Introduction
Trigeminal neuralgia (TN), also known as tic doulourux, is a severe paroxysmal facial pain occurring within the trigeminal distribution of the facial region.Reference Lunsford and Sheehan1 This condition is characterised by sudden, electric shock-like pain that is localised to one or more divisions of the trigeminal nerve and has sudden onset and termination.Reference Obermann2 Initial therapeutic intervention involves medical approaches as the primary treatment line. Nevertheless, the efficacy of these medications can diminish, leading to potential side effects. Surgical interventions such as microvascular decompression, radiofrequency rhizotomy, glycerol rhizotomy and balloon compression carry inherent risks, including bleeding, cerebrospinal fluid leakage and infection.Reference Esparza-Moreno, Garcáa-Garduño, Ballesteros-Zebadúa, Lárraga-Gutiérrez, Moreno-Jiménez and Celis-López3 Another alternative is stereotactic radiosurgery (SRS), which is the least invasive procedure and has demonstrated significant pain relief with minimal side effects.Reference Maesawa, Salame, Flickinger, Pirris, Kondziolka and Lunsford4
Historically, Gamma Knife-based (Elekta, Stockholm, Sweden) treatment has been deemed suitable for SRS treatment and is considered the gold standard in terms of sparing normal healthy tissue.Reference Tuleasca, Régis and Sahgal5 However, it has been shown that dedicated linear accelerators (linac) can achieve effectiveness comparable to that of Gamma Knife. TrueBeam STx (Varian Medical System, Palo Alto, CA, USA) is a linear accelerator capable of delivering SRS treatment with high efficacy and safety. This equipment can deliver flattening filter-free (FFF) beams that increase dose rates and reduce scatter, leakage and out-of-field doses.Reference Hrbacek, Lang and Klöck6 In addition, the linac is integrated with a high-resolution multileaf collimator (HD-MLC), 6D robotic couch and the possibility of acquiring high-resolution cone-beam computed tomography (CBCT). When integrated with Novalis (Brainlab, Munich, Germany) technology, the linac can utilise optical/thermal imaging and two oblique planar stereoscopic kilovoltage (KV) images for patient positioning, monitoring and tumour targeting through the ExacTrac Dynamic system (ETD-Brainlab).Reference Da Silva Mendes, Reiner and Huang7 The dose generated by the KV imaging is considered negligible compared to the therapeutic radiation dose.Reference Murphy, Balter and Balter8 Furthermore, Brainlab incorporates a frameless radiosurgery system for patient immobilisation, demonstrating high detection accuracy and sub-degree positioning accuracy compared to frame-based techniques.Reference Gevaert, Verellen and Tournel9
Conventionally, TN linac treatment employs 6 MV beams with either cones or MLCs.Reference Rashid, Pintea, Kinfe, Surber, Hamm and Boström10,Reference Stevens, Lobb and Yenice11 Moreover, research has shown the efficacy of 6 MV-FFF beams for TN SRS, indicating that increased dose rates can shorten treatment times and reduce patient discomfort.Reference Mhatre, Chadha, Kumar and Talapatra12 In this study, a 28% reduction in treatment time, along with lower doses to organs at risk (OARs), was achieved with 6 MV-FFF compared with 6 MV-SRS.
Initially, the use of 10 MV-FFF does not seem easy to implement because it generates a higher lateral dose, as it can produce more lateral electrons than 6 MV-FFF for a single field.Reference Oh, Kang, Yea, Kim and Oh13 However, this effect can be mitigated by integrating different arcs, allowing the use of 10 MV-FFF due of its ability to produce a higher dose rate. To our knowledge, no studies to date have compared the dosimetric characteristics of 10 MV-FFF and 6 MV-FFF beams for TN SRS using cones.
The objective of this study was to compare the dosimetric properties of 10 MV-FFF beam versus 6 MV-FFF beam for TN SRS, aiming to determine whether the higher energy beam offers potential benefits in terms of treatment time reduction, patient comfort and dose distribution. By exploring these differences, this study adds a new perspective to the ongoing debate on the most effective beam energy for TN SRS treatment, with implications for clinical practice and patient care.
Methods and Materials
Linac technology overview
TrueBeam Novalis STx, developed by Varian and Brainlab, is a high-precision linear accelerator system specifically designed for image guidance stereotactic treatments targeting both cranial and extracranial regions. It is equipped to deliver treatment using conical stereotactic beams, which are particularly effective in TN treatment. At our institution, we used a 6 MV-FFF beam for SRS TN treatment, delivering a dose rate of 1400 MU/min through a 4 mm stereotactic cone. The prescribed dose for these treatments is 90 Gy at the isocentre.
Mechanical precision
The accuracy of the linac isocentre was verified before every TN treatment using the Winston-Lutz (WL) test.Reference Schell, Bova and Larson14 This test ensures precise alignment of the isocentre and is carried out automatically by running an XML file in developer mode, taking approximately four minutes per session.Reference Alarcon and Venencia15 A total of 16 images obtained using different combinations of gantry, collimator and couch angles with a portal imaging device (EPID) were analysed using RIT (Radiological Imaging Technology, Colorado Springs, CO, USA) software to confirm submillimetre accuracy.
Image guidance procedure
Following the completion of the WL test, the ExacTrac Dynamic IGRT (image-guided radiation therapy) system was calibrated. This calibration ensured an accurate correlation between the thermal and 3D camera images, X-ray correction images and alignment of the isocentre with both the surface tracking and X-ray systems. Moreover, the quality assurance of the entire IGRT system follows the TG-142 protocol.Reference Klein, Hanley and Bayouth16
The approved treatment plan was transferred to both the ARIA v15.6 (Varian) and the ExacTrac Dynamic (ETD) systems. The ETD settings were adjusted to align with the clinical protocol for TN treatment, incorporating a spatial tolerance of 0.5 mm and an angular tolerance of 0.5° along with X-ray verification throughout the procedure. To improve the visibility of bone structures, kV and mA levels were optimised, and specific OARs were designated for monitoring using digital reconstructed radiographs (DRRs).
In addition, 3D visualisation was fine-tuned to improve the overall quality of the DRR images. Prior to the start of treatment, the patient was positioned using the 3D surface camera of the ETD system, with a region of the patient’s face selected for continuous monitoring. Dual X-ray images from the ETD system were acquired and fused with the DRR to confirm patient positioning. CBCT was performed to verify patient alignment and laterality. Throughout the treatment period, X-ray images were obtained to ensure accurate patient positioning for all couch angles.
Treatment plan
Eleven patients previously treated with TN radiosurgery were selected and anonymised. Table 1 summarises the characteristics of the patients, and the clinical response to treatment for all patients was evaluated according to the Barrow Neurological Institute (BNI) scale,Reference Rogers, Shetter, Fiedler, Smith, Han and Speiser17 as shown in Table 2. All plans used 6 MV-FFF beams and were planned with Elements Cranial v4.0 TPS (Brainlab, Munich, Germany). Plans were re-planned with 10 MV-FFF beams using the same dose prescription and isocentre position. Plan dose comparisons were performed based on dose gradient (distance from the isocentre to the 45 Gy isodose), volumes at 20 Gy and 10 Gy, maximum brainstem dose, and dose to 10% of the brainstem.
Characteristics of patients treated with 6 MV-FFF beam
BNI pain intensity score
The patient simulation included CT imaging from Somatom Go. Up (Siemens Healthineers, Erlangen, Germany) with 0.6 a slice thickness. The patients were immobilised using a frameless SRS mask fixation system (Brainlab, Munich, Germany). The patients underwent to high-resolution magnetic resonance imaging (MRI) on the T1, T2, T2-weighted and FIESTA scans with a slice thickness of 0.5 mm.
The MRI images underwent rigid fusion with the CT reference images, followed by a cranial distortion correction using Elements (Brainlab). OARs were delineated using the Anatomical Mapping on Elements, which employs a Synthetic Tissue Model segmentation algorithm.Reference Krüger, Kurtev-Rittstieg, Kägi, Naseri, Hägele-Link and Brugger18 The radiation oncologist conducted the final revision of the anatomically contoured OARs.
Trigeminal neuralgia treatment plan technique
Treatment plans were performed using the Elements Cranial SRS cone v.4.0 (Brainlab) treatment planning system (TPS) with 12 non-coplanar arcs of 110° with couch spacing between 10° and 20°, a 4 mm conical cone and a 6 MV-FFF beam, as shown in Figure 1. Modifications were made to the start and stop angles of the arcs to avoid entry through the eyes.
A treatment plan for TN involving 12 non-coplanar arcs on Elements Cranial Cone.
Determination of the isocentre at the TN was performed by an experienced team, including a medical physicist, neurosurgeon and radio-oncologist. The planned isocentre was positioned at the retrogasserian segment of the trigeminal nerve within the cisternal space, and the final location was determined to have a brainstem maximum dose of less than 25 Gy. The pencil Beam algorithm with heterogeneity correction was used for dose calculations with a 1 mm3 grid size.
Patient-specific quality assurance (PSQA) was performed using an independent monitor unit (MU) verification for each arc.Reference Stern, Heaton and Fraser19 An Excel spreadsheet was developed specifically for MU verification, considering dosimetric parameters such as dose calibration for the corresponding energy, tissue maximum ratio (TMR), scatter factors and absorbed dose at the depth of interest. A tolerance limit of 3% was considered acceptable according to the institutional criteria. Additionally, during the commissioning process, measurements were performed using EBT3 radiochromic films to validate the accuracy of the dosimetric calculations.Reference Vacca, Caussa, Filipuzzi, Garrigo and Venencia20
To calculate treatment plans with a 10 MV-FFF beam, a new machine was created in the TPS, and dosimetric measurements were carried out for stereotactic cones of 4, 5, 6 and 7.5 mm. The TMR, scatter factors (SFs) and dose profiles were measured at a dose rate of 2400 MU/min. According to the manufacturer’s guidelines, the jaw position was set at 14 × 14 mm2.
TPS beam modelling
Nominal linac output
The reference nominal linac output (NLO) for 10 MV-FFF was 0.744 Gy/100 UM, and for 6 MV-FFF, it was 0.657 Gy/100 UM, both for a 10 × 10 cm2 field size, 1000 mm source surface distance (SSD) and a depth of 100 mm.
Tissue maximum ratio
The TMR was measured using a Razor diode within a Blue Phantom (IBA Dosimetry, Schwarzenbruck, Germany) water tank. The diode was positioned parallel to the beam axis at the isocentre, with the active area perpendicular to the beam at the effective point of the measurement. Prior to the measurements, the coincidence between the axis of the radiation beam and the detector was examined by evaluating the alignment of the lateral and vertical profile doses at 5 cm and 20 cm depths. The phantom was filled with water during measurements. TMR measurements were acquired every 2 mm from 0 mm to 25 mm depth and every 5 mm from 25 mm to 250 mm depth.
Scatter factor
Scatter factors were measured using PTW 60012 (PTW-Freiburg, Freiburg, Germany) and Razor diodes (IBA Dosimetry, GmbH, Schwarzenbruck, Germany) in conjunction with radiochromic EBT3 films (Ashland, Wilmington, DE, USA) following the TRS-483 protocol. The measurement set-up with diodes was at isocentre in the Blue Phantom, with an SSD of 900 mm and a depth of 100 mm. For all cones, the jaws were set to a 14 × 14 mm2 field size, according to the Brainlab technical reference.21 Diode measurements were performed using the daisy chain method with an intermediate field size of 30 × 30 mm2. A CC04 ion chamber (IBA) was used for measurements of intermediate and 10 × 10 mm2 field sizes. For each measurement, Equation (1) was applied to obtain the scatter factor and correction factors
$k_{Qclin,Qref}^{fclin,\;fref} = k\;$
were applied to correct the scatter factor for each diode. Table 3 shows the correction factors
$k_{Qclin,Qref}^{fclin,\;fref} = k$
for the PTW 60012 and Razor diodes for 10 MV-FFF, using the field size obtained from the measured profiles with the corresponding equivalent square field size Sclin:
$$\small {{S_t}\left( c \right) = {{D_{HDR}^{Isoc}\left( {c,r = 0,{d_{meas}}} \right)} \over {D_{HDR}^{Isoc}\left( {30x30\;m{m^2},r = 0,{d_{meas}}} \right)}}{\rm{*}}{{D_{IC}^{Isoc}\left( {30x30\;m{m^2},r = 0,{d_{meas}}} \right)} \over {D_{IC}^{Isoc}\left( {100x100\;m{m^2},r = 0,{d_{meas}}} \right)}}}$$
Field output correction factors
$k_{Qclin,Qref}^{fclin,fref} = k$
for 10 MV-FFF using PTW 60012 and IBA razor diodes, following the TRS-483 protocol
where
${S_t}$
is the total scatter factor,
$c$
is the cone diameter at isocentre (mm),
$r$
is the radial distance from the central axis to the point of interest (mm),
${d_{meas}}$
is the depth of point in tissue (mm) and
$D_{HDR}^{Isoc}$
and
$D_{IC}^{Isoc}$
are the dose readings for the detector and ionisation chamber, respectively.21
The EBT3 films were uniformly cut to the required size and positioned between solid water slab phantoms (RW3 slab, PTW, Freiburg, Germany) to ensure consistent measurements. The films were placed at isocentre level, with 10 slabs of 10 mm thickness each, forming a total depth of 100 mm. An additional 50 mm slab thickness was used to generate backscatter. For each measurement, 1000 MU was delivered. All films were obtained from the same batch, and the recommended handling procedures were strictly followed to maintain accuracy and consistency.Reference Howard, Herman and Grams22 Prior to irradiation, the films were carefully handled under low-light conditions to avoid premature exposure, and gloves were worn to prevent contamination.
Over the course of three weeks, two uniformly cut film pieces were irradiated for each cone, resulting in a total of six irradiated pieces per cone. After a 24-hour waiting period post-irradiation, the films were scanned using a flatbed Epson Expression 10000 XL scanner (Sieko Epson Corp., Suwa, Japan). All films were scanned in the same orientation to minimise scan-related artefacts and ensure uniformity. Two duplicate scans were obtained for each film piece to ensure reproducibility.
The calibration curve for the films was established over a dose range of 0–10 Gy, with reference doses used for accurate calibration. The scanned images were processed and converted into dose data using RIT software. The cone film doses were calculated as the average dose within a 1 × 1 mm2 region of interest (ROI) positioned at the centre of the beam axis. For each cone, the average dose was obtained from three measurements and the final reported value included the mean and standard deviation, providing a statistical representation of the dose distribution.
Dose profiles
Dose profiles were obtained using an IBA Razor diode and EBT3 film. The diode profiles were measured with the detector parallel to the beam axis, with the respective effective points located at the measurement depth. All measurements were conducted in step of 0.1 mm with an acquisition time of 0.5 s per step. A reference detector was not used. The dose profiles obtained from the EBT3 films were derived from the same films used for SF measurements.
Results
The dose distribution comparison between 6 MV-FFF and 10 MV-FFF treatment plans indicated a slight increase in low-dose volumes for 10 MV-FFF, as illustrated in Figure 2. In the case of the 11 patients, 10 MV-FFF plans demonstrated a slight increase of 3.8% in the dose gradient, along with increments of 17.1 and 17.8% for the volumes receiving 20 Gy and 10 Gy, respectively, in comparison to the 6 MV-FFF plans (Figures 3a, 3b, and 3c). Furthermore, for the same patients, 10 MV-FFF plans exhibited a 6.5% increase in the maximum dose to the brainstem and an 18.1% increase in the dose to 10% of the brainstem volume compared to the 6 MV-FFF plans (Figures 4a and 4b).
TN treatment plan dose comparison between 6 MV-FFF and 10 MV-FFF using a cone diameter of 4 mm.
Comparison of a) dose gradient (mm), b) volume of 20 Gy (cc) and c) volume of 10 Gy (cc) between 6 MV-FFF and 10 MV-FFF for 11 plans.
Comparison of a) brainstem Dmax (Gy) and b) brainstem D10% (Gy) between 6 MV-FFF and 10 MV-FFF for 11 plans.
The average MU per arc for the 6 MV-FFF plans was 1930 ± 38 MU, and the average MU per arc for 10 MV-FFF was 1961 ± 31 MU (Figure 5a). The time for arc 1 was recorded for all plans, showing an average time of 1.39 ± 0.03 minutes for 6 MV-FFF and 0.84 ± 0.02 minutes for 10 MV-FFF, Figure 5b. A reduction of 40% in treatment time per arc was obtained when using 10 MV-FFF in comparison to 6 MV-FFF. Considering couch rotation, the total treatment delivery time was reduced by 28% for the 10 MV-FFF plans.
Comparison of a) average MU per arc and b) time for arc 1 for 6 MV-FFF and 10 MV-FFF plans.
The tissue phantom ratio (TPR) for the 10 MV-FFF beam exhibits an average increase of 10% at a depth of 100 mm compared to the 6 MV-FFF beams (Figure 6). The dose profile indicates an average penumbra (80–20%) width increase of 0.3 mm [0.26–0.35 mm] for the 10 MV-FFF, as seen in Figure 7. According to this figure, the width at the 80% dose level is greater for the 6 MV-FFF than for the 10 MV-FFF, whereas at the 20% dose level, the width for the 10 MV-FFF exceeds that of the 6 MV-FFF. This increase in width is largely attributed to the enhanced width development for the 10 MV-FFF compared to the 6 MV-FFF below the 50% dose level. Likewise, the full width at half maximum (FWHM) measurements for each cone size, as presented in the second and third column of Table 4, exhibited minimal fluctuations across the studied cone sizes for both the 6 MV-FFF and 10 MV-FFF beams.
TPR for cone diameters of 4, 5, 6 and 7.5 mm for 6 MV-FFF and 10 MV-FFF.
Dose profiles (SSD = 900 mm and d = 100 mm) for 6 MV-FFF and 10 MV-FFF for cone diameter of 4, 5, 6 and 7.5 mm.
Dose profile FWHM and scatter factor measured with EBT3 film (SSD = 900 mm and d = 100 mm) for 6 MV-FFF and 10 MV-FFF with cone diameters of 4, 5, 6 and 7.5 mm
The measurement of the scatter factor with EBT3 film is presented in the four and five columns of Table 4, which includes the average and standard deviation of six measurements taken over three weeks. Considering the measurements with the PTW 60012 and Razor detectors, the average scatter factor, along with one standard deviation (SD), is indicated by the black dotted lines in Figure 8. For all cone sizes, the average scatter factor for the 6 MV-FFF beam was greater than that for the 10 MV-FFF beam by 15, 13.5, 12.7 and 10.3% for cones of 4, 5, 6 and 7 mm, respectively. All scatter factor measurements fell within the reference data range provided by Brainlab.
Scatter factor (SSD = 900 mm and d = 100 mm) for 6 MV-FFF and 10 MV-FFF for cone diameter of 4, 5, 6 and 7.5 mm. The bars indicate ±1 SD.
Discussion
As discernible from the results, there was an increase in the lateral dose for the 10 MV-FFF because of the greater width below the 50% dose threshold in a single field, while maintaining nearly the same FWHM. This result is comparable to the values obtained by Wiant et al.,Reference Wiant, Terrell, Maurer, Yount and Sintay23 who found similar FWHM values for different cone sizes in both 6 MV-FFF and 10 MV-FFF beams. However, this increase in width was not observed when 12 arcs were aligned with the isocentre point, as shown in Figure 1. The slight average increase in the dose gradient by 0.1 mm and the volumes of 20 Gy and 10 Gy by 0.04 cc with a 10 MV-FFF beam compared to a 6 MV-FFF beam do not affect the institutional dose constraints. The maximum dose and D10% of the brainstem exhibited average increases of 1.53 Gy and 0.29 Gy, respectively. Nevertheless, a shift of less than 0.3 mm in the isocentre position may be required to achieve the brainstem dose for 6 MV-FFF plans, but this adjustment is not necessary because the doses remain within institutional tolerance for TN treatment. The average increase of 31 UM/arc with the 10 MV-FFF beam over the 6 MV-FFF beam did not affect the arc treatment time, as the dose rate of 2400 UM/min was utilised with 10 MV-FFF. The reduction in treatment delivery time per arc achieved with 10 MV-FFF, by 40% compared to 6 MV-FFF, represents the most significant benefit, enhancing treatment efficiency while maintaining plan quality. This result is comparable to the 30% reduction in delivery time per arc between 10 MV-FFF and 6 MV-FFF reported by Neupane et al.,Reference Neupane, Shang, Kassel, Muhammad, Leventouri and Williams24 using a fixed small multileaf collimator on an Edge (Varian Medical system) linac in 10 TN patients. Moreover, fast treatment can mitigate patient discomfort during treatment and alleviate the pain associated with TN. Given the reduced treatment time, adjustments to the dose prescription should be considered to account for faster delivery.
The differences in TPR, penumbra width (80%–20%) and SF are mainly attributed to energy. The higher TPR value for the 10 MV-FFF beam compared to the 6 MV-FFF beam correspond to higher values of TPR 20/10 for 10 MV-FFF over 6 MV-FFF, as illustrated by Sahani et al.Reference Sahani, Sharma and Sharma25 The increment in penumbra width for 10 MV-FFF compared to 6 MV-FFF was consistent with the findings of the study conducted by Foster et al.,Reference Foster, Speiser and Solberg26 who reported higher penumbra values for 10 MV-FFF than for 6 MV-FFF for fields shaped by both jaws and MLC. Moreover, we found similar FWHM values between 10 MV-FFF and 6 MV-FFF for all cones studied. A similar result was obtained by Wiant et al.Reference Wiant, Terrell, Maurer, Yount and Sintay23 comparing both energies across different cone sizes. The scatter factor is higher for 6 MV-FFF than for 10 MV-FFF for cones of 4, 5, 6 and 7.5 mm, as observed by Cheng et al.Reference Cheng, Ning, Arora, Zhuge and Miller27 in their comparison of output factors using Monte Carlo simulation and measurement for 6 MV-FFF and 10 MV-FFF beams in SRS systems. Additionally, Wiant et al.Reference Wiant, Terrell, Maurer, Yount and Sintay23 obtained similar increases in the scatter factor for 10 MV-FFF over 6 MV-FFF for cone sizes of 4, 6, 7.5, 10, 12.5 and 15 mm.
Conclusion
For TN SRS procedures, treatment plans using 10 MV-FFF beam with stereotactic cone demonstrated reduced delivery times compared to 6 MV-FFF plans. However, this resulted in a slight increase in MUs, low-dose volume and brainstem dose. The faster treatment is mainly attributed to the higher dose rate of the 10 MV-FFF beam (2400 MU/min) versus the 6 MV-FFF beam (1400 MU/min) in the TrueBeam Novalis STx linac. The increased dose is linked to the broader penumbra (80–20%) of the 10 MV-FFF beam, which is consistent with other research findings. While 10 MV-FFF plans meet restriction brainstem dose requirements, to achieve the brainstem dose generated by 6 MV-FFF plans, a minor isocentre shift of less than 0.3 mm is required. Additional research is needed to evaluate the potential dose-related effects of shortening the total treatment time.
Acknowledgements
None.
Financial support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Competing interests
None.
Open access
