Journal of Physics: Conference

Series

PAPER • OPEN ACCESS You may also like

- Glass beads and Ge-doped optical fibres
Monte Carlo based corrections for the dosimetry of as thermoluminescence dosimeters for
small field photon dosimetry
x-ray microbeams with diamond detectors S M Jafari, A I Alalawi, M Hussein et al.

- Technical advances in x-ray microbeam

radiation therapy
To cite this article: Richard P. Hugtenburg and Dimitri D. H. Reynard 2020 J. Phys.: Conf. Ser. 1662 Stefan Bartzsch, Stéphanie Corde, Jeffrey
012013 C Crosbie et al.

- Combinatorial Corrosion Studies of PVD

Deposited, Thermally Interdiffused,
Magnesium Coated Galvanized Steel
Zak Barrett, Hamilton N. McMurray and
View the article online for updates and enhancements. Geraint Williams

This content was downloaded from IP address 35.233.216.72 on 25/10/2024 at 17:53

MMND ITRO 2020 IOP Publishing
Journal of Physics: Conference Series 1662 (2020) 012013 doi:10.1088/1742-6596/1662/1/012013

Monte Carlo based corrections for the dosimetry of x-ray

microbeams with diamond detectors

Richard P. Hugtenburg1,2 and Dimitri D. H. Reynard1,3

1
Swansea University Medical School, Singleton Park, Swansea, SA2 8PP, UK
2
Department of Medical Physics and Clinical Engineering, Singleton Park, Swansea
Bay University Health Board, Swansea, SA2 8QA, UK
3
Université Grenoble Alpes, 15 rue des Universités, St Martin D'Heres, Grenoble,
38400, France

E-mail: r.p.hugtenburg@swansea.ac.uk

Abstract. X-ray microbeams are a potential, novel mode of radiation therapy and dosimetry
methods are under development that require micrometric spatial precision. The microDiamond
detector has the requisite resolution and is composed of diamond which is closely tissue-
equivalent. The high density of diamond however perturbs of secondary electrons and Monte
Carlo methods are needed to determine corrections to accurately measure clinical parameters.
The PENELOPE Monte Carlo code has been used to calculate corrections for the output factor
(OF) and peak-to-valley dose ratio (PVDR). A high-performance computing (HPC) system
was found to be necessary and the calculation took 72 hours when performed on a cluster of
100 CPUs. The correction for the output factor was found to be 1.009±0.016 (2 s.d.). The
correction factor for the peak-to-valley ratio was found to be 1.144±0.013 (2 s.d.) and was
larger due to Compton scattering of the microbeam in the extracameral components of the
detector, in particular the 300 micron bulk diamond crystal. It was found that considerable
improvements in efficiency could be achieved without loss of precision by switching off
electron transport for electrons that are generated far from the sensitive element of the detector.

1. Introduction
Radiation therapy with X-ray microbeams is a promising avenue for the treatment of radioresistant
tumours. X-rays from a 3rd generation synchrotron light-source has been proposed as method of
achieving sufficient dose-rate for the microbeam delivery to be unaffected by patient motion (see
recent reviews by Bräuer-Krisch et al 2015 [1] and Bartzsch et al 2020 [2]). Clinical aspects are
discussed in the review of Grotsky et al 2015 [3]. Microbeam treatment fields are characterised by a
large variation between high dose regions, due to the primary radiation (the peaks), and low dose
regions, due to scattered radiation (the valleys). The difference in X-ray spectral quality and
distribution of these dose components presents a considerable challenge to treatment planning and
verification; in particular requiring that dosimeters have high orders of tissue-equivalence in addition
to achieving the resolution necessary to resolve microbeams.
In this work we consider the use of the microDiamond (PTW, Freiburg, DE) detector, a
dosimeter that has been shown to have good dose-rate linearity [4] as well as being able to accurately
resolving microbeams [5]. The microDiamond detector is expected to have excellent tissue-
equivalence due to its construction from chemical vapour deposition (CVD) diamond (Z = 6), however

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
MMND ITRO 2020 IOP Publishing
Journal of Physics: Conference Series 1662 (2020) 012013 doi:10.1088/1742-6596/1662/1/012013

the presence of the diamond bulk, with a density of 3.51 g cm-3, generates perturbations in the
secondary electron fluence, affecting the determination of shape of the microbeam as well as its
response to photon radiation. Several other detectors have been considered for their use in microbeam
dosimetry, with most achieving sufficiently high precision, but few achieving comparable tissue-
equivalence. Of note is recent success in improving the precision of the readout from radiochromic
film, which is closely tissue-equivalent in terms of effective atomic number, Zeff = 6.73 [6], and
density (ρ ~ 1.4 g/cm3), with a confocal microscopy [7].
The effects of departure from tissue equivalence can be accounted for with Monte Carlo
modelling and a formalism that has been proposed by Bouchard et al 2015 [8], that offers a general
approach to determining detector correction factors that accounts for the contribution from both the
detector sensitive element and the extra-cameral components surrounding the sensitive volume, has
been used here. Monte Carlo can be used in this way to predict the response of the detector due to
variation in the photon spectrum with position and electron transport in the vicinity of the sensitive
volume, however due to the complex structure of the diamond detector and the small sensitive volume,
the approach requires the use of parallel Monte Carlo methods via a high-performance computer
(HPC) system.
Two clinically important parameters are the output factor (OF) and the peak-to-valley dose ratio
(PVDR). The output factor enables the dose to be calculated in the microbeam (the peak dose) from
measurements of a broad-beam of similar quality, where an ionisation chamber measurement can be
used in combination with a traceable dose-standard. The peak-to-valley dose ratio is used to calculate
the dose between microbeams. As the valley-dose is due to scattered radiation it is diffuse and is
expected to have similar clinical effect to a broad-beam. While the precision required for the
microbeam dose is not currently known, it is reasonable to assume that the valley dose would need to
be determined with comparable precision to conventional (broad-beam) radiation therapy, i.e. of order
of 3%. Interestingly the scattering process that generates the valley dose is strongly influenced by the
high degree of polarisation of the X-ray synchrotron beam source, which creates a distinct asymmetry
in the shape of the valley dose distribution [10]. Polarisation has not been included in the modelling as
the contribution from the square fields considered here to a central point of measurement of the OF
and PVDR is independent of the polarisation. Calculations of the correction factors that need to be
employed for measurements of the OF and PVDR with the microDiamond detector are presented and
level of uncertainties in both cases will be discussed.

2. Methods
A model of the microDiamond has been built using the Monte Carlo code PENELOPE, consisting of a
layer of 1 μm thick aluminium that forms the Schottky diode on the surface of the sensitive layer of
diamond of dimension 1.4 μm thick, 2.2 mm diameter positioned on a diamond bulk of thickness 300
μm and encased in epoxy resin and polyether ketone polymer (PEEK) and contained within an
aluminium and RW3 water-equivalent plastic housing (figure 1). Apart from the aluminium layer, no
attempt has been made to model the electrical contacts or stem.
The microbeam field has been modelled with an overall dimension of 2×2 cm2, with 50 μm
microbeams repeated every 400 μm. The 50 microbeams were assumed to be parallel and incident
perpendicular to the water phantom. This approximation is expected to be reasonable given the large
distance from the synchrotron wiggler insertion device to the phantom. A photon spectrum derived for
the ID17 beamline at ESRF in a previous study [11] has been used. The X-ray energy ranges from 50
keV to 300 keV with a peak intensity at 90 keV. The assumption of negligible primary beam between
microbeams may be unreasonable if the collimator system forming the microbeam is not sufficiently
thick and in practice it may be necessary to determine a contribution to the valley dose due to leakage.
Calculations were performed for the detector placed in central peak and valley positions at a
depth of 2 cm in a water phantom. Use was made of splitting in order to lend importance to photon
and electron trajectories traversing the small sensitive volume of the diamond, with the calculation
taking 72 hours of real-time with 200 processors, achieving a statistical precision of 2% (2 s.d). The

2
MMND ITRO 2020 IOP Publishing
Journal of Physics: Conference Series 1662 (2020) 012013 doi:10.1088/1742-6596/1662/1/012013

calculation is repeated with all of the components of the detector replaced with water and one further
calculation of the detector placed in an open 2×2 cm2 field was required to compute the output factor.

Figure 1. Scale model used with the Figure 2. Relative response of the
PENELOPE Monte Carlo code. The diamond microDiamond as a function of distance from
bulk (magenta) is surrounded by epoxy resin the center of a single microbeam. The
(blue) and PEEK plastic (green), contained in response depends on the direction of
an aluminium lined (red) RW3 (black) secondary electrons entering the sensitive
housing of overall diameter 6.9 mm. The volume of the diamond. The inset shows the
detector is surround by water (blue) and arrangement of the beam and the orientation
positioned at a depth of 2 cm. The sensitive of the detector realtively to the x-axis, where
diamond layer above the diamond bulk and the beam in the example given is at a position
the aluminium surface are not visible in the -0.02 cm from the sensitive volume.
diagram.

3. Results
The relative response of the microDiamond to a water-equivalent medium was calculated in 200 μm
steps for a range of distances from -1 to 1 cm from a single 50 μm wide microbeam (figure 2). The
contribution for each microbeam to the peak and to the valley is determined by summing the energy
deposition from all 50 microbeams. The correction factor determined for the measurement of the
central peak was found to be 1.103±0.007, due predominantly to the microbeam that straddles the
sensitive volume. The lower cross-section of carbon versus water at these energies reduces the
response of the diamond detector while the higher density of the diamond bulk on the range of
Compton electrons increases it slightly. The correction factor for the open 2×2 cm2 field was found to
be 1.089±0.010, therefore the overall correction to the output factor is 1.009±0.016. The correction
factor for the center-most valley is 0.964±0.007 and the overall correction factor for the peak-to-valley
ratio was 1.144±0.013.
Although the overall correction factors for the valley dose measurements is small, large
variations in the response to microbeams in the vicinity of the sensitive volume are evident, and
balance to an extent. These variations are shown to be due to extracameral components of the detector
which have a particularly strong influence on the response of the detector when measuring
immediately adjacent valleys. It is found that Compton scattering of a microbeam traversing the
diamond bulk leads to enhancement factor of 1.9, while Compton scattering in the PEEK plastic
(density = 1.32 g cm-3) and RW3 (density = 1.045 g cm-3) components leads to a smaller, but non-
negligible enhancement of the response of approximately 1.1, where it is noted that RW3 has a small
(0.2%) but non-negligible titanium (Z=22) content. Asymmetry in the response at distances further
from the sensitive volume is due to two effects, the difference in scattering conditions for the diamond
detector which increases the response and the reduced response to those photons when they interact in
the vicinity the diamond sensitive layer, which decreases the response.

3
MMND ITRO 2020 IOP Publishing
Journal of Physics: Conference Series 1662 (2020) 012013 doi:10.1088/1742-6596/1662/1/012013

Although the calculation time was expected to large, as the sensitive volume presents a cross-
section that is 3.5 million times smaller than the open field, it has been shown that control of the range
of secondary electrons can additionally boost the speed of the calculation substantially. For example it
is reasonable to assume that electron transport is only important in the vicinity of the microbeam and
that electrons generated by the Compton scatter of the primary microbeams will not reach the sensitive
volume for all microbeams, apart from the central microbeam. The speed-up associated with these
additional approximations is a factor of 20, which would reduce the overall calculation time to 720
hours, easily accessible on modern HPC systems.

4. Conclusions
Despite the excellent tissue-equivalence of the diamond detector a significant correction is shown to
be required due to the perturbation of electron fluence in the vicinity of the diamond crystal. Monte
Carlo calculations on a HPC system have shown that the corrections are, in the case of the clinically
important peak-to-valley ratio, significant and as expected time-consuming. The analysis points to
potential improvements in the efficiency of the calculation and confidence in the use of diamond
detectors as a convenient on-line dosimetry method.

5. References
[1] Bartzsch S, Corde S, Crosbie JC, Day L, Donzelli M, Krisch M, Lerch M, Pellicioli P, Smyth
LM, Tehei M 2020 Technical advances in x-ray microbeam radiation therapy. Physics in
Medicine & Biology 65 02TR01
[2] Bräuer-Krisch E, Adam JF, Alagoz E, Bartzsch S, Crosbie J, DeWagter C, Dipuglia A, Donzelli
M, Doran S, Fournier P, Kalef-Ezra J 2015 Medical physics aspects of the synchrotron radiation
therapies: Microbeam radiation therapy (MRT) and synchrotron stereotactic radiotherapy (SSRT)
Physica Medica 31 68-83
[3] Grotzer MA, Schültke E, Bräuer-Krisch E, Laissue JA 2015 Microbeam radiation therapy:
clinical perspectives Physica Medica 31 564-7
[4] Brualla-González L, Gómez F, Pombar M, Pardo-Montero J 2015 Dose rate dependence of the
PTW 60019 microDiamond detector in high dose-per-pulse pulsed beams Phys. Med. Biol. 2015
61 N11
[5] Livingstone, J, Stevenson AW, Butler DJ, Häusermann D, Adam, JF 2016 Characterization of a
synthetic single crystal diamond detector for dosimetry in spatially fractionated synchrotron x‐ray
fields. Medical physics, 43(7), 4283-4293. 2016 Medical Physics 43 4283–4293
[6] Butson M, Niroomand-Rad A 2017 Historical background, development, and construction of
radiochromic films. In Radiochromic Film CRC Press
[7] Pellicioli P, Bartzsch S, Donzelli M, Krisch M, Bräuer-Krisch E 2019 High resolution
radiochromic film dosimetry: Comparison of a microdensitometer and an optical microscope.
Physica Medica 65 106-13
[8] Bouchard H, Kamio Y, Palmans H, Seuntjens J, Duane S 2015 Detector dose response in
megavoltage small photon beams. ii. pencil beam perturbation effects. Medical Physics, 42 6048–
6061
[10] Hugtenburg RP, Adegunloye AS, Bradley DA 2010 X-ray microbeam radiation therapy
calculations, including polarisation effects, with the Monte Carlo code EGS5. Nucl. Instrum.
Meth. A 619 221-4
[11] Martıń ez-Rovira I, Sempau J, Prezado Y 2012 Development and commissioning of a monte carlo
photon beam model for the forthcoming clinical trials in microbeam radiation therapy. Medical
Physics 39 119–131