Technical note: Vendor‐agnostic water phantom for 3D dosimetry of complex fields in particle therapy

Abstract Purpose Three‐dimensional (3D) dosimetry is a necessity to validate patient‐specific treatment plans in particle therapy as well as to facilitate the development of novel treatment modalities. Therefore, a vendor‐agnostic water phantom was developed and verified to measure high resolution 3D dose distributions. Methods The system was experimentally validated at the Marburger Ionenstrahl‐Therapiezentrum using two ionization chamber array detectors (PTW Octavius 1500XDR and 1000P) with 150.68 MeV proton and 285.35 MeV/u 12C beams. The dose distribution of several monoenergetic and complex scanned fields were measured with different step sizes to assess the reproducibility, absolute positioning accuracy, and general performance of the system. Results The developed system was successfully validated and used to automatically measure high resolution 3D dose distributions. The reproducibility in depth was better than ±25 micron. The roll and tilt uncertainty of the detector was estimated to be smaller than ±3 mrad. Conclusions The presented system performed fully automated, high resolution 3D dosimetry, suitable for the validation of complex radiation fields in particle therapy. The measurement quality is comparable to commercially available systems.


| INTRODUCTION
State-of-the-art particle therapy can deliver highly conformal dose distributions to target volumes while sparing healthy tissues due to their advantageous inverse depth dose distributions and increased biological effectiveness compared to other forms of external radiation therapy. 1 To leverage the physical and biological advantages of particles for better health outcomes for patients, strict patient-specific quality assurance (PSQA) is a necessity. 2 Classic three-dimensional (3D) dosimetry concepts for PSQA, employing multiple pin-point ionization chambers, 3 radiographic films or gels 4 have either limited spatial resolution (pin-point ionization chambers) or a complex dependence on linear energy transfer (LET) and particle energy (radiographic films or gels). Medical two-dimensional ionization chamber arrays 5 or novel approaches using GEM-based detectors 6 offer a comparatively high spatial resolution and immediate data readout, but are often not directly usable within a water phantom and/or are comprised of non-water-equivalent materials. Even though the stopping power of materials such as water-equivalent plastics are comparable to water, nuclear interactions of heavier ions, such as carbon, and scattering of light ions, such as protons, will be different in these materials and will have detrimental effects on the comparability of the measurement outcome. 7 Additionally, high-resolution, 3D dosimetry is an absolute necessity for benchmarking and for further optimization of different beam application modalities. Two-dimensional passive range modulators (2DRM), like 2D SOBP modulators 8 or ripple filters, 9 are still under active development and three dimensional passive range modulators (3DRM) are showing great potential for the treatment of moving tumors 10 and may be crucial for new treatment modalities, such as FLASH irradiations. 11,12 To the best of our knowledge, only the company IBA (Louvainla-Neuve, Belgium) offers a commercial medical product 13 suitable for 3D dosimetry in water for particle therapy, based on a two-dimensional ionization chamber array. 14 However, this dosimetry system is highly integrated into the IBA ecosystem of accelerators and quality assurance solutions making it difficult to use at non-IBA medical accelerators, such as the SIEMENS-built medical ion beam facilities in Marburg (MIT), Shanghai (SPHIC) or Heidelberg (HIT). The applicability of such a highly optimized medical dosimetry system to the development of different beam application modalities, such as 2DRMs and 3DRMs, is hindered further by the inability to extend its capabilities or to optimize the system for a specific experimental measurement task. Therefore, a universal and vendor-agnostic water column for 3D dosimetry in particle therapy was developed [WatER column for 2D ioNization chambEr aRray detectors (WERNER)]. The system is comprised of a PMMA water tank with a stepper-driven watertight detector holder, a standard medical two-dimensional ionization chamber array as well as analog and digital input and outputs (AD I/O). The full system is controlled via a LabVIEW application and synchronized to the dose delivery system of the medical accelerator using commonly available digital signals. In this work, we describe the current iteration of WERNER, optimized for the use at MIT and for the PTW 2D ionization chamber arrays OCTAVIUS 1500XDR and 1000P. The system recorded more than 50 000 individual dose points in <5 min per measurement to validate and benchmark complex fields generated by different 2DRMs and 3DRMs.

| MATERIALS AND METHODS
The development of the universal water column for 2D ionization chamber array detectors is described in detail in the following paragraphs. A schematic drawing of the full system is shown in Fig. 1.

2.A | Mechanical setup
The water column consists of a cuboid-shaped PMMA tank (inner dimensions: 400 × 335 × 350 mm 3 ) and 20 mm thick outer walls for mechanical stability. To minimize the nonwater material thickness, a 5 mm thick PMMA beam entrance window (dimensions: 220 × 220 mm 2 ) is embedded into the front-facing wall. Two ball screw actuators (GUNDA ELS-R25), connected to a single stepper motor (GUNDA SM23Hz) via a drive belt, are mounted to the side of the water column to ensure highly precise and synchronous movement of the detector holder over the full length of the water column. Motor movement is governed by an external stepper motor driver (GUNDA BAC100), which is controlled by a LabVIEW application via RS485. The watertight PMMA detector holder is connected to both actuators via a spring-loaded connection to prevent excessive forces during fast acceleration and, additionally, is stabilized by a top-mounted aluminum bracket to prevent oscillation. The detector stands on a raised platform within its holder to offer protection in case of water leakage and to guarantee a reproducible and tight fit to the 5 mm front plate of the holder. To minimize the production of waves in the water tank, the nominal movement speed of the detector was set to 10 mm/s.

2.B | Array detectors
All tests were conducted with two different PTW ionization chamber array detectors: the OCTAVIUS 1500XDR and the OCTAVIUS 1000P. Both detectors are specified for use in particle therapy and share a common detector housing geometry, readout characteristics and readout software and are, therefore, easily interchangeable during an experiment. over an area of 11 × 11 cm 2 . In the inner 5.5 × 5.5 cm 2 area ionization chamber spacing is 2.5 mm, whereas the spacing outside is F I G . 1. Schematic view of the developed water column. A full three-dimensional representation can be found in the supplements vided by the manufacturer, and are, therefore, only suitable for relative dosimetry of 12 C beams. Absolute dose calibration of the detectors is possible using external reference fields, but is not within the scope of this work.

2.C | Measurement control
The PTW two-dimensional ionization chamber array is operated with a beta version of its standard software -BeamAdjust (V2.

| RESULTS
The described water column was validated both mechanically, using high precision dial gauges (Garant 43 2210_10/58), and a two-dimensional inclinometer (DigiPas DWL-1500XY) as well as experimentally, exploiting the sharp dose gradients used in carbon therapy. The mechanical validation could only be performed relative to the detector holder and is therefore not able to verify the absolute position of the detector within its holder. In general, the mechanical validation was in good agreement with the validation using particle beams and, therefore, is not discussed in detail. However, it is important to note that the same tools and techniques are used for the mechanical calibration after each assembly. The impact of improper mechanical calibration is highlighted in Fig. 2.    Fig. 4. The measurement consists of around 60,000 individual dose points and was measured in <8 min.
In order to compare the performance of WERNER to a commercially available medical system, we performed additional measurements with 150.68 MeV protons and the PTW PeakFinder. The PeakFinder is a water column system, designed for high precision Bragg curve measurements and is routinely used for quality assur- are the gold-standard in absolute dosimetry in water, a full characterization of a complex dose distribution, as needed for the development of for example, 3DRMs, is not feasible due to a significantly lower amount of measurement points per unit time compared to a system such as WERNER. Even though dosimetry approaches using radiographic films or gels offer a high spatial resolution and fast measurement time, the measurement quality and linearity, especially for carbon ions, is highly limited 16 and not suitable for the intended purpose of the presented system. However, it is important to note that, compared to those systems, a substantial increase in the number of measurement steps to attain a higher measurement resolution will increase the measurement time for a system like WERNER. Therefore, we recommend that WERNER be used with a variable step size, employing larger steps for low dose gradient regions and smaller steps for high dose gradient regions.
Promising novel approaches, like GEM-based systems, 17   around 30 to 60 min, including the full assembly of the system, the filling with demineralized (VE) water and its calibration. It is worth noting that error values obtained during mechanical calibration cannot be directly translated to an absolute measurement error: they are only used as an indicator for the goodness of the mechanical alignment and the irradiation of a monoenergetic scanned field of an appropriate size is strongly advisable as an in-beam calibration after assembly.
Due to the versatile design of the system and its applicability at a variety of particle accelerators, specific parameters of WERNER, such as detector resolution and measurement speed, are highly dependent on the specific use case and measurement setup. The presented system can potentially work with a multitude of different medical array detectors and is limited mainly by their size, weight, and readout characteristics (time resolved readout mode or external trigger signal are required). By using external signals to control the measurement logic, WERNER can be synchronized to additional measurement equipment, if needed.

| CONCLUSION
A vendor-agnostic water phantom for 2D ionization chamber array detectors was developed and verified with beam using 150.68 MeV protons and 285.35 MeV/u 12 C, provided by the medical synchrotron at MIT. The presented system is suitable to be used for the fast characterization of complex dose distributions, as found in many particle therapy related contexts.