge 0.9997$) between the BCT predictions and alanine results for absorbed doses ranging from 25 to 45 Gy. This finding further supports the suitability of the shielded BCT for dose monitoring in preclinical applications, where broad dose‐response curves are often required. Compared to results obtained using a non‐shielded BCT under similar conditions, this again represents a substantial improvement. Oesterle et al. reported deviations as high as 4.10% and 5.61% for 6 and 9 MeV UHDR electron beams, respectively, with variable beam parameters. ⁸ Here again, this improvement is most likely due to the electrostatic shield, as our previous work has shown that BCT signal perturbations from phantom charge build‐up are influenced by both the total NP and the pulse length and can reach as high as 14% for our specific setup, ¹⁶ similar to the one of Oesterle et al. ⁸

While the results presented in this work demonstrate the strong potential of shielded BCTs for accurately tracking and reporting dose and dose rate in electron FLASH‐RT, certain limitations remain. Firstly, several UHDR linacs are equipped with internal, unshielded BCTs integrated into the linac head. Although non‐shielded BCTs could theoretically achieve comparable accuracy if placed at a large distance from electrically grounded and conductive phantoms, our previous work has shown that even distances of several tens of centimeters do not fully eliminate perturbations. ¹⁶ Moreover, many UHDR linacs–such as the Mobetron, S.I.T system, and FLASHknife–use plastic components for beam collimation, which are prone to charge accumulation and have been demonstrated to perturb the BCT reading, even without a phantom. ¹⁶ Therefore, retrofitting linac heads with self‐shielded BCTs or introducing an electrostatic shield upstream of plastic collimators appears essential for accurate dose monitoring in nearly all foreseeable scenarios.

Secondly, BCTs suffer from signal droop, a gradual decline in signal strength over time during a pulse induced by the core's magnetization build up in the presence of a constant beam current. However, the impact of this effect is likely negligible in most scenarios, as the ACCT model used in this work has a droop of only 0.66%/ms, and UHDR electron pulses typically have durations on the order of microseconds, making the associated signal loss (and the added uncertainty for variable pulse durations) minimal.

Lastly, BCTs are unable to measure some essential beam parameters, including energy, flatness, and symmetry, which limits their ability to confirm stability in these parameters and may affect dose accuracy. To address this issue, validation of beam energy, flatness, and symmetry with an additional device before delivery could be implemented. ³² , ³³ Alternatively, integrating complementary real‐time monitoring detectors might offer a solution for tracking these parameters. ³⁴ , ³⁵ Another important consideration is that, although the BCT directly measures the beam signal, its calibration under reference conditions must be transferred to the different irradiation setups via accurate dose calculations, ideally using a validated Monte Carlo beam model. ³¹ , ³⁶ Similarly to transmission chambers in conventional radiotherapy, this step is crucial to ensure the accuracy of the dose report by accounting for variations in field size, SSD, and medium heterogeneity, thereby ensuring the reliability of BCT dose reporting in preclinical and clinical applications. Although alanine dosimetry was employed in this work to validate calibration stability, its delayed readout precludes daily verification in a clinical setup. In practice, dosimeters such as the flashDiamond ¹³ , ¹⁹ or ultra‐thin ionization chambers ³⁷ could enable immediate dose measurement under UHDR conditions, thus offering a practical means for rapid BCT calibration checks.

5. CONCLUSION

This study demonstrates for the first time the potential of electrostatically shielded BCTs as reliable, real‐time dosimetry tools for UHDR electron beam applications in FLASH radiotherapy. By establishing a robust calibration using alanine dosimeters and evaluating the shielded BCT performance under variable pulse parameters, this study shows that the shielded BCT can provide stable and accurate dose measurements over both short and long timescales. The device's sensitivity to transient machine dysfunctions and its ability to maintain dose fidelity despite moderate variations in beam parameters underscore its potential utility in preclinical and clinical settings.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

APPENDIX A. DOSE RESPONSE LINEARITY

The linearity of the shielded BCT response as a function of dose was evaluated for doses below 20 Gy using a plastic scintillation detector (PSD, RP200, MedScint, Qc) previously validated for relative dosimetry in UHDR electron beams. ³⁸ The PSD was first calibrated in terms of dose against the alanine in our reference conditions (Table 1) using the shielded BCT reading to correct for beam output fluctuations between the two conditions. Then, the dose for irradiations ranging between 1 and 10 pulses was recorded by the PSD while monitoring the beam with the shielded BCT. Results are shown in Figure A1, indicating a strong linearity ($R^2\ge 0.9999$) of the calibration factor for doses below 20 Gy.

FIGURE A1.

Dose predicted by the calibrated BCT versus measured by the Plastic Scintillation Detector (PSD) as a function of the total dose delivered for UHDR beams of (a) 6 MeV and (b) 9 MeV. Error bars not shown as they are substantially smaller than the symbols.

Bernelin T, Muir B, Renaud J, et al. Characterization of a shielded beam current transformer for ultra‐high dose rate (FLASH) electron beam monitoring and dose reporting. Med Phys. 2025;52:e17927. 10.1002/mp.17927