Download as pdf or txt
Download as pdf or txt
You are on page 1of 38

This article was downloaded by: 10.3.97.

143
On: 15 Nov 2023
Access details: subscription number
Publisher: CRC Press
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: 5 Howick Place, London SW1P 1WG, UK

Handbook of Radiotherapy Physics


Theory and Practice, Second Edition
W. P. M. Mayles, A. E. Nahum, J.-C. Rosenwald

Principles and Basic Concepts in Radiation Dosimetry

Publication details
https://www.routledgehandbooks.com/doi/10.1201/9780429201493-7
Alan Nahum
Published online on: 31 Dec 2021

How to cite :- Alan Nahum. 31 Dec 2021, Principles and Basic Concepts in Radiation Dosimetry from:
Handbook of Radiotherapy Physics, Theory and Practice, Second Edition CRC Press
Accessed on: 15 Nov 2023
https://www.routledgehandbooks.com/doi/10.1201/9780429201493-7

PLEASE SCROLL DOWN FOR DOCUMENT

Full terms and conditions of use: https://www.routledgehandbooks.com/legal-notices/terms

This Document PDF may be used for research, teaching and private study purposes. Any substantial or systematic reproductions,
re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contents will be complete or
accurate or up to date. The publisher shall not be liable for an loss, actions, claims, proceedings, demand or costs or damages
whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.
Chapter 5
Principles and Basic Concepts in Radiation Dosimetry
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

Alan Nahum*

CONTENTS
5.1 Introduction............................................................................................................................................................. 56
5.2 The Stochastic Nature of Energy Deposition............................................................................................................ 56
5.3 Definitions of Dosimetric Quantities........................................................................................................................ 57
5.3.1 Absorbed Dose.............................................................................................................................................. 57
5.3.2 Kerma (and Exposure)................................................................................................................................... 58
5.3.3 Particle Fluence............................................................................................................................................. 59
5.3.4 Energy Fluence.............................................................................................................................................. 59
5.3.5 Planar Fluence............................................................................................................................................... 60
5.3.6 Fluence Rate.................................................................................................................................................. 60
5.4 Relations between Fluence and Dosimetric Quantities for Photons........................................................................... 60
5.4.1 Relation between Photon Fluence and Kerma................................................................................................ 60
5.4.2 Relation between Kerma and Absorbed Dose................................................................................................ 61
5.5 Charged-Particle Equilibrium (CPE); Partial CPE.................................................................................................... 62
5.6 Relation between Fluence and Dose for Charged Particles........................................................................................ 65
5.6.1 Stopping Power and CEMA.......................................................................................................................... 65
5.6.2 Delta-Ray Equilibrium.................................................................................................................................. 66
5.7 Cavity Theory........................................................................................................................................................... 66
5.7.1 Introduction.................................................................................................................................................. 66
5.7.2 The Fano Theorem........................................................................................................................................ 66
5.7.3 Cavity Theory for ‘Large’ Photon Detectors.................................................................................................. 67
5.7.4 Bragg–Gray Cavity Theory............................................................................................................................ 69
5.7.4.1 Introduction.................................................................................................................................... 69
5.7.4.2 Theory............................................................................................................................................ 70
5.7.4.3 When Does a Detector Behave in a Bragg–Gray Manner?................................................................ 70
5.7.4.4 Bragg-Gray (Unrestricted) Stopping-Power Ratios, Water-to-Detector............................................ 72
5.7.4.5 Is a Low-density Cavity in a Medium Irradiated by Kilovoltage X-Rays a Bragg–Gray Cavity?......... 72
5.7.5 The Spencer–Attix Modification of Bragg–Gray Theory................................................................................ 73
5.7.5.1 Introduction.................................................................................................................................... 73
5.7.5.2 Theory............................................................................................................................................ 74
5.7.5.3 Practical Implications of Spencer–Attix Theory............................................................................... 75
5.7.6 Departures from ‘Perfect’ Bragg–Gray Behaviour For Ionisation Chambers.................................................. 76
5.7.6.1 Introduction.................................................................................................................................... 76
5.7.6.2 The Various Sources of Ion-Chamber Perturbation.......................................................................... 77
5.7.7 General or ‘Burlin’ Cavity Theory................................................................................................................. 77
5.7.8 The Response of a Detector as it Varies from ‘Bragg–Gray’ to ‘Large Photon’............................................... 78
5.8 The Special Case of Small, Non-equilibrium, Megavoltage Photon Fields................................................................. 79
5.8.1 Introduction.................................................................................................................................................. 79
5.8.2 Absorbed Dose in a Uniform Medium Irradiated by a Small Photon Field..................................................... 79
5.8.3 Detector Response in Small Photon Fields..................................................................................................... 80

* Contributions by John Fenwick and Sudhir Kumar are gratefully


acknowledged.

55
56 Part A: Fundamentals

5.8.3.1 General............................................................................................................................................ 80
5.8.3.2 Response as a Function of Detector Type......................................................................................... 81
5.8.3.3 Small-Field Perturbation Factor for Air-filled Ionisation Chambers.................................................. 83
5.8.3.4 The Physics of ‘Bragg–Gray Breakdown’ in Small Fields.................................................................. 84
5.8.4 Mass-density Compensation to Improve Small-field Detector Response......................................................... 86

experimental and theoretical aspects of radiation dosimetry


5.1 Introduction can be found in Greening (1981) and especially Andreo
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

et al. (2017). Whyte (1959), Kase and Nelson (1978), Johns


As described in Chapter 3, charged particles (including and Cunningham (1983), Rajan (1992) and Metcalfe et al.
those generated from uncharged-particle interactions – (1997) are also useful references.
see Chapter 4) lose energy when they interact with matter. From the formal definitions of the quantities such as
This energy is transferred to and absorbed by the irradi- absorbed dose and kerma given later in this chapter, one might
ated medium. This energy absorption is the basis for the imagine that radiation dosimetry is an old, well-established
biological effects of radiotherapy that may ultimately result discipline. In reality, its history is rather tortuous. Following
in tumour eradication (see Chapter 6). The absorption of Röntgen’s discovery of x-rays in 1895, the first radiation unit
(charged-particle kinetic) energy is expressed in terms of the proposed was based on its ionising power. Christen (1914)
fundamental quantity absorbed dose; this chapter is dedi- was ahead of his time in advocating the quantity dose, defined
cated to radiation dosimetry, i.e. the determination (either as ‘radiant energy per unit volume’. In 1928 the International
experimentally or theoretically) of the absorbed dose in mat- X-ray Unit Committee defined the Röntgen as:
ter irradiated by ionising radiation.
The success of radiotherapy depends on accurate knowl- ‘the quantity of x-radiation which, when second-
edge of the absorbed dose delivered to the patient under- ary electrons are fully utilised and the wall effect
going treatment. So-called ‘dose–response curves’ for of the chamber is avoided, produces in one cubic
tumour control and for normal-tissue damage (see Part B centimetre of atmospheric air at 0 °C and 76 cm
and Chapter 44) are relatively steep, especially for certain mercury pressure, such a degree of conductivity
tumours and normal tissues. Consequently, a difference that one electrostatic unit of charge is measured at
of only a few per cent in the ‘target’ dose may separate saturation current.’
uncomplicated tumour eradication (or ‘control’) from
either failure to control the tumour through underdosage This quantity later became known as exposure (see
or, assuming the dose is increased everywhere in the treat- Section 5.3.2).
ment plan by the same factor, serious normal-tissue dam- It was not until 1950 that the International Commission
age through overdosage. on Radiation Units and Measurements (ICRU) formalised
There are several different steps involved in the determina- the definition of dose in terms of ‘the quantity of energy
tion of the absorbed dose distribution in the patient. The first absorbed per unit mass (ergs per gram) of irradiated material
step in the radiotherapy clinic involves measurements with a at the point of interest’ (ICRU 1951). In 1954, the ICRU
detector (also known as a dosimeter) in a phantom (often water, finally approved the term absorbed dose, with its unit the rad
sometimes water-like plastic) placed in the radiation beam (or defined as ‘100 ergs per gram’. The modern unit, the gray
more generally, radiation field). Such measurements include (Gy), is exactly 100 rad (1 cGy = 1 rad). The meaning and for-
determining the dose at a reference depth in a reference-size mal definition of absorbed dose were further refined, and its
field (see Chapter 19), relative doses at many positions in the modern form (see Section 5.3.1) given in ICRU report 33
phantom to map out a complete dose distribution (see Chapter (ICRU 1980) and carried through to report 60 (ICRU 1998)
20), and so-called in vivo doses on the patient’s skin during and report 85, revised as report 85a (ICRU 2011).
treatment (see Section 48.3).
The detector yields the dose Ddet to its own sensitive mate-
rial via a calibration coefficient derived from a quantity of
charge, light, film blackening etc. The dose required is that at 5.2 The Stochastic Nature of
the position r(x, y, z) in the medium, D med (r), in the absence of Energy Deposition
the detector. The conversion of Ddet to Dmed is a fundamental
step requiring knowledge of the theoretical aspects of radia- Absorbed dose is the result of the deposition of energy in
tion dosimetry. The same is true of the calculation of the dose matter by radiation tracks. This is an inherently random pro-
distribution inside the (heterogeneous) patient (see Part F). cess, as Chapters 3 and 4 make clear. Consider a situation
This chapter covers the fundamental ideas and principles where the energy, E, deposited in an ever-decreasing volume
involved in radiation dosimetry, independently of the par- (and therefore mass) centred on the same point in a uniform
ticular detector being used. The characteristics of the vari- medium, is determined for a series of irradiations of identical
ous detectors or dosimeters used in radiotherapy, as well as duration. A plot is made of E divided by the mass m of the
the numerical values of the relevant quantities, are covered volume for each separate irradiation. Figure 5.1 shows the
principally in Part D. Comprehensive treatments of the expected result.
Chapter 5: Principles and Basic Concepts in Radiation Dosimetry 57

The unit of absorbed dose is the gray (Gy), which is equal


to 1 Joule per kilogram (J kg−1). ICRU (2011) defines the
energy imparted, ε, as the sum of all energy deposits (see later)
in the volume.
The mean energy imparted, e , by ionising radiation to the
matter in a volume is defined by the ICRU (2011) as:

E/m e = Rin - Rout + SQ (5.2)


Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

where
R in is the mean energy (excluding rest mass energies) of
D all those charged and uncharged ionising particles
that enter the volume
Rout is the mean energy (excluding rest mass energies) of
all those charged and uncharged ionising particles
Log m
that leave the volume
FIGURE 5.1 Energy deposited per unit mass, plotted against the ∑Q is the sum of all changes (decreases: positive sign,
mass, m, of the scoring volume, as this volume is gradually reduced in increases: negative sign) of the rest mass energy of
size for a given incident radiation fluence (as defined in Section 5.3.3). nuclei and elementary particles in any nuclear trans-
The shaded portion (cloud) represents the range where statistical fluctu- formations that occur in the volume
ations become important as the volume (i.e. m) becomes smaller. (From R in and Rout are known as radiant energies.
Rossi, H. H., Radiation Dosimetry, Vol. 1, Academic Press, New York,
1968.)
Figure 5.2 illustrates the concept of the stochastic quantity
energy deposit. In Figure 5.2a, illustrating a Compton inter-
For large volumes (i.e. large Log m), E/m has the same action (see Section 4.3.2.1) within the volume V, the energy
value, D, for repeated measurements, almost independent of deposit ε is given by:
the volume.* Below a certain value of m, fluctuations start
to appear, which increase as m decreases. These are due to e = hn 1 - ( hn 2 + hn 3 + E ¢ ) (5.3)
the fact that energy is deposited through a finite number
of interactions along particle tracks; some volumes, if small where E' is the kinetic energy of the charged particle – of ini-
enough, will contain few or no tracks, while others may con- tial kinetic energy E – on leaving the volume V. Note that the
tain many. Therefore, for a given particle fluence (see Section photon hn4 does not appear in Equation 5.3, as this photon
5.3.3), below some magnitude of the elementary volume, the is not emitted within the volume V. The ΣQ term is also not
randomness of energy deposition will become apparent. involved here.
The absorbed dose discussed in what follows is non-stochas- The volume in Figure 5.2b involves γ-ray emission (hn1)
tic; i.e. it is assumed to be determined in a mass element suffi- from a radioactive atom, pair production (kinetic energies
ciently large for the fluctuations to be negligible. The subject E1 and E2), and annihilation radiation as the positron comes
of fluctuations in the dose is generally dealt with under the to rest (see Section 3.8). The energy deposit in this case is
heading of microdosimetry (ICRU 1983; Goodhead 1987) given by:
and falls outside the scope of this chapter (see Section 6.11.5).
e = 0 - 1.022 MeV + SQ (5.4)

5.3 Definitions of Dosimetric Quantities where

5.3.1 Absorbed Dose SQ = hn 1 - 2mec 2 + 2mec 2 = hn 1 (5.5)


The absorbed dose is defined by ICRU (2011) as follows:
In Equation 5.4, the zero corresponds to the R in term;
The absorbed dose, D, is ‘the quotient of de by dm,
where de is the mean energy imparted by ionising
‘1.022 MeV’ corresponds to the Rout term, comprising the
radiation to matter of mass dm’: two annihilation γ-rays. The γ-ray energy hn1 arises from
a decrease in the rest mass of the nucleus; the −2mec2 term
de is due to the creation of an electron–positron pair, i.e. an
D= (5.1) increase in rest mass energy. Finally, the +2mec2 term is due
dm
to the annihilation of an electron and a positron. Careful
thought is required in such examples.
Note that the quantity ε, defined as the sum of (single-
* For photons, the fall-off at large values of m or Log m corresponds to
attenuation of the radiation when the path length across the volume
interaction) energy deposits in a volume, is stochastic; in
becomes comparable with the mean free path for the particle interac- Equation 5.2, consistent with the definition of absorbed
tions; this consideration is not relevant here. dose, its mean value is used, which is non-stochastic.
58 Part A: Fundamentals

hn = 0.5II MeV
V
e
E'
hn4 V
hn1 E
hn3
hn2 q+
E2
hn1
E1
q-
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

hn = 0.5II MeV

(a) (b)

FIGURE 5.2 An illustration of the concept of the energy imparted to an elementary volume by radiation. (a) Compton interaction from an incom-
ing photon impinging on the volume V. (b) γ-ray emission from a radioactive atom resulting in pair production. (From Attix, F. H., Introduction
to Radiological Physics and Radiation Dosimetry, Wiley, New York, 1986. With permission.)

5.3.2 Kerma (and Exposure) both. Like kerma, it describes energy transferred to


electrons in a small volume at the point of interest,
The quantity kerma* can be thought of as a step towards but for exposure air is always taken as the reference
absorbed dose. Conceptually, it is very close to exposure, the material, whether or not it is actually present. Like
first radiation quantity to be formally defined (see Section kerma rate, the exposure rate is used to characterize
5.1 and e.g. Greening 1981). Medical physicists working x- and γ-ray fields, and in fact has continued to be
with external-beam radiation therapy were, until around used for this purpose in preference to kerma rate.
2000, familiar with kerma (in air in a 60Co γ-ray beam) in Exposure is similar to absorbed dose in one respect:
the context of the calibration of an ionisation chamber at a energy spent by the secondary electrons in generat-
National Standards Laboratory (see Section 19.3).† Kerma is ing bremsstrahlung contributes neither to the local
defined by the ICRU (2011) as follows: absorbed dose nor to the exposure.

The kerma, K, is ‘the quotient of dEtr by dm, where Figure 5.3 illustrates the concept of kerma (and exposure).
dEtr is the mean sum of the initial kinetic ener- It is important to realise that it is the initial kinetic energies
gies of all the charged particles liberated in a mass that are involved; the eventual fate of the charged particles
dm of a material by the uncharged particles incident (e.g. whether they leave or do not leave the elementary vol-
on dm’: ume) has no influence on kerma. In the volume in the figure,
dE tr the initial kinetic energies of the two electrons labelled e1
K = (5.6) contribute to the kerma, as both were generated in the vol-
dm
ume. The fact that one of these electrons leaves the volume
The unit of kerma is the same as that of absorbed dose, with a residual kinetic energy E1 is irrelevant. The kinetic
J kg–1 or gray (Gy). Kerma applies only to indirectly ionis- energy E2 of the electron labelled e2 entering the volume
ing (or uncharged) particles, which for our purposes, almost makes no contribution to kerma, as this electron was gener-
always means photons, although neutrons also fall into this ated outside the volume.
category.
The quantity exposure is closely related to air kerma.
Exposure, usually denoted by X, is the quotient of dQ by dm, ,m
where dQ is the absolute value of the total charge of the ions Air
of one sign produced in air when all the electrons (negatrons e1
and positrons) liberated by photons in air of mass dm are
X or
completely stopped in air. Until the late 1970s, all ionisation
chambers were calibrated in terms of exposure; subsequently, rays
this was replaced by air kerma.
Attix (1968) described exposure thus:
e2
E2 E1
Exposure is conceptually identical neither to kerma
nor absorbed dose, but it has characteristics of e1

FIGURE 5.3 Illustration of the concepts of kerma (and exposure).


The two electrons labelled e1 contribute to the kerma; but the electron
* ‘kerma’ is an acronym for ‘kinetic energy released per unit mass’. labelled e2 does not. (Reproduced from Attix, F. H., Introduction to
† The quantity kerma in air is still in official use to specify the ‘strength’ Radiological Physics and Radiation Dosimetry, Wiley, New York, 1986.
of the radioactive sources used in brachytherapy (see Section 51.3). With permission.)
Chapter 5: Principles and Basic Concepts in Radiation Dosimetry 59

It is emphasised that kerma includes the energy that the Let N be the number of particles striking a finite sphere
charged particles may re-radiate in the form of bremsstrah- surrounding point P (during a finite time interval). If the
lung photons (i.e. radiation losses). Kerma can be therefore sphere is reduced to an infinitesimal one at P with a cross
partitioned as follows (Attix 1979; Andreo et al. 2017): sectional area of da, then the fluence Φ is given by:

K = K col + K rad (5.7) dN


F = (5.11)
da
where This is usually expressed in units of m–2 or cm–2 (ICRU
‘col’ refers to ‘collisions’ with the atomic orbital electrons 2011). Note that fluence is a scalar quantity – the direction
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

(i.e. electronic losses -see Section 3.2) of the radiation is not taken into account.
‘rad’ refers to radiation losses from interactions with the
atomic nuclei (see Section 3.4). We will be making significant use of fluence, differential in
energy E, often written as ΦE:
The collision kerma,* Kcol, is related to the (total) kerma by
dF
K col = K (1 - g ) (5.8) FE = (5.12)
dE
where g is the fraction of the initial kinetic energy of the elec- where dF is the fluence of particles of energy between E and
trons that is re-radiated as bremsstrahlung (in the particular E + dE.
medium of interest).† E max

Exposure and air kerma can be related to each other as The (total) fluence can be obtained from F =
ò0
F E dE .
follows. The product of the charge dQ (refer to the defini-
tion of exposure given earlier) and the mean energy required Similarly, the fluence, differential in solid angle Ω, can be
to produce one ion pair in air, often written as Wair/e (see defined as:
ICRU 2011 and Section 16.1), yields the electronic part of
dF
the energy transferred, i.e. dEtr (1 − g), and therefore, FW = (5.13)
dW
X (Wair / e) = K air (1 − g ) (5.9) where dF is the fluence of particles propagating within a
or solid angle dW around a specified direction.‡
The two differential quantities can be combined, yielding
X (Wair / e) = (K col )air (5.10) F W ,E , the fluence differential in solid angle and energy.

5.3.3 Particle Fluence Fluence can also be expressed as the quotient of the sum
of the track lengths of the particles crossing the elementary
To calculate absorbed dose, we require quantities that describe sphere and the volume of the sphere§ (Chilton 1978; Papiez
the radiation field, known as field quantities (Greening 1981) and Battista 1994):
or radiometric quantities (ICRU 2011). Particle fluence
(ICRU 2011), the number of particles per unit area, is a key SDs
F = (5.14)
field quantity. The concept is illustrated in Figure 5.4. dV
This form is extremely useful when considering so-called cav-
Sphere Point P ity integrals (see Section 5.7) which yield kerma or absorbed
dose by integrating (over energy) the fluence, differential in
energy (over the volume of interest), multiplied by the rel-
Great circle, area da evant interaction coefficient.

5.3.4 Energy Fluence
Energy fluence is the product of fluence and particle energy.
Particle traversing Volume dV Let R be the value of the total (radiant) energy (excluding
sphere
rest mass energy) carried by all the N particles in Figure 5.4
(i.e. the radiant energy). Then, the energy fluence Ψ is given
FIGURE 5.4 Characterisation of the radiation field at a point P in terms
of the radiation traversing a sphere centred at P. (Adapted from Attix,
by (ICRU 2011):
F. H., Introduction to Radiological Physics and Radiation Dosimetry, dR
Wiley, New York, 1986.) Y = (5.15)
da
* This quantity is sometimes called electronic (or ionisation) kerma ‡ The specification of a direction requires two variables; in spheri-
which would be consistent with the change from collision to elec- cal coordinate systems with polar angle θ and azimuthal angle φ,
tronic stopping power – see Section 3.2.2. However, ICRU Report dΩ = sinθ dθ dφ.
90 (2014) has recommended the continued use of collision kerma - see § This can be proved by noting that for a volume V and external sur-
Rogers and Townson (2019). face area S, the mean chord length is 4V/S (i.e. 4r/3 for a sphere of
† g ≈ 0.003 in the case of cobalt-60 γ-rays in air (IAEA 1997). radius r).
60 Part A: Fundamentals

If particles of only a single energy, E, are present, then R = E N


and Ψ = E Φ. 5.4 Relations between Fluence and
Exactly as for fluence, the energy fluence can be expressed Dosimetric Quantities for Photons
as differential in energy, solid angle or both.
This section relates quantities concerning energy trans-
ferred to (i.e. kerma) or deposited in (i.e. absorbed dose)
5.3.5 Planar Fluence a medium to the fluence that characterises the radiation
Planar fluence is the number of particles crossing a fixed field. This will first be done for uncharged particles (pho-
plane in either direction (i.e. summed by scalar addition) per tons), starting with the relation between fluence and kerma
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

unit area of the plane (see Figure 5.5). One can also define a (Section 5.4.1) before moving on to the relation between
vector quantity corresponding to net flow, but this is of lim- fluence and absorbed dose (Section 5.4.2) via the impor-
ited use in dosimetry.* Planar fluence is a particularly useful tant concept of charged particle equilibrium (Section 5.5).
concept when dealing with beams of charged particles. In Subsequently, the corresponding fluence–dose relation for
certain situations, e.g. at small depths in a parallel electron charged particles will be derived (Section 5.6). Armed with
or proton beam, to first order, the planar fluence remains these fluence–absorbed dose relationships, we will then
constant as the depth increases; in contrast, fluence gener- be in a position to derive expressions describing detector
ally increases due to the change in direction of the charged or dosimeter response, generally known as cavity theory
particle tracks (see Section 3.7). The difference between the (Section 5.7).
two quantities is illustrated in Figure 5.5.
5.4.1 Relation between Photon Fluence and Kerma
5.3.6 Fluence Rate A small (or elementary) volume ΔV of material ‘med’, mass
All the quantities in Section 5.3 can be defined per unit Δm and density r = r med, is traversed by several photon tracks
dF of energy hn (see Figure 5.6).
time, e.g. the fluence rate F = and the energy fluence rate To extract energy from particle tracks and transfer it to the
dt
dY medium, we require an interaction coefficient. We start from
Y = .
dt the definition of the mass energy transfer coefficient m tr (hn )/r
(see Section 4.5.3 and ICRU 2011):
Incoming beam m tr (hn ) 1 dRtr
= (5.16)
r r dl R
where dRtr is the mean (radiant) energy that is transferred
to the kinetic energy of charged particles by interactions of
uncharged particles, of incident radiant energy R, in travers-
ing a distance dl (R = N hn for N photons traversing ΔV). It
follows that the mean energy transferred to charged particles
by the ith photon of track length dli is:
Scattering medium
dE tr,i = hn dli m tr (hn ) (5.17)

Scattered rays

FIGURE 5.5 Schematic 2D illustration of the concept of planar flu-


ence. The number of particles crossing the two horizontal lines (of
equal length) is the same. This shows that the planar fluence in the
original beam direction remains constant, whereas the fluence, i.e. the
total track length in the two circles, is clearly greater on the down-
stream side of the scattering medium. (Adapted from Whyte, G. N.,
Principles of Radiation Dosimetry, Wiley, New York, 1959.) FIGURE 5.6 Schematic representation of photon tracks, each of energy
hn, crossing an elementary volume ΔV of material ‘med’, with density
* In radiation dosimetry, the effects of individual energy deposits are rmed. (Adapted from Andreo, P., Burns, D. T., Nahum, A. E., Seuntjens,
added irrespective of the direction of the trajectories of the particles J. and Attix, F. H., Fundamentals of Ionizing Radiation Dosimetry,
giving rise to them, as Section 5.3.1 makes clear. Wiley-VCH, Weinheim, 2017.)
Chapter 5: Principles and Basic Concepts in Radiation Dosimetry 61

and therefore, the sum of energy transferred to charged-


where g med , the fraction of the energy transferred to charged
particle kinetic energy in the volume is:
particles that is converted into bremsstrahlung, is a medium-
specific average value for the photon energy spectrum and
S dE tr,i = hn S dli m tr (hn )
i i can be calculated from Equations 5.20 and 5.22.
Dividing both sides by the mass of the volume Δm, writing
5.4.2 Relation between Kerma
Dm = rDV on the right-hand side of the equation, and re-
grouping some of the terms: and Absorbed Dose
Having established a relationship between kerma and fluence
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

S dE tr,i S dli m tr (hn ) S dli æ m (hn ) ö in the previous section, if absorbed dose can be related to
tr
i
= hn i
= hn i ç ÷
Dm r DV DV è r ø med kerma, then a relationship between absorbed dose and fluence
for photons can also be established. The absorbed dose in the
Recognising that on the left-hand side the sum of energy medium, Dmed, is given by the mean value of energy imparted
transferred divided by the mass is kerma K med (cf. Equation to an elementary volume (per unit mass), whereas kerma is the
5.6) and that the sum of the tracklengths divided by the vol- energy transferred to this volume (per unit mass). However,
ume is the fluence Φ (cf. Equation 5.14), we arrive at: the charged particles can leave the elementary volume, taking
their residual kinetic energy with them. This is illustrated in
æ m (hn ) ö
K med = ç tr ÷ hn F med (5.18) Figure 5.7. Note that the quantity denoted in the figure by
è r ø med E trn is the net energy transferred to electrons within the layer,
where the medium ‘med’ has been made explicit. Alternatively, excluding that part of the initial kinetic energy converted into
in terms of the energy fluence Ψ: bremsstrahlung photons; this is equal to Etr(l − g), as we have
already seen.
æ m (hn ) ö Referring to Figure 5.7, the (net) kinetic energy leaving
K med = ç tr ÷ Y med (5.19) n
è r ø med the layer is E out , and the (net) kinetic energy entering the
layer on charged particles is E inn . The energy deposited in the
Up to this point, it has been assumed that the photons layer, ε, is given by:
crossing the elementary volume all have the same energy
e = E inn + Etrn - E out
n
(5.24)
hn. The equivalent expression for kerma for a photon spec-
trum is: where the scattered (Compton) photon and the bremsstrah-
lung photon are assumed to leave the layer without interact-
hn max
æ m (hn ) ö ing. If the electron track that leaves the layer is now
K med =
ò hn ´ (F hn )med ç tr
è r ø med
÷ dhn (5.20) compensated by an identical track entering the layer, i.e.
0
E inn = E out
n
in Equation 5.24, then it follows that:
where F hn is the photon fluence differential in energy e = E trn (5.25)
( F hn º dF /dhn ); Equation 5.20 is a fundamental relation-
ship in radiation dosimetry.

Additionally, we require the relationship between photon E nin


fluence and collision kerma, Kcol (see Section 5.3.2). Referring e– Bremsstrahlung
to Equations 5.19 and 5.20, the mass energy transfer coef-
ficient μtr/r is replaced by the mass energy absorption coeffi-
cient μen/r, as the energy absorbed excludes the fraction of the Incoming photons
e–
initial kinetic energy of charged particles that is converted to
bremsstrahlung photons (see Section 4.5.3 and ICRU 2011). E nout
For a single photon energy, we have: n
E tr

æ men (hn ) ö
( K col )med = ç ÷ hn F med (5.21) Scattered photon
è r ø med

and for a spectrum of photon energies: FIGURE 5.7 Schematic illustration of how secondary electrons lib-
hn max
erated by photons can transfer their kinetic energy E trn to a thin layer,
æ m (hn ) ö while some leave that layer and others enter the layer from outside. If the
( K col )med = ò 0
hn ´ (F hn )med ç en
è r ø med
÷ dhn (5.22) energy leaving the layer, E outn
, is exactly replaced by the energy entering,
E inn , then charged particle equilibrium is said to exist, and absorbed dose
The two types of kerma are related by: can be equated to collision kerma. This equality is represented sche-
matically in the figure: the section of the full length of the track of the
secondary electron starting outside the thin layer is exactly equal to the
(K col )med = K med (1 - g med ) (5.23) length of the two partial tracks of electrons (in bold) inside the layer.
62 Part A: Fundamentals

The equality between the kinetic energy of charged particles Range R


leaving a volume and the kinetic energy of those entering 100 100 100 100
the same volume is known as charged particle equilibrium
(CPE); Equation 5.25 expresses the fact that under CPE, the
A B C D E F G
energy imparted to the volume, ε, is equal to the net energy
Kerma
transferred to this same volume, Etrn. Absorbed dose
Dividing both sides of Equation 5.25 by the mass of the Kerma
layer (or volume element), and changing from stochastic to
average quantities, we can write: Build up Electronic
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

region equilibrium
CPE
D med = (K col )med (5.26)
Depth
Equation 5.26 expresses the important result:
FIGURE 5.8 Diagram showing the build-up to charged-particle equi-
Under the special condition of charged-particle librium for the idealised case of no attenuation of the photon beam
equilibrium, the absorbed dose is equal to the colli- and one straight electron track generated in each voxel labelled A to
sion kerma. G. (Taken from Johns, H. E. and Cunningham, J. R., The Physics of
Radiology, 4th Ed., Charles, C., Ed., Thomas Publisher, Springfield,
Replacing collision kerma by absorbed dose in Equation IL, 1983. With permission.)
5.21, for the case of monoenergetic photons we can there-
fore write: In each voxel, one electron is generated, and therefore the
kerma will be constant, as it is assumed that there is negligi-
CPE æ m (hn ) ö
D med = ç en ble photon attenuation. Only a small fraction of the electron’s
÷ hn F med (5.27)
è r ømed kinetic energy is deposited in voxel A; there is clearly no com-
pensation for the part of the track that leaves the volume, and
In the case of a spectrum of photon energies, (Φhv)med this
therefore the dose in voxel A is relatively low. In voxel B, a new
becomes:
electron starts, but in addition, there is also part of the elec-
hn max tron track that started upstream in voxel A. Hence, the dose in
CPE æ m (hn ) ö
D med =
ò
0
hn ´ (F hn )med ç en
è r ø med
÷ dhn (5.28) voxel B is higher than that in voxel A. In voxel C, the dose is
higher still. In voxel D, where the electron that started in voxel
A comes to rest, the sum of the sections of the separate electron
Equation 5.27 and Equation 5.28 are fundamental relation-
tracks is exactly equal to a complete electron track (which cor-
ships in radiation dosimetry; they will be exploited in Section
responds to the energy transferred, i.e. to kerma). Therefore,
5.7.3.
CPE is first attained in voxel D. Subsequent voxels (E, F, G
Section 5.5 explains how CPE can be realised under cer-
etc.) will contain patterns of electron tracks identical to those in
tain conditions.
D, and therefore CPE must also be present in these voxels. In
voxel D, the absorbed dose is equal to kerma (strictly, collision
kerma), and this will also be the case in the subsequent voxels
5.5 Charged-Particle Equilibrium E, F, G etc., provided that photon attenuation is negligible.
(CPE); Partial CPE In a real photon beam it is impossible to achieve true CPE;
photon attenuation reduces the photon fluence with depth,
Charged-particle equilibrium (CPE), also known as elec- and therefore the number of secondary particles (electrons)
tronic equilibrium* is said to exist in a volume V in an irradi- starting at different depths will not be constant.
ated medium: Even though in many situations strict CPE may not exist,
it can be well approximated, such as at depths beyond the
if each charged particle of a given type and energy dose maximum in media irradiated by photons below around
leaving V is replaced by an identical particle of the 1 MeV in energy. At higher energies, the equals sign in
same energy entering V. Equation 5.26 can be replaced by the proportionality sym-
bol. This situation is termed partial charged particle equilib-
Figure 5.8 illustrates schematically how CPE can be achieved rium† or PCPE (see Chapter 4 in Andreo et al. 2017):
in a photon-irradiated medium. Naturally, the figure is an
over-simplification, as in practice there will be a spectrum PCPE

of secondary-electron energies and directions. However, the D µ K col


argument is not essentially altered by showing only one elec-
tron starting in each voxel (labelled A to G in the figure) and See also Equation 5.30.
travelling in a straight line.
† Partial CPE was previously often termed transient CPE (e.g. Attix
* Other charged particles may be involved (e.g. positrons), but in 1979, 1986) but in Chapter 4 of Andreo et al. (2017), it is argued that
the photon beams employed in radiotherapy, (secondary) electrons transient is inappropriate, as it refers to time; the term quasi or partial
predominate. CPE is preferred.
Chapter 5: Principles and Basic Concepts in Radiation Dosimetry 63

TABLE 5.1 electrons give rise to an entirely negligible amount of (sec-


Approximate Thickness of Water Required to Establish Partial ondary) bremsstrahlung in water. Hence, the quantity g (see
Charged-Particle Equilibrium Equations 5.8 and 5.23) is very close to zero, and (total)
kerma and collision kerma cannot be separated. Even the
Maximum Approximate
highest-energy secondary electrons have extremely short
Energy of Thickness of Water for Approximate Photon
Photons (MeV) Equilibrium (mm) Attenuation (%) ranges; to show the build-up of dose from the surface (≈0–
0.017 cm), the depth scale in the figure only extends to
0.3 0.1 0.03 0.08 cm (equivalent to 0.08 g cm−2 for water), and over this
0.6 0.4 0.1 distance, photon attenuation is negligible. Summarising,
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

1 0.8 0.3 the assumptions involved in the pedagogical Figure 5.8 are
2 2.5 0.8 amply fulfilled by the photon beams of Figure 5.9.
3 8 2 In Figure 5.10, the in-water depth dependence of absorbed
6 15 4 dose, total and collision kerma is shown for a much higher
8 25 6 photon beam energy (25 MeV). If radiative interactions
10 30 7 and scattered photons are ignored, it can be shown that
15 50 9 (Greening, 1981):
20 60 11
PCPE
30 80 13
D » K col ´ (1 + m x ) (5.29)
For bremsstrahlung beams of different maximum energies; the final column
gives approximately the attenuation of photons in that thickness of water. where
Source: From Greening, J. R., Fundamentals of Radiation Dosimetry, μ is the common slope of the D, K and Kcol curves
Adam Hilger, Bristol, 1981. With permission.
x is the mean distance the secondary charged particles
carry their energy in the direction of the primary rays
while depositing it as dose (see Chapter 4 in Andreo
et al. 2017).

The constant of proportionality between dose and colli-


sion kerma is conventionally denoted by β:

D med = b med ´ (K col )med (5.30)

FIGURE 5.9 Absorbed dose and kerma divided by the incident photon
fluence (Gy cm 2) versus depth in water for broad, parallel photon beams
of energy 250 keV and quality 250 kV, obtained from Monte-Carlo sim-
ulations.* (Reproduced from Kumar, S., Deshpande, D. D. and Nahum,
A. E., Phys. Med. Biol., 60 (2), 501–519, 2015. With permission.)

Table 5.1 shows the amount of photon attenuation over


the distance in water required to achieve partial electronic
equilibrium for photon beams of different energies. It can be
seen that the degree of CPE failure increases as the photon
energy increases.
Figure 5.9 shows the depth dependence (in water) of FIGURE 5.10 Monte-Carlo calculations of the variation with depth
kerma K and dose D in relatively low-energy mono- and of dose D, kerma K and collision kerma K col along the central axis
poly-energetic photon beams. At these relatively low pho- for a broad, monoenergetic 25 MeV photon beam incident on water.
ton energies the even lower energies of the secondary The partial CPE region beyond the depth of maximum secondary
electron range, R max = 11.2 cm, is emphasised; the displacement
* It is usual to refer to the ‘quality’ (related to penetration) of a polyen- x between the collision kerma and dose curves is also indicated.
ergetic photon beam obtained via the bremsstrahlung interaction pro- Note that the scale on the ordinate is logarithmic, and therefore
cess, by stating the ‘voltage’ (e.g. kV or MV) that would produce such the values where D ≈ 0, cannot be shown. Data from Kumar et al.
a beam, i.e. of the maximum energy of the spectrum (see Sections (2015a). (Reproduced from Chapter 4 in Andreo, P., Burns, D.
22.2 and 23.2.1); using keV or MeV as unit indicates that the beam T., Nahum, E. E., Seuntjens, J. and Attix, F. H., Fundamentals of
is monoenergetic. Monte-Carlo simulation of radiation transport is Ionizing Radiation Dosimetry, Wiley-VCH, Weinheim, 2017. With
covered in Chapter 30. permission.)
64 Part A: Fundamentals

Figure 5.10 identifies the distance x and the PCPE region, The depth x1 at which kerma and dose are equal is therefore:
where the D, K and Kcol curves first become parallel to each
other. At this high photon energy, g is appreciable, resulting ln ( me /m p )
in a clear separation between kerma and collision kerma. x1 = (5.35)
me - m p
When integrated out to large depth, the area under the
collision kerma curve is closely equal to the area under the
Greening also showed that the maximum value of the dose
absorbed dose curve. However, it is clear from Figure 5.10
occurred at depth x1.
that the area under the kerma curve exceeds that under the
These features can be seen in Figure 5.10, although the
collision kerma curve, and therefore kerma violates energy
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

high energy, 25 MeV, of the photons gives rise to appreciable


conservation. This issue has been explored in depth in Kumar
secondary bremsstrahlung, causing the collision kerma to
and Nahum (2016).
cross the dose curve at a depth slightly less than that of Dmax.
Kumar et al. (2015a) propose the following approximate
For photon beams of very small cross section (‘small
expression for (D/Kcol):
fields’, see Section 5.8), there can be major departures from
(partial) CPE, causing the dose to differ significantly from
æ D ö the collision kerma at all depths. Figure 5.11 shows the
çK ÷ = b med ratio of the absorbed dose D to the kerma K for spectra
è col ø med (5.31) (for point sources) of 6 MV, 10 MV and 15 MV obtained
æ m (hn ) é ù ÷ö by Monte-Carlo simulation (see Chapter 30);* it can be
» çç 1 + eff
r
{
êë0.5 ´ Rcsda (E 0)hn } û ÷ø
med ú seen that D/K decreases rapidly for field sizes less than
è
~3 cm × 3 cm. This decrease is due to the onset of (lateral)
electronic disequilibrium as the field width drops below
where that required to encompass the lateral excursions of the
µeff is an effective attenuation coefficient that takes higher-energy secondary electrons (see Section 5.8; Scott
account of the build-up due to scattered photons et al. 2009; IPEM 2010).
and is therefore smaller than the narrow-beam Dutreix et al. (1965) contains an excellent, pedagogical
attenuation coefficient µ (see Section 4.5.1) discussion of the concept of charged-particle equilibrium.
hn is the kerma-weighted mean energy over the pho-
ton-fluence spectrum (the term in the square brack-
ets corresponds to the x of Equation 5.29)
E 0 is the mean initial energy of the (secondary) elec-
trons set in motion by these photons
Rcsda is the continuous-slowing-down-approximation
(CSDA) range of these electrons expressed as mass
per unit area (e.g. g cm–2) (see Section 3.5.3).

Greening (1981) has given an instructive analytical treat-


ment of the depth dependence of kerma and dose. A parallel
beam of high-energy photons of energy fluence Y 0 at the
surface of the medium is assumed with an effective linear
attenuation coefficient µp. The secondary electron energy flu-
ence decreases in the beam direction with an effective lin-
ear attenuation coefficient µe, and secondary bremsstrahlung
can be ignored. Greening showed that the absorbed dose at
depth x1 is given by:
x1
m tr me FIGURE 5.11 Monte-Carlo-derived ratios of absorbed dose to kerma
Dx1 = Y 0
r ò exp(-m x)exp éë-m (x
0
p e 1 - x )ùûdx (5.32) (D/K) on the central axis at 10 cm depth in a large cylindrical water
phantom for megavoltage beam qualities versus field size defined at 100
cm source-to-phantom surface distance; the sides of the square fields
yielding: are 0.25 cm, 0.5 cm, 0.75 cm, 1 cm, 1.5 cm, 2 cm, 3 cm and 10 cm,
respectively; the error bars are ±2 standard deviations (k=2) and corre-
m tr me spond to statistical (type A) uncertainties. (Reproduced from Kumar, S.,
Dx1 = Y 0 éexp(- m p x 1) - exp(- me x 1)ùû (5.33)
r (me - m p ) ë Deshpande, D. D. and Nahum, A. E., Phys. Med. Biol., 60 (2), 501–519,
2015. With permission.)

Now the kerma at depth x1 is given by:

m tr * The quantity shown in Figure 5.11 is D/K rather than D/Kcol, and
K x1 =Y 0 exp(- m p x 1) (5.34) therefore, the ratio at large field sizes and high energies is slightly below
r unity, as kerma includes secondary bremsstrahlung (cf. Figure 5.10).
Chapter 5: Principles and Basic Concepts in Radiation Dosimetry 65

The total local energy transfer to an elementary volume dV


5.6 Relation between Fluence and is therefore:
Dose for Charged Particles S dEl ,i = S el S dli
i i
5.6.1 Stopping Power and CEMA Dividing both sides by the mass of the elementary volume
The previous two sections have concerned uncharged par- and replacing dm by r dV on the right-hand side of the
ticles, i.e. photons.* In this section, equivalent relationships equation:
are derived between the fluence of charged particles (assumed
S dEl ,i S el S dli æ S ö é S dl i ù
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

here to be electrons) and absorbed dose. i


= i
= ç el ÷ ê i ú (5.37)
Figure 5.12 shows, schematically, a number of electron dm r dV è r ø med êë dV úû
tracks crossing a small volume ΔV of medium ‘med’, density
r = r med and mass Δm. In contrast to photons, for charged The term in the square brackets is the total tracklength per
particles the relevant quantity is stopping power, the energy unit volume, i.e. the average electron fluence‡ in the elemen-
lost (quasi-continuously) per unit track length (see Section tary volume (Equation 5.14), which we can write as Φmed.
3.2.2); we denote this type of local energy loss by dE ℓ to Thus:
distinguish it from dEtr employed for photons. We are pri-
S dE ,i æS ö
marily interested in energy deposited locally in the volume, i
= ç el ÷ Fmed (5.38a)
so the appropriate quantity is the electronic stopping power, dm è r ø med
Sel, rather than the total stopping power, as the latter includes The quantity cema§ for charged particles was proposed by
energy lost in the form of bremsstrahlung photons, and these Kellerer et al. (1992) as an equivalent to kerma for uncharged
would escape the volume of interest† (this is analogous to the particles. ICRU (2011) defines cema as:
difference between collision kerma and kerma).
For any of the electron tracks in ΔV, we can write ‘the quotient of the mean energy lost in electronic
interactions in a mass dm of a material by dm,
dEl = S el dl (5.36) except secondary electrons incident on dm.’

‘Secondary electrons’ refers here to the knock-on electrons


(or delta-rays) generated by the incident primary electrons
(see Section 3.2) – their kinetic energies are automatically
included in dEl through the use of the unrestricted elec-
tronic stopping power ¶ in Equation 5.38a. Cema is given by
ΣdEl/dm and is therefore equal to the product of electron
fluence and mass electronic stopping power:

æS ö
C med = F med ç el ÷ (5.38b)
è r ø med

Cema is not necessarily equal to absorbed dose, as some of


the delta-rays can leave ΔV, in an analagous way to second-
ary electrons when photons constitute the primary radiation.
For cema to equal absorbed dose, we require that charged-
particle kinetic energy leaving the elementary volume is
replaced by an exactly equal amount entering this volume
FIGURE 5.12 A schematic representation of electron tracks crossing an and being deposited in it or imparted to it (cf. Figure 5.7).
elementary volume ΔV of material ‘med’ with density r; it is assumed This is known as delta-ray (or knock-on electron) equilibrium.
that the mean chord length across the volume is much smaller than the If this form of equilibrium is present, then absorbed dose can
ranges of the electrons (represented as straight lines). Bremsstrahlung be equated to cema, and we can write, for a medium ‘med’,
photons (labelled k i) are emitted from electron tracks 2, 5 and 6.
(Adapted from Andreo, P., Burns, D. T., Nahum, E. E., Seuntjens, d - eqm
æS ö
J. and Attix, F. H., Fundamentals of Ionizing Radiation Dosimetry, D med = F ç el ÷ (5.39)
Wiley-VCH, Weinheim, 2017.) è r ø med

* Neutron interactions are covered in Chapter 4. Neutron dosimetry has


some similarities with photon dosimetry (e.g. the use of the kerma con- ‡ If the photon spectrum includes energies above the threshold for pair
cept). However, it is of limited interest in current radiotherapy and is production (1.022 MeV, see Section 4.3.3) the particle fluence will
therefore not covered in this book (see Chapter 19 in Andreo et al. 2017). include positrons as well as electrons.
† Note that the secondary electrons liberated within the elementary § The name ‘cema’ is the acronym for ‘converted energy per unit mass’.

volume by bremsstrahlung photons are included in the total electron ¶ The term ‘unrestricted’ indicates that there is no restriction on the
fluence; some of the incident electrons may also originate from brems- magnitude of the energy transfer, via kinetic energy of delta-rays or
strahlung losses elsewhere in the medium. atomic excitation, to the medium (see Sections 3.2.4, 3.2.5 and 5.7.5).
66 Part A: Fundamentals

or, in the case of a spectrum of electron energies: theory, which relates Ddet to Dmed, is known as cavity theory.
The principal aim of cavity theory is to determine the factor
E max f Q, given by:
d - eqm
æ S el (E ) ö
D med =
òF E ç
è r ø med
÷ dE (5.40)
D 
0 f Q =  med  (5.41)
 Ddet  Q
where ΦE is the electron fluence, differential in energy (see
for an arbitrary detector ‘det’, in an arbitrary medium ‘med’,
Section 5.3.3).
irradiated by radiation of quality Q (here, photons or elec-
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

trons). Figure 5.13 illustrates the situation schematically.


5.6.2 Delta-Ray Equilibrium This section contains the key ideas and expressions involved
in applying cavity theory to practical radiotherapy dosimetry.
Naturally, delta-ray equilibrium must always exist if charged- Extensive use is made of some of the results obtained in the
particle equilibrium exists. However, if the primary radia- previous sections of this chapter – in particular, the concept
tion consists of charged particles, CPE can never be achieved of particle fluence and its relationship to absorbed dose. A
except in the rather special case of uniformly distributed β more detailed treatment is given in Chapter 9 of Andreo et
sources in a large medium. For the beams of high-energy al. (2017). Burlin (1968) is also worth consulting.
electrons used in radiotherapy, the energy of the primary The ‘Fano theorem’ is arguably the most fundamental
electrons decreases continuously with depth, and hence there form of cavity theory and is covered first (see Section 5.7.2).
can never be CPE. However, the ranges of most delta-rays are This is followed by the two most important situations
extremely short (see Section 3.2), and almost all the energy is where it is possible to derive an exact expression for f Q.
transferred through electronic losses; i.e. the cema is mostly The first of these is that of a so-called ‘large detector’ in
deposited locally. This is demonstrated by the closeness a photon-irradiated medium (see Section 5.7.3); ‘large’
of L D S el to unity even for very small values of the cut-off refers to its size in relation to the ranges of the second-
energy Δ (see Figure 3.8).* Consequently, delta-ray equilib- ary particles, i.e. electrons. The Bragg–Gray cavity theory,
rium is generally present to a high degree in an electron- for ‘small’ detectors, is covered in Section 5.7.4, and the
irradiated medium. more exact Spencer–Attix cavity theory in Section 5.7.5
One situation where delta-ray equilibrium is definitely Ion-chamber correction factors or perturbations to the
not a good approximation is very close to the surface in a Bragg–Gray-based dose ratio are described in Section
medium irradiated by an electron beam. The appreciable 5.7.6. The approximate Burlin or ‘general’ cavity theory,
range, in the forward direction, of the most energetic delta- for intermediate-size detectors, is described in Section
rays results in a small but discernible delta-ray build-up (see 5.7.7. A Monte-Carlo-based example of the response of a
Figure 3.16).† ‘fictitious’ detector, involving all three cavity theories, is
given in Section 5.7.8. The final, separate Section 5.8 is
devoted to the dosimetry of ‘small’, sub-equilibrium mega-
5.7 Cavity Theory voltage photon beams.

5.7.1 Introduction
When a measurement is made with a detector, the detector 5.7.2 The Fano Theorem
material will in general differ from that of the medium into Fano (1954) proved in a formal, mathematical way that:
which it is introduced. As stated in Section 5.1, the signal
from a radiation detector used as a dosimeter is proportional In a medium of given composition exposed to a
to the energy absorbed in its sensitive material and thus, to uniform fluence of primary radiation (such as x-rays
the absorbed dose in this material, Ddet. The step from the or neutrons) the fluence of secondary radiation is
raw detector signal to Ddet at some reference radiation quality, also uniform and independent of the density of the
generally known as calibration, is dealt with in Chapter 19. medium as well as of the density variations from
The detector can be thought of as a cavity introduced point to point.
into the (uniform or ‘undisturbed’) medium under irradia-
tion; the name ‘cavity’ stems from the fact that gas-filled Figure 5.14 illustrates schematically the situation to which
ionisation chambers dominated the early development of the the Fano theorem applies. A medium large enough for CPE
subject (Greening 1981; Whyte 1959), and the associated to be established is irradiated by indirectly ionising radiation,
giving rise to secondary-particle tracks (shown as straight
lines for simplicity). The central region has a higher density
* See Section 3.2.5 for definition of L Δ . but with its atomic composition identical to the region sur-
† As illustrated in Figure 3.16 and Figure 5.5, the increase in the pri-
rounding it. More electron tracks per unit volume are started
mary electron fluence with depth, due to the increasing obliquity of in this central region, but each track is correspondingly
the electron tracks, remains the prominent phenomenon to explain
the increase of dose with depth in a medium irradiated by a ‘broad’
shorter due to the higher stopping power. Fano’s theorem
electron beam. predicts that the electron fluence (i.e. total track length per
Chapter 5: Principles and Basic Concepts in Radiation Dosimetry 67

Quality Q Quality Q
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

Ddet µ signal Dmed = fQDdet

(a) (b)

FIGURE 5.13 The general situation of a detector introduced into a medium (a), yielding the dose averaged over the sensitive volume of the detec-
tor, Ddet , for a given exposure to radiation of quality Q and then (b) being converted into the dose Dmed at ×, in the absence of the detector, by
multiplying by the cavity-theory factor f Q.

Electron tracks in medium with density variation beam attenuation. Reference is made to the Fano theo-
rem in Section 5.7.4.3 and in particular, in explaining
ion-chamber ‘Bragg–Gray breakdown’ in small, sub-
equilibrium megavoltage photon fields (see Section 5.8.3;
Figure 5.35).
The Fano theorem implicitly assumes that the mass
stopping power of a medium is independent of its density.
At relativistic energies (i.e. above around 0.5 MeV for
electrons), this is no longer the case due to the density
(or polarisation) effect (see Section 3.2.3). However, this
High density ‘ violation’ of the Fano-theorem preconditions probably
­
has no practical importance. Fortunately, Bragg–Gray cav-
ity theory (Section 5.7.4) can be applied to the response
of a low-density cavity (e.g. air) in megavoltage photon
beams.
Low density

FIGURE 5.14 Schematic illustration of the situation to which the Fano 5.7.3 Cavity Theory for ‘Large’ Photon Detectors
theorem applies: the total track length of secondary electrons per unit Imagine a uniform medium irradiated by a beam of monoen-
volume, i.e. the electron fluence, remains the same, although the den-
sity of the photon-irradiated medium may vary. The electron tracks are
ergetic photons with energy fluence Y med,z at depth z.
shown as straight lines for simplicity. Provided that this depth is sufficient for partial CPE to be
established, by combining Equations 5.21 and 5.30, the dose
unit volume) in the central region will be exactly the same as in the uniform medium at depth z can be written as:
that in the outer region, irrespective of the size or density of
the central region. Greening (1981) and Chapter 9 in Andreo
PCPE æm ö
D med,z = b med Y med,z ç en ÷ (5.42)
et al. (2017) contain more detailed treatments of the Fano è r ø med
theorem.
An example of the application of the Fano theorem is Suppose now that a detector is placed with its centre at
in the dosimetry of low-energy x-ray beams, where it is depth z and that the volume of the sensitive material of the
not possible to make a gas cavity small enough to fulfil detector, ‘det’, is large enough for there to be PCPE in this
the Bragg–Gray condition (see Section 5.7.4, especially volume; i.e. its dimensions are greater than the maximum
Table 5.2). If the low-density (gas) cavity had the same range of the secondary electrons generated in this material.
atomic condition as the surrounding medium (including Figure 5.15 illustrates the situation schematically.
that of the chamber wall), then there would be no limita- The photons interact with the detector material, giving
tion on the size of the cavity (e.g. water vapour surrounded rise to secondary electron tracks. Within a narrow ‘rind’
by water-equivalent chamber walls in a water phantom). region on either side of the detector edge there cannot be
Of course, in practice, corrections would be required for (partial) CPE as the electrons in this region are generated
68 Part A: Fundamentals

med
D med,z æ men ö
= çç ÷÷
Ddet è r ødet
hn max (5.46)

=
ò0
hn max
hn ´ (F hn )med,z ( men (hn )/r )med dhn

ò0
hn ´ (F hn )med,z ( men (hn )/r )det dhn
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

This expression, often written simply as (men r)med, det , is


known as the mass–energy absorption coefficient ratio. The
photon fluence spectrum at a depth in the medium can
generally only be obtained by Monte-Carlo simulation (see
Chapter 30).
As has already been noted, the detector should be small
FIGURE 5.15 The deposition of energy by photon-liberated second-
ary electrons in a (wall-less) detector that is ‘large’ compared with the
with respect to the attenuation of photons within it, i.e. to
secondary electron ranges (for simplicity represented by straight lines). the photon mean free path s = 1/μ; otherwise, it is difficult to
(Reproduced from Andreo, P., Burns, D. T., Nahum, E. E., Seuntjens, fulfil the requirement that the energy fluence in the detector
J. and Attix, F. H., Fundamentals of Ionizing Radiation Dosimetry, is negligibly different from that in the medium. It may be
Wiley-VCH, Weinheim, 2017. With permission.) necessary to add a perturbation correction factor to correct
for the differences between photon attenuation in the detec-
partly in the surrounding medium and partly in the detector.
tor and that in the medium it replaces. Mobit et al. (2000)
Away from the detector edge PCPE will exist in the detector
have made a detailed Monte-Carlo-based study of this issue.
material. The absorbed dose in the detector, with the bar
The use of small thermoluminescent dosimeters (TLDs)
indicating that this is an average over the detector volume,
(see Section 17.2) in kilovoltage photon fields is an example of
will therefore be given by:
a large-detector situation; the maximum secondary electron
PCPE æm ö range for kV x-rays in the common TLD material lithium
Ddet = b det Y det ç en ÷ (5.43) fluoride is of the order of 0.5 mm (i.e. less than the dosimeter
è r ødet thickness). Figure 5.16 shows how the mass–energy absorp-
tion coefficient ratio (Equations 5.45 or 5.46) varies with
where the energy fluence is that in the detector material photon energy for three different thermoluminescent materi-
(strictly an average over the detector volume) and the mass– als. The rapid variation below 100 keV is due to the strong
energy absorption coefficient applies to the detector mate- dependence of the photoelectric effect on atomic number
rial ‘det’. The ‘large detector’ condition is essentially fulfilled (see Section 4.3.1). The virtually flat portion of the curves
when the volume of the no-CPE ‘rind’ region is only a small corresponds to the energy region where the Compton effect
fraction of the total volume of the sensitive material of the dominates, and therefore, the interaction coefficients depend
detector. only on electron density (see Section 4.3.2), their ratio being
Combining Equations 5.42 and 5.43, we arrive at: independent of photon energy.
D med,z æ b med öY med,z (men r)med 1.2
=ç ÷ (5.44)
Ddet è b det ø Y det (men r)det
If it is assumed that the photon (energy) fluence in the detec-
tor differs negligibly from the photon (energy) fluence exist- 1.1 A
(men /r) TLD / (men /r)H O
2

ing in the undisturbed medium, then we can set Ψdet equal to


Ψmed,z; this is justified for detector dimensions much smaller
than the photon mean path, 1/µdet, or if the detector and the 1.0
medium have very similar density and atomic composition. If
C
additionally the ratio of βs is negligibly different from unity,
then the ratio of the absorbed doses, or the cavity theory fac- 0.9
tor f Q , can be simplified to:
B
D (m r)med 0.8
fQ = med,z = en (5.45) 0.01 0.1 1 10
Ddet (men r)det
Photon energy (MeV)
In all real situations, there will be a spectrum of photon ener-
gies. This will be the case even for a monoenergetic photon FIGURE 5.16 Variation of the mass–energy absorption coefficient ratio
source, as Compton-scattered photons of lower energy will relative to water for different thermoluminescent materials as a func-
tion of photon energy. A: Lithium fluoride, B: lithium tetraborate
always be present at a depth in the medium. From Equations
Li 2B 4O7, C: Li 2B 4O7 + 0.3% Mn. (From Greening, J. R., Fundamentals
5.28 and 5.45, the full expression becomes: of Radiation Dosimetry, Adam Hilger, Bristol, 1981. With permission.)
Chapter 5: Principles and Basic Concepts in Radiation Dosimetry 69

Tissue Bone electron transport, is also shown. The reader will understand
after studying Section 5.7.4 that the ratio of the doses exactly
Dose on either side of the interface will be given by the (mass)
or Kerma electronic stopping-power ratio, as the electron fluences must
be equal. Johns and Cunningham (1983) discuss in detail
interface effects in photon-irradiated media.
In the megavoltage energy range, the appreciable ranges
of the secondary electrons make it impossible for radiation
detectors to fulfil the large photon detector condition without
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

becoming impractically large. We shall see that the (μen/r)


ratio appears as a component in the (approximate) expression
FIGURE 5.17 Variation of absorbed dose at an interface between soft for the perturbation factor that corrects for the non-water-
tissue and bone, irradiated by kilovoltage photons but assuming neg- equivalence of ionisation chamber walls (Section 19.4.2.3)
ligible photon attenuation. There is charged particle equilibrium in
and more significantly, in the formalism for absorbed dose
both media except very close to the interface, where electron-scattering
effects will occur. The dashed lines indicate kerma where it differs from
determination in kilovoltage x-ray beams (Section 19.8).
absorbed dose.

The large photon cavity result can be exploited to calculate 5.7.4 Bragg–Gray Cavity Theory
the change in the absorbed dose close to an interface, e.g. 5.7.4.1 Introduction
going from tissue to bone in a patient irradiated by a low-
energy x-ray beam. Figure 5.17 is only schematic; the effect We will now focus on a different class of detector, one that
of photon attenuation, which is appreciable at low ener- is small compared with the ranges of the electrons.* In
gies, has been assumed to be negligible. The equilibrium- Figure 5.18a, a uniform medium is irradiated by a photon
dose levels in the tissue and the bone will be proportional beam; the tracks of the secondary electrons are shown sche-
matically. In Figure 5.18b, a detector has been introduced that
to (men /r)tissue and (men /r)bone, respectively (evaluated over
is ‘small’ in the sense that the majority of the electrons lose only
the photon energy-fluence spectrum). Extremely close to
a tiny fraction of their energy in passing through this detector.
the interface (i.e. within the range of the maximum energy
In the situation depicted in the figure, the size and compo-
secondary electrons), there cannot be PCPE; the increased
sition of the detector are such that, in any direction, the path
electron backscattering from the bone, due to its higher
through the detector corresponds to an entirely negligible
effective atomic number (see Section 3.6), produces the com-
fraction of the build-up depth required for the establishment
plex behaviour indicated by the dose curves on either side of
of (partial) CPE. Therefore, the absorbed dose in this small
the interface. Kerma, which, by definition, is unaffected by
detector is definitely not given by the product of photon

Photon Photon
beam beam

Uniform medium
(a) (b)

FIGURE 5.18 A (megavoltage) photon beam is incident on a uniform medium (a). The tracks of the secondary electrons liberated in the medium
are indicated schematically. A low-density detector is inserted into the medium (b). If the electron-fluence in the detector is negligibly different (in
energy distribution and magnitude) from that in the undisturbed medium at the same position, then the cavity can be said to behave in a ‘Bragg–
Gray’ manner. (Reproduced from Andreo, P., Burns, D. T., Nahum, E. E., Seuntjens, J. and Attix, F. H., Fundamentals of Ionizing Radiation
Dosimetry, Wiley-VCH, Weinheim, 2017. With permission.)

* The primary radiation can be either electrons or photons; in the latter


case, the electrons are termed secondary.
70 Part A: Fundamentals

energy fluence and (μen/r)det. Instead, the absorbed dose in ought to negligibly modify the number, energy and direc-
the detector material and in the medium can be derived from tion of electrons present in a photon-irradiated medium
the respective electron fluences present in these materials. (there were no electron beams at the time). Gray showed
that the ratio of energy lost per unit mass in the medium
to the energy lost per unit mass in the gas was equal to the
5.7.4.2 Theory
ratio of the mass electronic stopping powers, medium to
From Equation 5.39, the ratio of doses, medium to detector, gas, smed,gas, and then made the assumption that the ratio of
at the depth z in the medium is given by energy absorbed in the respective media was also equal to
smed,gas. This is equivalent to assuming delta-ray equilibrium
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

D med,z F med,z ´ (S el r)med (see Sections 5.6.1 and 5.6.2) and is discussed further in
= (5.47)
Ddet F det ´ (S el r)det Section 5.7.5.
In radiotherapy, we also deal with beams of electrons and
If it is assumed that the electron fluence in the (sensitive vol- heavy charged particles. In these cases, the primary particles
ume of the) detector is identical to the electron fluence exist- are electrons (or protons or carbon ions), and there are no
ing in the ‘undisturbed’ medium, i.e. photon-generated secondary electrons (except for the small
bremsstrahlung contribution). The Bragg–Gray idea can also
F det = F med,z (5.48) be applied to detectors in these charged-particle beams.

then Equation 5.47 can be simplified to


5.7.4.3 When Does a Detector Behave
D med (S el (E ) r)med in a Bragg–Gray Manner?
= º s med,det (5.49)
Ddet (S el (E ) r)det For a detector to respond as a Bragg–Gray (BG) cavity:
This expression is known as the (mass) stopping-power ratio, The charged-particle fluence in the sensitive vol-
med
generally written as smed,det or alternatively as [S el (E )/r]det . ume of the detector (or cavity) must be identical in
In all practical cases, there will be a spectrum of electron magnitude and energy distribution to that in the
energies, and the full expression for the stopping-power ratio uniform medium at the reference point.
becomes
We will now consider the requirements for a ‘cavity’ (i.e. a
E max detector) to behave in a Bragg–Gray manner in a charged-
ò
pr
(F E )med ,z ´ (S el (E ) r )med dE particle or photon field.
D med
= 0
E max
º s BG
med ,det (5.50)
Ddet
ò
pr
(F )
E med ,z ´ (S el (E ) r)det dE I. The cavity must be small compared with the ranges of
0
the charged particles crossing it.
where the energy dependence of the electronic stopping
power has been made explicit, and it is understood that This is equivalent to requiring that the energy losses of the
(ΦE)med,z refers to the medium in the absence of the detector in charged particles crossing the cavity are a very small fraction
both the numerator and the denominator. It is emphasised of the kinetic energies of the charged particles, i.e. DE  E .
that this is the fluence of primary electrons only* – no delta- It is difficult to convert this into concrete numbers, but we
rays are involved; hence the superscript ‘pr’. For reasons that will assume that DE £ 0.01E , i.e. a maximum 1% energy loss
will become apparent in Section 5.7.5, the stopping-power in traversing the detector.
ratio evaluated according to Equation 5.50 is convention- For a 1 MeV electron, with (dE/ds)tot = 1.86 MeV cm 2 g−1 in
ally denoted by s med, water (see Tables L2 at the end of the book), this translates
det (Nahum 1978; ICRU 1984b; Andreo
BG

et al. 2017). into a detector thickness <5.5 × 10 −3 g cm−2; this becomes


Detectors that behave in this fashion in photon-irra- <4.7 × 10 −2 g cm−2 for a 10 MeV electron, i.e. still imprac-
diated media, i.e. that respond to, but do not ‘disturb’, ticably thin for detector sensitive material of unit density or
the electron fluence that exists at some reference point in greater (but see Figure 5.20). For a 4 MV x-ray beam, the
the irradiated medium in the absence of any detector, are average secondary electron energy is of the order of 0.5 MeV,
known as Bragg–Gray cavities, after the pioneering work and the ‘Bragg–Gray’ detector would have to be even thinner.
of Bragg (1912) and Gray (1929, 1936). As we have seen
If the primary particles are photons, then there is an addi-
in the introduction to this chapter, gas-filled detectors
tional requirement:
dominated all the early radiation measurement techniques.
Bragg presented qualitative arguments and Gray reasoned II. The absorbed dose in the cavity must originate from
in a more quantitative way that a small gas-filled cavity the charged particles incident on it.
* Although the electrons liberated in photon interactions with matter
This means that any contribution to the dose in the cavity
(e.g. Compton scattering) are conventionally termed secondary, in this
context it is useful to label these (e.g. Compton) electrons as primary due to photon interactions in the cavity must be negligible. It
to distinguish them from the delta-rays (or knock-on electrons) liber- is essentially a corollary to the first requirement. If the cav-
ated by these primary electrons while slowing down. ity is small enough to fulfil requirement I, then the build-up
Chapter 5: Principles and Basic Concepts in Radiation Dosimetry 71

of dose due to interactions in the cavity material itself will fluence in the medium ‘displaced’ by the gaseous detector
also be very small. If, however, this build-up is appreciable, result in deviations from Bragg–Gray behaviour. However,
then the charged-particle fluence in the cavity will differ these ‘perturbations’ can be minimised by intelligent ionisa-
from that in the undisturbed medium, thereby violating tion chamber design – see Section 5.7.6.
Bragg–Gray behaviour (except when the detector material It should be noted that the atomic composition and/or
and the medium are perfectly matched – a Fano-theorem- density of the detector cavity (i.e. the part of the detector giv-
like situation). ing rise to the signal) is implicitly assumed to differ from that
At this point, it is useful to distinguish between solid-state of the medium. If this is not the case, then rather obviously,
detectors – see Chapter 17 (e.g. diamond, diode, TLD, liquid there will be no difference in the respective electron-fluence
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

ionisation chambers) and gas-filled detectors – see Chapter spectra under any circumstances; consequently, a (real or
16 (i.e. air-filled ionisation chambers). In the former case, the imagined) perfectly medium-equivalent detector will always
fulfilment of the requirements means that the dimensions of fulfil Bragg–Gray conditions, as Bouchard et al. (2015) have
any such detector must be extremely small, especially in the emphasised.
beam direction (e.g. single-crystal diamond) – any variation An air-filled ionisation chamber of the dimensions used
in dose either across the detector or across the volume of in radiotherapy (see Chapter 16), irradiated by a megavolt-
medium replaced by the detector will therefore be negligible. age photon beam of a ‘standard’ (i.e. not ‘small’) field size,
In the case of a photon field, there is therefore no need for is a clear case of a Bragg–Gray cavity. Figure 5.19 compares
(partial) CPE; such an extremely small detector will behave the (secondary) electron-fluence spectrum in a small water
in a Bragg–Gray manner even in the build-up region of a voxel to that in a small air cavity at the same position in a
high-energy photon beam. water phantom irradiated by a 6 MV x-ray beam of field size
3 cm × 3 cm; the two curves lie on top of one another.
Consider now a gas-filled detector (with a medium-equivalent Figure 5.20 compares (total) electron-fluence spectra in
wall for simplicity), e.g. ionisation chambers with volumes a (synthetic) diamond detector of 2.20 mm diameter and
of practical dimensions of a few millimetres (see Tables 16.2 0.001 mm height/thickness with that in a small volume of
and 16.4). The volume of medium ‘replaced’ by such a detec- water in a water phantom irradiated by a 6 MV x-ray beam.
tor will in general no longer be extremely small, but due to For the 10 cm × 10 cm field, the spectra match ‘perfectly’, and
the extremely low density of the gas (e.g. air), there is no the water-to-detector electron-fluence ratio, integrated over
problem meeting requirement II at megavoltage qualities the spectrum, is within 0.2% of unity. Thus, this extremely
(see Table 5.2 and Figure 5.22). However, dose variations thin ‘solid-state’ detector also behaves as a quasi-perfect
across the ‘displaced medium’ can be appreciable (especially Bragg–Gray detector at this field size and beam quality.‡
in the build-up region in any beam and at any depth in very
small fields – see Section 5.8.3). Therefore, in this important
special case of a (very) low-density detector, CPE is required*
in order to make dose variation across the ‘replaced’ medium
manageably small† (see Sections 5.7.6 and 19.4.2). An alter-
native way of arriving at the necessity for (partial) CPE is to
invoke the Fano theorem (see Section 5.7.2).

Summarising, for a gas-filled detector when the primary


particles are photons there is the additional requirement:

III. (Partial) Charged-particle equilibrium should be


present in the medium at the position of the detector

The fact that in real megavoltage beams the equilibrium can


only ever be ‘partial’ due to photon attenuation (see Section
5.5, especially Figure 5.10) gives rise to minor departures
from perfect Bragg–Gray behaviour, corrected by ‘perturba-
FIGURE 5.19 Monte-Carlo-derived electron-fluence spectra (all gen-
tion’ correction factors – see Sections 5.7.6 and 19.4.2. erations of secondary electrons) scored in a ‘point like’ water ‘voxel’
If the primary radiation is a charged-particle beam, then (0.5 mm diameter, 0.5 mm thickness) and in a ‘PinPoint’ air cavity
there can never be CPE; variations in the charged-particle (2.3 mm diameter, 2.6 mm length) both at 5 cm depth on the beam
central axis in a large water phantom irradiated by a 6 MV photon beam
of field size 3 cm × 3 cm. (Reproduced from Kumar, S., Fenwick, J. D.,
* Greening (1981) emphasised that Gray’s original theory required Underwood, T. S., Deshpande, D. D., Scott, A. J. and Nahum, A. E.,
CPE. Phys. Med. Biol., 60 (20), 8187–8212, 2015. With permission.)
† In real photon beams, there will only ever be partial CPE due to pho-

ton attenuation – see Section 5.5. This minor departure from ‘perfect’
CPE causes the response of small air cavities to deviate from ‘perfect’ ‡ The Bragg–Gray-like response of this diamond detector is to be expected,
Bragg–Gray behaviour; this is generally corrected by either a displace- given that its thickness in the beam ­d irection (3.51 × 10 −4 g cm−2)
ment perturbation factor or a shift in the effective point of measure- easily fulfils the detector thickness ≤5.5 × 10 −3 g cm−2 requirement
ment (see Section 19.4.2.2). derived earlier.
72 Part A: Fundamentals
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

FIGURE 5.20 Monte-Carlo-calculated electron-fluence spectra (all


FIGURE 5.21 Variation of the unrestricted mass electronic stopping-
generations of secondary electrons) at 10 cm depth in water irradiated
power ratio, water to medium, as a function of electron energy, evaluated
by a 6 MV beam in 10 cm × 10 cm and 0.5 cm × 0.5 cm fields. The
from Equation 5.49 for materials of dosimetric interest. (Reproduced
red (wavy) lines represent the spectrum in the extremely thin diamond
from Andreo, P., Burns, D. T., Nahum, E. E., Seuntjens, J. and Attix,
detector (see text), the black lines represent the spectrum at the same
F. H., Fundamentals of Ionizing Radiation Dosimetry, Wiley-VCH,
depth in a small volume of water. (Reproduced from Benmakhlouf, H.
Weinheim, 2017. With permission.)
and Andreo, P., Med. Phys., 44 (2), 713–724, 2017. With permission.)

The figure also shows the corresponding data for a ‘small’ due to the reduction in electronic stopping power brought
0.5 cm × 0.5 cm field – in such a field, PCPE will not be about by the density or polarisation effect in condensed
present (see Section 5.8.2). The small difference between the media (i.e. water compared with air) at relativistic ener-
electron fluence in the very thin diamond detector and that gies – see Section 3.2.3 and especially Figure 3.6. Values
in the undisturbed medium indicates at most a minor depar- of smed,air for other media ‘med’ would have shown a sim-
ture from Bragg–Gray behaviour. This is an example of how a ilar strong dependence on electron energy. The marked
very small (non-gaseous) detector can exhibit a Bragg–Gray- dependence of s w,air on electron energy translates into the
like response in the absence of CPE, as discussed previously. strong depth dependence of air-filled ionisation chamber
What is the situation for a high-energy electron beam? The response (essentially D w/D air) in a water phantom irradi-
(primary) electrons will easily cross the small air cavity of a ated by a megavoltage electron beam (see Figure 19.9).
typical ionisation chamber and photon interactions in the sen- By contrast, in a megavoltage photon beam, s w,air exhibits
sitive volume of the detector (including any bremsstrahlung- an entirely negligible depth dependence (except possibly
induced interactions) will clearly not be a problem. However, in the build-up region), as partial CPE ensures that the
CPE is never present in a medium irradiated by an electron secondary ­ electron energy distribution remains essen-
beam (see Section 5.5). To summarise a moderately complex tially unchanged with depth (Andreo and Nahum 1985;
situation, Bragg–Gray conditions are well approximated by Andreo 1988).
certain designs of ionisation chamber, particularly adequately
guarded ‘plane-parallel’ designs with an appropriate choice 5.7.4.5 Is a Low-density Cavity in a
of the effective point of measurement (see Section 5.7.6). Medium Irradiated by Kilovoltage
For Farmer-type cylindrical designs, the change in the angu- X-Rays a Bragg–Gray Cavity?
lar distribution of electrons with depth (in-scattering – see
Section 19.4.2.4) gives rise to medium-to-air-cavity electron- Should we expect that an air-filled ionisation chamber will
fluence differences (or ‘perturbation’) that require correction act as a Bragg–Gray cavity in a medium irradiated by a kilo-
via an appropriate factor (see Section 5.7.6). voltage x-ray beam? (see also Section 19.8 and Chapter 22).
The energies of the secondary electrons will be significantly
lower than for megavoltage photons, and the build-up depth
5.7.4.4 Bragg-Gray (Unrestricted) Stopping-Power (in water) is only a fraction of that in a megavoltage beam
Ratios, Water-to-Detector (see Figure 5.9). Will the signal produced by an ionisation
Values of the water-to-medium unrestricted mass elec- chamber still be overwhelmingly due to secondary electrons
tronic stopping-power ratio s w,air, as a function of electron incident from the surrounding medium, as opposed to elec-
energy, for a variety of media including air are shown in trons liberated by photon interactions with the air molecules?
Table 5.2 gives values of ( Dair )tot , the ratio of the dose
cav
Figure 5.21. Several features can be noted. First, for all
the materials, except aluminium and air, this ratio varies to a small air cavity, as defined in the table caption, due to
very little with electron energy, especially above ~1 MeV. photon interactions in the air to the dose corresponding to
The major exception is air. The strong decrease of s w,air full build-up in air, calculated by Monte-Carlo simulation
with increasing electron energy, starting at ~0.5 MeV, is (Ma and Nahum 1991).
Chapter 5: Principles and Basic Concepts in Radiation Dosimetry 73

TABLE 5.2
tot for an Air Cavity
Calculated Dose Ratio (Dair )cav
Exposed to Different Photon Beam Qualities
Incident Beam
(Dair )cav
tot
Tube HVL (Dair )cav
tot (5 cm depth
Potential (mm Al) (in vacuum) in water)
50 kV 1.62 0.26 0.18
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

150 kV 5.01 0.27 0.29


240 kV 7.30 0.27 0.27
240 kV 17.4 0.18 0.23
60Co 0.0037 0.0060
  4 MV 0.0014 0.0098
The calculated ratio (Dair)cav
tot is the dose due to direct photon
interactions with the air of the cavity divided by the total
absorbed dose in the cavity. The air cavity is a 6 mm diameter
and 6 mm thick cylinder, located either in vacuum or at a
depth of 5 cm in water. (Dair )cav
tot is expected to be very small in FIGURE 5.22 Ratio of the dose from photon interactions in the air
a Bragg–Gray cavity. cavity to the total dose, (Dair )cav / (Dair )tot , (in water at 5 cm depth)
Source: From Ma, C.-M. and Nahum, A. E., Phys. Med. Biol., for air cavities of typical ion-chamber dimensions and for a Farmer
36, 413–428, 1991. With permission. chamber, for low-energy photons. The radiation quality is expressed
in terms of the half-value layer (HVL) of Cu (lower axis) and Al (upper
axis). The data points are for a cylindrical air cavity of 6 mm thick-
In a Bragg–Gray cavity, ( Dair )tot ought to be very close
cav
ness and diameter with monoenergetic photons (open circles, dashed
to zero. It can be seen in the table that this is the case line), following the Monte-Carlo (MC) calculations of Ma and Nahum
(1991), and for a Farmer chamber as described in Figure 16.6 (trian-
for the 60Co γ-ray and 4 MV x-ray qualities but not for
gles, solid line: monoenergetic photons; filled circles, kV spectra: MC
any of the kilovoltage x-ray qualities. These results dem- work by J.Wulff and P. Andreo, private communication). (Reproduced
onstrate that air-filled ionisation chambers do not fulfil from Andreo, P., Burns, D. T., Nahum, E. E., Seuntjens, J. and Attix,
Bragg–Gray assumptions at any kilovoltage x-ray quality. F. H., Fundamentals of Ionizing Radiation Dosimetry, Wiley-VCH,
Figure 5.22 shows essentially the same dose ratio but over Weinheim, 2017. With permission.)
a much wider range of kilovoltage beam qualities, specifi-
cally for an air volume with the dimensions of a Farmer
chamber (see also Chapter 9 in Andreo et al. 2017). The 5.7.5 The Spencer–Attix Modification
absorbed dose fraction reaches a value of 0.6 at a half-value of Bragg–Gray Theory
layer (HVL) of 0.3 mm Al - i.e. the response of the air
cavity is approaching that of a ‘large photon detector’ (see 5.7.5.1 Introduction
Section 5.7.3). In the previous section, delta-ray equilibrium was implic-
What is the explanation for the shape of the curves in itly assumed – it is a pre-requisite for the strict validity of
Figure 5.22? Basically, the value of ( Dair )tot is critically
cav
the stopping-power ratio as evaluated in Equations 5.49
dependent on the mean range of the secondary electrons. and 5.50. Bragg and Gray effectively assumed that all the
If the electron ranges are long, the cavity will tend towards delta-rays energy arising from electronic losses in the cav-
ity was deposited in the cavity. Expressed alternatively,
Bragg–Gray behaviour, i.e. ( Dair )tot  1 (cf. Table 5.2); if
cav

the charged particles lose all their energy by continuous


the electron ranges are (very) short, ( Dair )tot will approach
cav
slowing down, i.e. in a large number of extremely small
unity, and the cavity will exhibit ‘large photon’ behaviour. energy-transfer events. Greening (1981) emphasised that
At high photon energies (i.e. high kV), Compton inter- this was one of the pre-conditions for Gray’s cavity theory
actions dominate, and the mean initial energies of the (Gray 1929).
Compton electrons are a much lower than that of the pho- Experiments by Attix et al. (1958) employing a series
ton energy (see Figure 4.5 and Section 4.3.2), with cor- of flat, air-filled ionisation chambers with walls made of
respondingly short ranges. As the photon energy decreases, C, Al, Cu, Sn and Pb respectively, i.e. ranging from low
the photoelectric-effect begins to take over, and the result- and air-like (C) to high (Cu–Pb) atomic numbers, showed
ing photo-electrons have kinetic energies essentially equal clearly that in a 412 keV γ-ray field (from 198Au), the
to those of the photons that give rise to them (see Section charge collected per unit mass of gas (or ‘mass ionisation’)
4.3.1); paradoxically, therefore there is an increase in the depended on the wall separation (which varied from tenths
initial kinetic energy of the secondary electrons. Hence, of a millimetre to 10 mm). However, if the instrument
as the photon energy decreases, Figure 5.22 shows an ini- were responding strictly as a Bragg–Gray cavity, the mass
tial increase in ( Dair )tot followed by a decrease before a final
cav
ionisation should be constant. Attix et al. suggested that
increase at very low photon energies. delta-ray transport might explain this dependence on wall
separation.
74 Part A: Fundamentals

5.7.5.2 Theory E max

Spencer and Attix (1955) proposed an extension of Bragg–


D med =
ò (F )
2D
tot
E
med
(L D (E ) r)med dE
Gray theory to take account of, in an approximate manner, (5.51)
the effect of the finite ranges of delta-rays. In their cavity the- 2D

ory, all the electrons above a cutoff energy Δ, whether ‘pri-


mary’ electrons or delta-rays, were considered to be part of
+
ò (F )
D
tot
E
med
(S S- A (E ) r)med dE

the (total) electron-fluence spectrum incident on the cavity.


All electrons liberated in the cavity (or in the same volume in where the term SS-A(E) is explained in the next paragraph.
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

the uniform medium) with energies below Δ were assumed to As the lower integration limit of the first integral is 2Δ, no
deposit 100% of their (kinetic) energy locally, i.e. inside the electron can fall below Δ in energy after a (single) electronic
cavity; Figure 5.23 illustrates this schematically. interaction; this is possible, however, for losses described
Spencer and Attix suggested that Δ be set equal to the by the second integral. Therefore, the electronic stopping
energy of electrons that have a range (in the cavity mate- power was modified to include not only the energy of the
rial) just sufficient to traverse the cavity. The energy trans- delta-rays (which by convention cannot exceed half that of
ferred to the cavity was therefore derived from the electronic the ‘primary’ electrons – see Section 3.2.2) but also that of
stopping power restricted to losses less than Δ, conventionally the ‘primary’ electrons if it is below Δ after the interaction –
denoted by L Δ (see Section 3.2.5). The dose local to the ‘cav- this ‘special stopping power’ is denoted by SS–A in Equation
ity’, in the case of the uniform medium, was given by: 5.51. The two integrals on the right-hand side of Equation
5.51 account for all the energy transferred to the medium in
E max
electronic interactions.*
ò (F ) tot
E (L D (E ) r)med dE The presence of the ‘special’ stopping power, SS–A , makes
med
D the evaluation of Equation 5.51, and hence that of the stop-
ping-power ratio, problematic. Nahum (1976, 1978) showed
(
where F Etot )med
is the total electron fluence, differential in that the contribution to the dose to the medium from elec-
energy, present in the volume of medium corresponding trons as they fell below Δ in energy, which he termed the
to the (detector) cavity. However, this expression did not dose due to the ‘track ends’, D med
TE , was approximated by:
include the contribution to the dose from ‘primary’ electrons
whose energies had dropped below Δ after the interaction. To TE
Dmed = éëF Etot ( D ) ùû éëS el ( D ) /r ùû D (5.52)
account for these contributions, Spencer and Attix modified med med
their ‘cavity integral’ to:
where
F Etot ( D ) is the electron fluence (all generations), differen-
tial in energy, at energy Δ.

The product of the first two terms in Equation 5.52 yields


the number of electrons slowing down past Δ per unit mass,
and multiplying this by Δ yields the dose due to these track
ends (Nahum 1976, 1978; Chapter 9 in Andreo et al. 2017).†
Therefore, the cavity integral for the dose to the uniform
medium becomes:

E max
SA
D med,D =
ò (F )
D
tot
E
med
(L D (E )/r)med dE
(5.53)
+ éëF tot
E (D )ùû éëS el (D )/r ùû med D
med

FIGURE 5.23 Schematic illustration of energy deposition in a volume


(‘cavity’) in the medium according to the Spencer–Attix two-group
The ‘track-end term’ constitutes between 5% and 10% of the
local-energy-deposition scheme. Photons are represented as straight total dose, depending on the value of Δ.
green lines. All generations of electrons are included in the fluence Φtot, Figure 5.24 shows the excellent agreement (at the 0.5%
down to energy Δ. The electrons deposit energy in the cavity ‘continu- level or better) between the dose computed from Equation
ously’ according to the restricted electronic stopping power L Δ; these
transfers are shown schematically as very short blue dotted lines. The
three ‘track ends’ of energy Δ (----) are accounted for explicitly. ①: * In the corresponding expression for the dose to the detector material, the
Compton electron falls below Δ. ②: Electron crosses the cavity, pro- two stopping powers in Equation 5.51 naturally pertain to this material.
ducing a bremsstrahlung photon that leaves the cavity. ③: Electron † Bouchard (2012) arrived at an identical expression for the track-end

slows down below Δ. ④: Photon crosses cavity without interacting. ⑤: dose in a more rigorous and elegant fashion. His paper contains a deri-
Electron crosses cavity, producing bremsstrahlung photon and delta-ray vation of the Spencer–Attix dose ratio starting from the Boltzmann
above Δ in energy. ⑥: Delta-ray drops below Δ in energy. Transport Equation (see Section 30.4.1).
Chapter 5: Principles and Basic Concepts in Radiation Dosimetry 75

The Spencer–Attix ratio, s med,det, D , predicts reasonably well the


variation of chamber response with gas pressure for chamber
walls of high-atomic-number materials irradiated by gamma
rays of energies high enough to fulfil Bragg–Gray conditions
(Greening 1957; Attix et al. 1958; Burlin 1962). It is empha-
sised that the simpler unrestricted or ‘Bragg–Gray’ stopping-
BG
power ratio, s med,det , predicts that chamber response should
be completely independent of gas pressure.
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

5.7.5.3 Practical Implications of
Spencer–Attix Theory
If the atomic composition and hence the effective atomic
number, Z, of the cavity and the medium are quite simi-
lar (e.g. air and water), the difference between s med,det, D and
FIGURE 5.24 Ratio of the Spencer–Attix dose to the Monte-Carlo BG
dose, (DS-A(Δ)/DMC), as a function of the Spencer–Attix Δ, for a water s med,det may be only of the order of 0.5% to 1.0% (Spencer and
volume (20 mm diameter, 6 mm thickness) along the beam central Attix, 1955; Chapter 9 in Andreo et al. 2017). For example,
axis in a cylindrical water phantom (30 cm diameter, thickness 30 cm), for Δ = 10 keV, Nahum (1978) showed that s water,air,D =10 keV was
at 2.5 cm depth for a 10 MeV electron beam, 5 cm depth for Co-60 BG
generally about 1% higher than s water,air for megavoltage pho-
γ-rays and 6 MV, 10 MV and 15 MV x-rays, and 10 cm depth for a
10 MeV monoenergetic photon beam. (Reproduced from Kumar, S.,
ton and electron beams. The SA−BG difference depends crit-
Deshpande, D. D. and Nahum, A. E., Phys. Med. Biol., 61 (7), 2680– ically on the degree of atomic-number ‘mismatch’ between
2704, 2016. With permission.) the detector and the medium, or more specifically, on the
difference between the respective I-values (see Section 3.2.2
5.53, for small values of Δ, and that scored in a Monte-Carlo and also Spencer and Attix 1955 and Chapter 9 in Andreo
simulation in a small volume of water at a depth in a large et al. 2017). As an example, Figure 5.27 indicates a differ-
water phantom irradiated by a range of radiation qualities.* ence between Spencer–Attix and Bragg–Gray formalism of
This agreement (at small Δ) is a demonstration of the validity ≈2.5% for the water-to-silicon ratio for a (simulated) gaseous
of Equations 5.52 and 5.53. Si detector in water irradiated by 6 MV photons.
The corresponding expression for the dose to the detector The value of Δ is difficult to define uniquely for any shape
cavity with a size corresponding to electron energy Δ is: of cavity given the quasi-isotropic directions of the electrons
E max as they enter the detector volume. Fortunately, s water,air,D varies

ò (F ) only slowly with Δ, which is conventionally set to 10 keV for


SA tot
Ddet,D = E (L D (E )/r)det dE
D
med
(5.54) the ionisation chambers employed in radiotherapy. Buckley
et al. (2003) found that Δ = 16 keV corresponded to the mean
+ éëF Etot (D)ùû éëS el (D) / r ùû det D chord length for a cylindrical ionisation chamber of Farmer
med
dimensions (air cavity 6.3 mm in diameter and 2 cm in
It is emphasised that the implicit assumption in Equation length) and that this value, rather than Δ = 10 keV, resulted
5.54 is that the electron-fluence spectrum (down to energy in better agreement between the Spencer–Attix stopping-
Δ) in the detector cavity is identical to that in the equivalent power ratio and the Monte-Carlo-derived dose ratio. Borg
volume of the (undisturbed) medium, cf. Equation 5.53; i.e. et al. (2000) concluded from their Monte-Carlo study that
the detector is acting in a Bragg–Gray manner. Spencer–Attix cavity theory with Δ = 10 keV was acceptable
Accordingly, the full expression for the Spencer–Attix within 0.5% for photon energies at 300 keV or above.
­stopping-power ratio (ICRU 1984b; Nahum 2009; Chapter 9 The use of the Spencer–Attix expression inherently
in Andreo et al. 2017) becomes: assumes that the electron-fluence spectra in the cavity and in
the medium are identical only down to energy Δ. This is less
med demanding than the original Bragg–Gray requirement that
æL ö D SA D
s med,det, D º ç D ÷÷ = med, the cavity and medium electron fluences be identical down
ç r SA
Ddet,
è ødet D
to the lowest electron energies present in the fluence spec-
trum. At energies below Δ, the electron fluence in the cavity
( )
E max

=
ò (F )
D
tot
E
med
(L D (E ) r)med dE + éëF Etot ( D ) ùû éëS el ( D ) r ùû D
med med material must consist overwhelmingly of (knock-on or delta)
dE + ( éëF D)
E max
electrons liberated in the cavity and the number and energy
ò (F ) ( D )ùû med éëSel ( D ) r ùû
tot tot
E (L D (E ) r)det E
D med det distribution of these electrons will inevitably be characteris-
tic of the detector material.
(5.55) BG
Should we expect smed,det, D to tend towards s med,det as the
cavity becomes larger (i.e. as Δ increases) or as it becomes
smaller (as Δ decreases)? Although it may initially seem coun-
* The reason for the decrease in D S-A(Δ)/D MC at relatively large values of ter-intuitive, it is large cavities that most closely correspond
Δ is discussed in Section 5.7.5.3. to the simpler unrestricted Bragg–Gray ratio (Nahum 1976;
76 Part A: Fundamentals

Chapter 9 in Andreo et al. 2017). The ranges of the majority


of delta-rays will be small compared with the dimensions of a
large cavity, which is consistent with the implicit Bragg–Gray
assumption of negligible delta-ray range. Conversely, if the
cavity is very small, then delta-rays generated in the medium
will be responsible for a much greater proportion of the cav-
ity dose.
However, if the cavity size is large, it may violate the
Bragg–Gray condition that a negligible fraction of the dose
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

arises from photon interactions in the cavity (see Section


5.7.4.3, requirement II). Referring to Figure 5.24, for all the
photon qualities, there is a value of Δ beyond which this ratio FIGURE 5.25 An ‘imaginary’ ionisation chamber illustrating the vari-
starts to fall significantly below unity. This is because in the ous potential sources of perturbation. (From Nahum, A. E., Phys. Med.
Spencer–Attix formalism (Equation 5.53), the integral over Biol., 41, 1531–1580, 1996. With permission.)
the electron-fluence spectrum only extends down to energy Δ;
any contributions due to photon-liberated electrons with However, as illustrated schematically in Figure 5.25, a
­initial energies below Δ are consequently not accounted for. practical ionisation chamber consists of more than a ‘bare’
This ‘missing’ dose corresponds to the second term of the air cavity. The effects of the various components of chamber
dose ratio given by Burlin cavity theory (Section 5.7.7). Note construction, indicated schematically in Figure 5.25, are gen-
that for 10 MeV electrons DS−A(Δ)/DMC remains essentially erally dealt with by introducing one or more perturbation cor-
at unity for all Δ, as the contribution due to photon-liberated rection factors into the expression for the absorbed dose, via
electrons is negligible. the composite factor ( pQ )ch in Equation 5.56. This factor cor-
Codes of Practice for reference dosimetry in megavoltage rects the ionisation chamber response for any deviations from
photon and electron beams (see Chapter 19) give values of perfect Bragg–Gray behaviour and is therefore defined by:
s w,air, D =10 keV evaluated according to Equation 5.55; these are
required to convert D air derived from the ionisation chamber D med,z = Dair,ch s med,air,D ( pQ )ch (5.56)
reading to absorbed dose to water.
where
Dair,ch  is the dose to air recorded by the ion chamber,
5.7.6 Departures from ‘Perfect’ Bragg–Gray i.e. by the instrument.
Behaviour For Ionisation Chambers
5.7.6.1 Introduction It is emphasised that the (Spencer–Attix) stopping-power
ratio s med,air,D (Equation 5.55) is derived on the assumption
As emphasised in Section 5.7.4, for a detector to behave as a of identical electron (and positron) fluence spectra in the
‘perfect’ Bragg–Gray cavity, the electron fluence in the sensi- medium and the air cavity, i.e. the Bragg–Gray condition.
tive medium of the detector must be identical (in magnitude The subscripts ‘ch’ and ‘Q’ emphasise that the perturbation
and energy distribution) to that at the position of interest factor depends on both the details of the chamber design and
in the uniform medium, i.e. in the absence of the detector. the radiation quality.
Consequently, a perfect Bragg–Gray detector would have to Denoting the fluence, differential in energy (of all genera-
be completely medium-equivalent (in composition and den- tions of electrons and positrons) in the medium, at depth z, by
sity). Under certain conditions, an air-filled ionisation cham-
ber, which can be likened to a small ‘gas bubble’, behaves as a
tot
(F Etot )med,z and that averaged over the air cavity by F E
air,ch
, ( )
good approximation to a Bragg–Gray cavity: for broad-beam and recalling the Spencer–Attix expression for absorbed dose
megavoltage photon and electron radiation, the basic Bragg– for a finite-size cavity (Equation 5.53), but without assuming
Gray conditions are adequately fulfilled by the ionisation that electron-fluence spectra in the air are identical to that in
chamber designs in common use in radiotherapy (see Section the medium, we can write:
16.3.2); in particular, the ranges in air of the vast majority
( )
E max
of ‘secondary’ electrons* set in motion by megavoltage pho- ò (F ) (L D (E ) r)med dE + éëF Etot ( D ) ùû éS el ( D ) r ù D
tot
D med, z med .z ë û med
E
med , z
= D
tons are much greater than the cavity dimensions† (Ma and
ò (F )
D air,ch E max
æ é tot ö
(L D (E ) r)air dE + ç F E ( D ) ù
úû air,ch ë el ( ) û air ÷ø
tot
éS D r ù D
Nahum 1991; Chapter 9 in Andreo et al. 2017). D
E
air,ch è êë

(5.57)
* The electrons (and positrons) set in motion by photons (through pair
production, Compton scattering and the photoelectric effect) are con- Since the absorbed dose to air predicted by Spencer–Attix
ventionally denoted as ‘secondary’, but here we wish to distinguish cavity theory, (Dair )SA , multiplied by the stopping-power
these photon-generated electrons from the delta-rays they in turn set ratio medium to air gives the dose to the medium, we can
in motion.
† The case of sub-equilibrium, small-field photon beams (see Section
derive (pQ)ch from equation 5.56 as:
5.8) demonstrates that the condition regarding the secondary elec-
tron ranges in air is not sufficient (see also Kumar et al. 2015b); CPE (Dair )SA
( pQ )ch = (5.58)
is also required (see Sections 5.5 and 5.7.4). (Dair )ch
Chapter 5: Principles and Basic Concepts in Radiation Dosimetry 77

Taking in Equation 5.57 the medium ‘med’ as ‘air’, we The topic of perturbation correction factors for (air-filled)
arrive at the following expression for the chamber perturba- ionisation chambers in radiotherapy has received a great deal
tion factor (Nahum 1996; Bouchard 2012): of attention due to the high demands placed on dosimetric
accuracy; Section 19.4.2 deals with it in more detail.
{ }
E max

ò (F ) (L D (E ) r)air dE + éëF Etot ( D ) ùû éS el ( D ) r ù D


tot
med.z ë û air
E
(pQ )ch = D med,z 5.7.7 General or ‘Burlin’ Cavity Theory
ò (F )
E max
ì ü
(L D (E ) r)air dE + í éF Etot ( D ) ù
úû air,ch ë el ( ) û air ýþ
tot
éS D r ù D Up to this point, two ‘extreme’ cases have been covered:
î êë
E
D air,ch
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

(5.59) 1. Detectors that are small compared with the elec-


tron ranges and that do not disturb the electron flu-
The electron fluence in the numerator is that in the medium ence present in the medium (Bragg–Gray cavities;
at the depth of interest, z; in the denominator it is the elec- Section 5.7.4)
tron fluence averaged over the air cavity of the ionisation 2. Detectors that are large compared with the second-
chamber. In principle, these two spectra can have different ary electron ranges and in which, therefore, par-
shapes – see Figure 5.35b. tial CPE is established (‘large photon detectors’;
Section 5.7.3).
5.7.6.2 The Various Sources of Ion-
Chamber Perturbation Many situations involve measuring the dose from photon (or
neutron) radiation using detectors that fall into neither of
As indicated in Figure 5.25 (see also Figures 16.6 and 16.7), these categories (Figure 5.26). In such cases, there is no exact
practical designs of ionisation chambers inevitably give rise expression for the dose ratio, Dmed/Ddet. ‘General’ cavity theory
to a number of different sources of perturbation, expressed was developed by Burlin (1966); the detector-to-medium dose
through various factors: ratio was approximated by a weighted mean of the stopping-
power ratio and the mass–energy absorption coefficient ratio:
pdis accounts for the displacement* of a certain volume of
det
medium by the air of the cavity; alternatively, this effect Ddet æ men ö
can be handled by assigning the chamber signal to an = ( f Q )det
med = wBG s det, med,D + (1 - wBG ) ç
ç r ÷÷ (5.61)
D med è ø med
effective point of measurement, Peff, displaced with
respect to the centre of the cavity. This effect applies to
where ωBG is a weighting factor that varies between unity for
high-energy photon and electron beams.
small (or Bragg–Gray) cavities and zero for large cavities (or
pwall accounts for the fact that the nature and composition of
photon detectors).
the chamber wall, which is responsible for a significant
Burlin estimated ωBG from the exponential attenuation of
proportion of the secondary electrons contributing to
the electron fluence entering the cavity through the wall,
the chamber signal, is different from that of the medium
balanced by the exponential build-up of the cavity-generated
(generally water). In the case of electron beams, this
electron fluence:
effect is generally assumed to be negligible (but see
Nahum 1996 for a theoretical analysis that challenged 

ò F e dx = 1 - e
eqm - me x
this). Buckley and Rogers (2006a, 2006b) found dif- F med med - me 
ferences up to 1.7% for plane-parallel chambers and wBG = eqm
= 0
(5.62)
F med 
m 
ò F dx
eqm e
0.6% for thimble chambers at the reference depth (with med
greater differences at other depths). 0

pcel accounts for the lack of air equivalence of the central where
electrode of cylindrical chambers. eqm
F med is the equilibrium electron fluence present in the
pcav accounts for the perturbation of the electron fluence in medium
the air of the cavity due to in-scattering; this effect is µe is the attenuation coefficient of the secondary elec-
important for electron beams but is negligible for pho- trons in the detector material
ton beams, as Harder (1974) demonstrated by reference  is the mean chord length in the detector.
to the Fano theorem (see Section 5.7.2).
The overall or ‘global’ perturbation correction factor, pQ, When applying his theory to air-filled cavities, Burlin used
is given approximately by the product of these perturbation the following expression, due to Loevinger (1956), for the
factors: attenuation of electrons from β-emitting radionuclides:

( pQ )ch = pdis pwall pcel pcav (5.60) 16 r


me = (5.63)
( E max - 0.036 )
1.4

where
* The word ‘displacement’ is somewhat ambiguous. It refers both to a
replacement of the medium (i.e. water) with air and to the shift of the r is the density (of air)
effective point of measurement (EPOM) according to the shape of the Emax is the maximum initial energy (in MeV) of the elec-
cavity and the direction of the ionising particles. trons in the β-ray spectrum (see Section 2.2.4).
78 Part A: Fundamentals

5.7.8 The Response of a Detector as it Varies


from ‘Bragg–Gray’ to ‘Large Photon’
Figure 5.27 shows how the response of a (fictitious) silicon
detector, D water/DSi, in water irradiated by a 6 MV x-ray
beam, computed by the Monte-Carlo usercode CAVnrc,
would vary as the detector density increased from that
FIGURE 5.26 A schematic illustration of an ‘intermediate’ or ‘Burlin’ of a gas (r = 0.005 g cm−3) past its real physical density
detector in a photon-irradiated medium; the secondary electron ranges (rSi = 2.33 g cm−3) to r = 10 g cm−3 (Kumar et al. 2016). The
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

are neither exclusively much greater (Bragg–Gray detector) nor much silicon detector was modelled as a cylinder of (constant)
smaller (large photon detector) than the detector dimensions. (Adapted diameter 2.3 mm and thickness 2 mm situated at 5.0 cm
from Chapter 9 in Andreo, P., Burns, D. T., Nahum, E. E., Seuntjens, depth in a large water phantom. The electron and photon
J. and Attix, F. H., Fundamentals of Ionizing Radiation Dosimetry, fluence spectra required to evaluate the Spencer–Attix,
Wiley-VCH, Weinheim, 2017.)
Bragg–Gray and Burlin dose ratios were computed with
usercode FLURZnrc. For each value of the detector density
Burlin and Chan (1969) suggested that e - met max = 0.01 a value of the Spencer–Attix Δ was estimated which satisfied:
yielded a better estimation of µe, where t max is the maximum
thickness of electron penetration, beyond which only 1% of éëRcsda (D)ùû Si
electrons reach. Janssens et al. (1974) found that 0.04 rather =
( Rcsda /R50 )
than 0.01 yielded improved agreement with experiment.
A more detailed treatment of the Burlin cavity is given in where
Chapter 9 in Andreo et al. (2017), including a brief discus-  is the mean chord length in the detector.
sion of the more complicated, but not necessarily more accu- Rcsda is the CSDA range in the detector material
rate, theories of Horowitz and Dubi (1982), Janssens (1983) R50 is the depth at which the dose is 50% of the maxi-
and Kearsley (1984). Andreo et al. concluded that analytical mum dose for a broad parallel electron beam inci-
theories would never be able to compete in accuracy with dent on the detector material.
Monte-Carlo simulation (Mobit et al. 1997). However,
Figure 5.27 in the next section demonstrates that the dose The ratio Rcsda/R50 was used to correct for the vast num-
ratio (Equation 5.61) given by Burlin theory can agree quite ber of changes in direction when the electrons slow down. It
well with the Monte-Carlo-derived value. was found to be around 1.95, almost independent of energy
Burlin cavity theory has recently been employed to esti- below 1 MeV.
mate dose deposition in cellular (micrometre-sized) targets At the lowest density, 0.005 g cm−3, i.e. a gaseous detector,
irradiated by kilovoltage photons (Oliver and Thomson Δ = 15 keV, the Monte-Carlo and Spencer–Attix dose ratios
2017). BG
are in excellent agreement, while the Bragg–Gray ratio, s w,Si ,

FIGURE 5.27 Left-hand axis: i) the variation of the Monte-Carlo dose ratio, (D w/DSi)MC (..□ ..), ii) the Spencer–Attix stopping power ratio,
water-to-silicon, sw,Si,Δ (--△--) and iii) the Burlin cavity ratio (--▽--), with the pseudo-density of a silicon cavity (0.23 cm diameter, 0.2 cm height)
located at 5 cm depth in a large water phantom irradiated by a 6 MV photon beam of radius 5.55 cm. Right-hand axis: the fraction of dose to the
silicon cavity due to direct photon interactions, (1 − ωBG), i) derived from Monte-Carlo (-.-o-.-) and ii) from the Burlin–Janssens formula (..△..).
BG
An arrow indicates the (density-independent) value of the unrestricted stopping-power ratio, s w,Si . A second arrow indicates the value of the ‘large

( ) ´(p )
w
ph
photon cavity’ dose ratio, corrected for photon fluence perturbation, men / r w,Si . (Adapted from Kumar, S., Deshpande, D. D. and
Si r =10 g cm -3
Nahum, A. E., Phys. Med. Biol., 61 (7), 2680–2704, 2016.)
Chapter 5: Principles and Basic Concepts in Radiation Dosimetry 79

is ≈2.5% lower. As the detector density is increased, the Field sizes at which CPE is violated can serve as a definition of
ranges of the secondary electrons in the silicon decrease, and ‘small fields’ (see Section 19.5.2.3 and especially Table 19.3).
the response of the detector moves into the ‘intermediate’ This lack of electronic equilibrium, together with the ‘geo-
or Burlin region. At a density of 0.5 g cm−3 the Burlin-theory metrical’ phenomenon of ‘source occlusion’ due to the finite
factor (1 − ωBG) – see Equation 5.62 – has risen to ≈0.1; thus size of the effective x-ray source, requires special consideration
≈10% of the detector dose comes from photon interactions in regarding firstly, the distribution of absorbed dose in a uni-
the detector material. At the (true) silicon density of 2.33 g cm−3 form medium such as water (Section 5.8.2) and secondly, the
this has increased to 40% and the Burlin ratio now lies signifi- response of detectors employed to determine the absorbed
cantly below the Spencer–Attix ratio. dose (McKerracher and Thwaites 1999; Sanchez-Doblado et
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

The ‘large photon detector’ ratio is approximately 1.114, al. 2007; Francescon et al. 2011; Scott et al. 2012; Bouchard et
which is ≈12% lower than the Spencer–Attix dose ratio at al. 2015; Kumar et al. 2015b; Benmakhlouf and Andreo 2017).
a typical gas density. The Monte-Carlo dose ratio begins
to approach 1.114 at the highest silicon pseudo-density,
5.8.2 Absorbed Dose in a Uniform Medium
10 g cm−3, where the Burlin-theory weighting factor (1 − ωBG)
has reached ≈0.8, meaning that 80% of the detector dose is
Irradiated by a Small Photon Field
due to photon interactions in the detector material. Megavoltage photon beams are created by electrons acceler-
ated in the section of a linear accelerator, reaching an energy of
several MeV and ‘hitting’ a thick tungsten target (see Section
11.2.12). The photon beam is the result of the bremsstrahlung
5.8 The Special Case of Small, Non-
interaction process (see Section 3.4). The effective source of the
equilibrium, Megavoltage Photon Fields bremsstrahlung radiation has a finite extent. Figure 5.28 illus-
5.8.1 Introduction trates how this extended source is partially blocked or occluded
when the opening of the collimator jaws is very narrow.
As a result of the treatment of small target volumes with corre- Scott et al. (2009) found that a 0.7 mm full width at half
spondingly small field sizes (see Chapter 40), narrow or ‘small’ maximum (FWHM) electron ‘pencil’ produced the best match
beams (often known as ‘small fields’) are being increasingly with measured output factors at a field width of 0.5 cm for a 15
employed and are deserving more attention (e.g. Alfonso et MV Varian 2100C linear accelerator (linac). This translated into
al. 2008; Das et al. 2008; IPEM 2010; Bouchard et al. 2015; a 2.5 mm ‘spot width’ of the photon fluence at the exit surface
ICRU 2017; IAEA 2017). In such ‘small’ megavoltage photon of the target; the extra broadening is due to the strong electron
fields, the distance from the beam central axis to the field edge is scattering in the high-Z tungsten target (see Section 3.6).
often less than the maximum lateral excursions of the secondary The effect of ‘source occlusion’ is demonstrated explicitly
electrons, and therefore CPE is never established at any depth. by the variation of water kerma as a function of the field

FIGURE 5.28 Schematic illustration of the effect of decreasing the field size to the point where the collimators obscure the ‘view’ of the extended
source (taking the FWHM of a 3D Gaussian distribution) even on the central axis. (Reproduced from Aspradakis, M. M., Byrne, J. P., Palmans,
H., Duane, S., Conway, J., Warrington, A. P. et al. IPEM Report 103. York: IPEM, 2010 with permission.)
80 Part A: Fundamentals
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

FIGURE 5.30 Schematic illustration of loss of charged particle equi-


FIGURE 5.29 Monte-Carlo derived output factors (OFs) for a 6 MV librium in a narrow field. Electrons labelled ‘𝟣’ leave the volume defined
photon beam with full linac-head geometry and point-source geometry by the edges of the narrow beam (the dashed divergent lines) but are
in terms of kerma ‘OF(kerma)’, and absorbed dose ‘OF(dose)’, on the not replaced by electrons labelled ‘2’ as it would be if the field size were
beam central axis at 10 cm depth in a cylindrical water phantom, versus larger (e.g. extending to the solid divergent lines). (Reproduced from
field size defined at 100 cm source–surface distance. The side lengths Gagliardi, G., Nahum, A. E. and Muren, L., ESTRO Newsletter Physics
of the square fields are 0.25 cm, 0.5 cm, 0.75 cm, 1 cm, 1.5 cm, 2 cm Corner, 80, 2011. With permission.)
and 3 cm, respectively. The error bars (masked by the symbols) are ±2
standard deviations (k=2) and correspond to statistical (Type A) uncer-
tainties. (Reproduced from Kumar, S., Fenwick, J. D., Underwood, T.
5.8.3 Detector Response in Small Photon Fields
S., Deshpande, D. D., Scott, A. J. and Nahum, A. E., Phys. Med. Biol., 5.8.3.1 General
60 (20), 8187–8212, 2015. With permission).
For a non-small megavoltage photon beam of a given quality,
size (‘output factor’ OF) corresponding to ‘full linac geom- dosimeter response, expressed as Dmed/Ddet, exhibits a negli-
etry’ (i.e. extended focal spot) of Figure 5.29 – there is a gible dependence on field size (Ding and Ding 2012; Taylor
dramatic decrease below a field size of ≈0.7 cm. In contrast, et al. 2012). Figure 5.31 shows the variation of (uncorrected)
the ‘OF(kerma) – point source’ curve decreases gradually detector response as the field size becomes ‘small’. The quan-
with decreasing field size. The reader is reminded that kerma tity on the y-axis is the response M f clin of the detector for a
relates directly to photon (energy) fluence – secondary elec- small ‘clinical’ field size f clin , decreasing from 3.0 cm × 3.0 cm
trons are not involved. to 0.5 cm × 0.5 cm, normalised to a (wide) ‘reference’ field
Figure 5.30 explains schematically how there cannot be f ref (typically 10 cm × 10 cm).†
(partial) CPE (see Section 5.5) on the beam axis if the field The figure shows that all the ionisation chambers ‘under-
width is insufficient to encompass the lateral displacement of read’; i.e. their response relative to a wide field decreases
secondary electrons. Electrons labelled ‘𝟣’ leave the volume faster than the absorbed dose to the medium (the ‘Monte-
defined by the edges of the narrow beam (the dashed diver- Carlo’ curve), and the degree of ‘under-reading’ increases as
gent lines) but are not replaced by electrons labelled ‘2’, as the the field size decreases. Note in particular that at the smallest
source of these virtual electrons lies outside the edges of the field size, 0.5 cm, the reading of the commonly employed
narrow beam, and therefore these electron tracks are never Farmer chamber appears to be a factor ~5 below the ‘Monte-
created. The effect of this loss of CPE* is clearly demon- Carlo’ curve.
strated by the two curves labelled ‘OF(dose)’ in Figure 5.29. The dose in water at the measurement point, Dw-point, in a
The ‘full linac geometry’ curve shows the combined effect of field fclin of quality Q clin, can be expressed as:
loss of CPE and source occlusion.
It follows that the absorbed dose on the central axis of a
(very) narrow megavoltage photon beam from a linear accel- (D w-point )Qfclin = M Qfclin N Qfrefref
((D f clin
)
w-point Q clin /M Qfclin
clin ) (5.64)
erator falls off rapidly with field size due to the combined
clin clin
((D f ref f ref
w-point )Q ref /M Q ref )
effect of source occlusion and lack of CPE, as demonstrated where
by the ‘Monte-Carlo’ curve of Figure 5.31. The challenge M is the reading of the detector, with the centre of its
is to determine accurately this strong decrease of absorbed sensitive volume at the measurement point
dose at sub-equilibrium field sizes. N  is the absorbed-dose-to-water calibration coefficient
for this detector in machine-specific reference field fmsr
of quality Q msr (see Section 19.5.2).

* This loss of CPE is sometimes described as lateral electronic disequi-


librium, although strictly either CPE exists or it does not, i.e. it is not † The subscript Q clin refers to the ‘clinical’ beam quality, which can be
a ‘directional’ concept. different from the reference quality Q.
Chapter 5: Principles and Basic Concepts in Radiation Dosimetry 81
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

FIGURE 5.31 Measured detector response at field size, fclin, normalised to that at the reference field size, as a function of fclin for three different
ionisation chamber types, for a diamond detector and for a diode in a 6 MV beam; the Monte-Carlo curve is assumed to represent the true ratio
of doses. (Adapted from Sanchez-Doblado, F., Hartmann, G. H., Pena, J., Rosello, J. V., Russiello, G. and Gonzalez-Castano, D. M., Phys. Med.,
23 (2), 58–66, 2007. by S. Duane [private communication].)

The final term is a field-specific correction factor, often written response in a small field;* the separate factors are defined
as k, accounting for the difference in response between the such that they act sequentially.
clinical and reference fields (see Section 19.5.3) It should be noted that the changes in the stopping-power
The variation of the ratio D/M as the field size decreases ratio for water to air, silicon or diamond, between field sizes
is usually obtained from Monte-Carlo simulations (e.g. of 4 cm × 4 cm and 0.5 cm × 0.5 cm, are in the region of 0.2%
Francescon et al. 2011 – see later; see also Section 19.5.4). to 0.6% (Ding and Ding 2012; Czarnecki and Zink 2013).
Bouchard et al. (2009) have given the following expression The stopping-power ratio in Equation 5.65 could imply
for the ratio of dose at a point in water to the mean dose in that the secondary electron fluence (spectrum) in the detec-
the sensitive volume of the detector: tor is negligibly different from that in the uniform medium;
i.e. the detector behaves substantially as a Bragg–Gray cav-
ity. However, this ratio can in principle be evaluated so that

{ }
f

{D } it takes full account of differences in the electron spectra


f w
w-point / Ddet = PZ Pr Pvol éëL D /r ùû (5.65)
Qf det
Qf
between the medium (water) and the detector.

5.8.3.2 Response as a Function of Detector Type


This expression, derived from Equation 5.56, contains the
following perturbation correction factors, which are needed Francescon et al. (2011) computed the response of a range
because the electron fluence in the detector differs from that of detector types (‘micro’ ionisation chambers, diodes and
at the point in water, i.e. deviates from perfect Bragg–Gray a liquid ionisation chamber) as a function of field sizes from
behaviour (Fenwick et al. 2013): 3.0 cm down to 0.5 cm using the Monte-Carlo egs_cham-
ber usercode. Figure 5.33 shows their results expressed in
PZ accounts for the effect of detector material of atomic terms of the correction factor, i.e. the change in D/M with
number Z (denoted as Pfl in Bouchard et al. (2009) field size (cf. Equation 5.64). Some clear trends can be seen
and in Fenwick et al. (2013); in the figure – the diodes ‘over-read’, therefore requiring
Pr accounts for the effect of detector material of density r;
P vol accounts for the averaging of the fluence across the * It should be noted that Figure 5.32 is cast in the framework of Bragg–
sensitive volume of the detector. Gray cavity theory. However, non-gaseous detectors do not generally
fulfill B–G conditions (see Section 5.7.4.3) even in large fields; in small
fields neither non-gaseous nor gaseous detectors do so. Consequently,
Figure 5.32 is a schematic description of the different sources small-field perturbation factors may correct for relatively large perturba-
of perturbation that contribute to the overall detector tions of the electron fluence (spectrum) existing in the uniform medium.
82 Part A: Fundamentals
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

FIGURE 5.32 Factors for the sequential conversion of mean dose in the sensitive volume of the detector, Ddet , to dose at a point in water, D w-point,
via the electron fluences, Φ, in the sensitive volume and intermediate structures. (Reproduced from Fenwick, J. D., Georgiou, G., Rowbottom, C.
G., Underwood, T. S. A., Kumar, S. and Nahum, A. E., Phys. Med. Biol., 63 (12), 125003, 2018. With permission.)

corrections factors less than unity, but the ‘under-reading’


of the ionisation chambers is clearly greater; the PinPoint
chamber requires a correction factor of ≈1.13 at the smallest
field size. Referring to Equation 5.60, the value of ( pQ )ch
must be close to 1.13. A departure from unity of this magni-
tude is far larger than that of the ion-chamber perturbation
correction factors encountered in non-small megavoltage
photon beams (see Sections 5.7.6 and 19.4.2).
Scott et al. (2012) carried out Monte-Carlo computations
of the response of ionisation chambers and of diamond and
diode solid-state detectors in a 15 MV Varian linac beam
(modelled as in Scott et al. 2008) as a function of field
size, from 10 cm × 10 cm (reference) all the way down to
0.25 cm × 0.25 cm. Their results were expressed in terms
of Fdetector = D water/Ddetector evaluated at 5 cm depth in water.
Figure 5.34 from Scott et al. (2012) confirmed the trend
with detector type which was found by Francescon et al.
(2011): strong under-reading by a PinPoint-chamber-like air
volume, moderate over-reading by the diode and the dia-
mond detector, as shown by the light-grey data points in FIGURE 5.33 This figure represents the correction factor
Figure 5.34 and curves in Figure 5.33. ((D w-point )Qfclin
clin ) (
/M Qfclin
clin )
/ (D w-point )Qf1010××1010 /M Qf1010××1010 for five detectors as
In an attempt to shed more light on the role played by the a function of the field size, for the 6 MV beams of Siemens (dotted
various detector characteristics (principally atomic number curve) and Elekta (full curve) linear accelerators. (Reproduced from
and density relative to those of water) in the variation of Francescon, P., Cora, S. and Satariano, N., Med. Phys., 38 (12), 6513–
6527, 2011. With permission.)
overall detector response with field size, Scott et al. (2012)
also simulated the response of detectors made of ‘water’
but with the physical densities of air, diamond and silicon, the PinPoint air cavity, this is sw,air,Δ ≈ 1.10 (cf. Figure 5.21),
respectively.* These variable-density-water-detector results and for the diamond and diode detectors, the values (≈1.15)
are shown as the black curves and data points in Figure 5.34. correspond to D water/Ddetector predicted by Burlin theory (see
Several features should be noted. At the largest (reference) Section 5.7.7), as neither detector acts in a 100% Bragg–
field size, the ‘water-detector’ curves all have an Fdetector of Gray manner.
unity, as expected. By contrast, the reference-field-size, real- By a careful examination of Figure 5.34, one can see that
detector values (light-grey) are given by the dose ratios, for each detector type, the water-detector (black) curve
D water/Ddetector, predicted by cavity theory (Section 5.7); for is displaced vertically (i.e. along the Fdetector axis) from the
‘true-detector’ (light-grey) curve by an amount that is almost
totally independent of the field size; i.e. the black/light-grey
* The (PEGS4) data files for the Monte-Carlo simulations of the arti-
ficial (water) substances corresponded to the densities of silicon,
pairs of curves remain parallel to each other. Expressed
diamond and air, but their mass stopping powers and mass energy another way, the relative change in detector response with
absorption coefficients were set equal to those of unit-density water. field size is virtually unaffected by replacing the true atomic
Chapter 5: Principles and Basic Concepts in Radiation Dosimetry 83
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

FIGURE 5.34 Monte-Carlo calculated dose-to-water to dose-to-detector-in-water ratios, Fdetector, obtained for diamond, diode and PinPoint
detector voxels (light grey data points and curves) and the detector-voxels filled with detector-density water (data points and curves in black). All
detectors were positioned on-axis at 5 cm depth in a water phantom and the results displayed with k=2 error bars, reflecting the statistical uncer-
tainties of the MC calculations. The field sizes were 0.25, 0.5, 0.75, 1.0, 1.5, 2.0, 3.0 and 10.0 cm. (Reproduced from Scott, A. J., Kumar, S.,
Nahum, A. E. and Fenwick, J. D., Phys. Med. Biol., 57 (14), 4461–4476, 2012. With permission.)

composition of each detector by water while keeping its physi- b show the dramatic difference in electron-fluence spectra
cal density equal to the true value. between a ‘PinPoint’ air cavity and a small water voxel at the
The findings illustrated in Figure 5.34 suggest very 0.25 cm × 0.25 cm field size in the 6 MV and 15 MV beams,
strongly, at least for a 15 MV x-ray beam, that the physi- respectively. As discussed in Section 5.7.4, small air cavities
cal density of the detector material is the main driver of the in megavoltage photon beams might be expected to behave
change in its response with field size, with the atomic com- in a Bragg–Gray manner irrespective of the field size, as the
position playing at most a very minor role. However, the dose contribution to the detector signal due to photon interac-
ratio at non-small field sizes is driven by medium-to-detector tions in the air cavity is negligible, and the ranges (in air)
differences in atomic composition (Andreo and Benmakhlouf of the photon-liberated electrons are much greater than the
2017) consistent with cavity theory. In a subsequent Monte- cavity dimensions.* Yet Figure 5.35 clearly indicates major
Carlo-based analysis of the response of a PTW 60017 diode differences between electron-fluence spectra in the air and
detector at both 6 MV and 15 MV, Fenwick et al. (2018) water volumes: Bragg–Gray behaviour can be said to have
confirmed the pivotal role of detector density in small-field ‘broken down’.
dosimeter response. They also showed that the electron flu- The electron-fluence spectra of Figure 5.35a can be com-
ence in the detector’s sensitive volume was perturbed more pared with the ‘perfect’ B–G behaviour of an air volume (of
by differences in atomic composition than by differences the same dimensions), but now in a 3 cm × 3 cm field, shown
in density; however, this Z-dependent perturbation was in Figure 5.19, i.e. negligible perturbation. Furthermore, the
independent of field size, in marked contrast to the density- 15 MV electron-fluence spectra in Figure 5.35b indicate that
dependent perturbation. the shape of the electron-fluence spectra in the air cavity dif-
Values of field-specific correction factors for a wide range fers significantly from that in water; i.e. the two distributions
of commercially available detectors can be found in Figures cannot be brought into coincidence by applying a single,
19.18 to 19.23; these show essentially the same trends as energy-independent perturbation correction factor.
Figures 5.33 and 5.34 – values below unity for solid-state The Monte-Carlo computations of Kumar et al. (2015b)
detectors and above unity for (air) ionisation chambers. yielded perturbation factors of 1.323 and 2.139 (=Φwater/Φair
integrated over energy) for the 6 MV and 15 MV beams,
respectively. Table 5.3 compares, for a PinPoint air cav-
5.8.3.3 Small-Field Perturbation Factor for ity, the EGSnrc-based Kumar et al. results with those of
Air-filled Ionisation Chambers Benmakhlouf et al. (2014) obtained with the PENELOPE
Monte-Carlo code (see Chapter 30).
Kumar et al. (2015b) employed the EGSnrc Monte-Carlo
code with the aim of gaining insight into the large devia-
* The very short ranges of the lowest-energy delta-rays are accounted
tions from Bragg–Gray behaviour exhibited by air-filled for implicitly by employing the Spencer–Attix stopping-power ratio –
ionisation chambers in very small fields. Figures 5.35a and see Section 5.7.5.
84 Part A: Fundamentals

TABLE 5.3
'Total' electron fluence differential in energy
per incident photon fluence (cm-2 MeV-1)
1.0x10-7 '0.25 x 0.25 cm2' 6 MV Comparison of Monte-Carlo-Derived Dose Ratios, Water-to-
Water voxel
Air, as a Function of Field Size for the PinPoint 3D Air Cavity
8.0x10-8 Pinpoint air cavity × 1.323 with Monte-Carlo-Calculated Output Correction Factor
Pinpoint air cavity
kQf clin , f msr
clin , Q msr
from Benmakhlouf et al. (2014).
6.0x10-8 é(Dw /Ddet )MC ù
kQfclin , fmsr ë û Qclin
clin , Qmsr
é(Dw /Ddet )MC ù
Beam Field Size (Benmakhlouf ë û Qmsr
4.0x10-8 Quality (cm × cm) et al. 2014) (Kumar et al. 2015b)
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

6 MV, FLG 0.5 × 0.5 1.147 1.144 ± 0.006


2.0x10-8
1×1 1.010 1.024 ± 0.007
2×2 1.000 1.000 ± 0.005
0.0
0.001 0.01 0.1 1 6 The uncertainties are ±2s (k = 2).
Energy (MeV) FLG: full linac geometry.
(a) The field size was defined at 100 cm source surface distance.
(Adapted from Kumar, S, Fenwick, J. D., Underwood, T. S. A.,
Deshpande, D. D., Scott, J. S. and Nahum, A. E., Phys. Med. Biol., 60
(20), 8187–8212, 2015. With permission.)
1.4x10-6
'Total' electron fluence differential in energy

'0.25 x 0.25 cm2' 15 MV


per incident photon fluence (cm-2 MeV-1)

1.2x10-6 The absorbed dose in either the medium or the detector can
Water voxel be approximated by the product of the electron fluence, F e,
1.0x10-6
(at ‘×’ or averaged over the detector) and the electronic stop-
Pinpoint air cavity × 2.139
Pinpoint air cavity
8.0x10-7 ping power, S el /r , averaged over the fluence (see e.g. Kumar
et al. 2015b):
6.0x10-7

4.0x10-7 { int
D w-point » éëF e (´)ùû ´ éëS el /r ùû
w w
int
}
{ }
-7
2.0x10 ext ext
+ éëF e (´)ùû ´ éëS el /r ùû (5.67)
w w
0.0
0.001 0.01 0.1 1 10 int
= éëD (´)ùû w + éëD (´)ùû w
ext
Energy (MeV)
(b)
and:

{ }
FIGURE 5.35 Monte-Carlo-derived electron-fluence spectra (all int int
generations of secondary electrons) scored in a ‘point-like’ water Ddet » éëF e (cav)ùû ´ éëS el /r ùû
det det
‘voxel’ (0.5 mm diameter, 0.5 mm thickness) and in a ‘PinPoint’ air
cavity (2.3 mm diameter, 2.6 mm length) both at 5 cm depth, in a
0.25 cm × 0.25 cm field, on the beam central axis in a large water phan-
tom irradiated by (a) 6 MV and (b) 15 MV photon beams. (Reproduced
{ ext
+ éëF e (cav)ùû ´ éëS el /r ùû
det det
ext
} (5.68)

int ext
from Kumar, S., Fenwick, J. D., Underwood, T. S., Deshpande, D. D., = éëD (cav)ùû det + éëD (cav)ùû det
Scott, A. J. and Nahum, A. E., Phys. Med. Biol., 60 (20), 8187–8212,
2015. With permission.)
where
5.8.3.4 The Physics of ‘Bragg–Gray ‘int’ signifies that this component of the electron fluence
Breakdown’ in Small Fields is generated internally, i.e. by photon interactions in
the equivalent water volume (w) or detector (det),
The signal generated by an arbitrary detector in water (for
respectively
simplicity) irradiated by a photon field can be analysed in
‘ext’ signifies that the electron fluence is generated exter-
terms of the electron fluence present in the detector, irrespec-
nally, i.e. by photon interactions in the water volume
tive of the origin(s) of this fluence. Referring to Figure 5.36,
external to the equivalent water volume (w) or detec-
the field on the left is denoted by fqCPE, indicating that this
tor (det), respectively.
field is wide enough for quasi-CPE to be present at ‘×’; the
sub-equilibrium field on the right is denoted by fsmall. Combining Equations 5.66 through 5.68 yields:
Taking Equation 5.64 as our starting point, and simplify-
ing the notation, we can write: éD wint + D wext ù é int ext
ù
kqCPE = ë int
small û small / ëDdet + Ddet û small
(5.69)
( ) éD w + D w ù é ù
ext int ext

small
D w-point / Ddet ë û qCPE / ëD det + D det û qCPE
kqCPE = small
(5.66)
(D w-point / Ddet )
qCPE
where, for simplicity, the bar over Ddet has been omitted and
‘w-point’ has been replaced by ‘w’.
Chapter 5: Principles and Basic Concepts in Radiation Dosimetry 85

(i) Uniform medium


fqCPE fsmall
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

(ii) Detector in medium


fqCPE fsmall

FIGURE 5.36 Schematic illustration of the response of a detector of arbitrary physical density in a narrow, sub-equilibrium megavoltage photon
e−
field; the measurement position, marked by ‘×’, lies at a depth greater than that of Dmax, and rmax is the maximum lateral distance any electron can
travel. (Adapted from Kumar, S., Fenwick, J. D., Underwood, T. S., Deshpande, D. D., Scott, A. J. and Nahum, A. E., Phys. Med. Biol., 60 (20),
8187–8212, 2015.)

• Case I – detector with density lower than water, e.g. Figure 5.36. First, the contribution from the external elec-
air ionisation chamber trons to the electron fluence in the air cavity and in the
water volume will be substantially equal. However, the dose
int
For such a detector, it can be assumed that Ddet ext
 Ddet – this at × in the water volume is augmented by the ‘internal’
is usually sufficient for Bragg–Gray behaviour; the denomi- electron fluence generated by photon interactions in this
nator, the dose ratio in a non-small beam, is the conventional water volume. In the air volume, this ‘internal’ component
water-to-detector stopping-power ratio. Consequently: is negligible.

éD wint + D wext ù é ext ù


û small / ëDdet û small • Case II – detector with density higher than water e.g.
(ksmall
)
qCPE low-density ȑ
éD wtot ù é tot ù diamond or diode detector
ë û qCPE / ëDdet û qCPE
(5.70)
For a detector with a density greater than that of the medium
éD wint + D wext ù é ext ù
»ë û small / ëDdet û small (water), we must start from the full Equation 5.69, but we
éës w ,det ùû qCPE can replace the denominator, for the qCPE field, by the cav-
ity-theory ratio, which we will write as f w,det to emphasise that
It is clear that this quantity is greater than unity, consistent this is not necessarily a Bragg–Gray detector:
with Figures 5.33 and 5.34. Therefore, there is now a signifi-
cant imbalance between the electron fluence in the undisturbed
éD wint + D wext ù é int ext
ù
û small / ëDdet + Ddet û small
medium and that in the gas cavity – one can say that Bragg–
Gray conditions have broken down.* Consequently, an (air-
(ksmall
)
qCPE high-density =ë
éë f w,det ùû qCPE
filled) ionisation chamber in a small, sub-equilibrium field will
‘under-read’, requiring a correction factor greater than unity. (5.71)
The significant under-reading of an air cavity in a small
int
field can also be understood from the right-hand side of The Ddet and D wint components are negligibly influenced
by field size, as in Case I. Therefore, it is the change in the
* The conversion of electron-fluence to absorbed dose is naturally taken Dext component that drives the variation in k with field size.
care of by cavity integrals (e.g. Equations 5.65 and 5.66) involving the Given that the detector now has a higher density than the
mass electronic stopping power of the detector material, in which the ext
water it replaces, Ddet must constitute a smaller fraction of
natural logarithm of the I-value (I increases with Z) plays a key role.
86 Part A: Fundamentals

the total dose than D wext i.e. the detector is more ‘Burlin-like’ 5.8.4 Mass-density Compensation to Improve
than the water it replaces. As the field size becomes smaller, Small-field Detector Response
D wint + D wext will therefore decrease faster than Ddet
int ext
+ Ddet , and It was shown in Section 5.8.3.2 that the physical density of
consequently, k will drop below unity, as Figure 5.33 dem- the sensitive detector material, relative to that of the medium
onstrates. Thus detectors with density greater than water will (water), plays the dominant role in detector response, as
‘over-read’ in small, sub-equilibrium fields. the field is reduced below the minimum width required for
In the case of a 100% photon detector (i.e. small with a very quasi-CPE in the uniform medium. This finding suggests
high density), the ‘ext’ terms will, by definition, be zero (elec- that by surrounding a low-density detector with high-density
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

trons from the medium make a negligible contribution to the material (e.g. a dense ion-chamber wall), or alternatively by
detector signal – see Section 5.7.3), and the detector signal adding a low-density material to a high-density detector, it
(∝ Ddet ) will therefore, to first order, be independent of field might be possible to reduce or even eliminate the dependence
size. Consequently, the k-factor will be entirely determined by of response on (sub-equilibrium) field size (Underwood
the decrease in the ‘external’ electron fluence entering the equiv- et al. 2012). This mass-density compensation concept was
alent volume of the medium (water) as the field width decreases demonstrated in simulation-based studies (Charles et al.
from ‘equilibrium’ (or reference) to ‘small’. For a given detector 2013; Underwood et al. 2013) and subsequently validated
volume and (small) field size, this corresponds to the lower limit experimentally by Charles et al. (2014), who modified two
of k, just as the air ionisation chamber corresponds to an upper unshielded diode detectors.
limit for a detector of that volume. It is worth pointing out that Figure 5.37 shows the design of the DiodeAir detector
in the hypothetical case of a high-density detector with a vol- proposed by Underwood et al. (2015). Monte-Carlo simu-
ume so small that there is negligible build-up in it and therefore, lation of the response of the detector as a function of field
negligible build-up in the (unit density) medium it displaces, size together with an experimental confirmation are shown
Bragg–Gray behaviour would be ‘restored’, as the electron flu- in Figure 5.38; it can be seen that an air gap of 1 mm results
ence in the detector and in the displaced medium would be in almost perfect field-size independence of the ‘DiodeAir’
overwhelmingly due to ‘external’ secondary electrons. response.
The description of small-field detector response given
here is complemented by the comprehensive small-field cav-
ity-theory review by Bouchard et al. (2015). These authors
emphasised that the lack of CPE alone does not give rise to
significant detector correction factors (i.e. departures from
‘classical’ cavity theory): a perfectly water-equivalent detec-
tor never requires corrections whether CPE is present or not.
Rather, it is the combination of lack of CPE in the small field
itself and the difference between the density of the detector
FIGURE 5.37 Schematic representation of the DiodeAir detector
cavity and the density of the medium that gives rise to the large design, with an air-gap thickness t. Prototypes with t = 0.5 mm, 1.0 mm,
correction or perturbation factors. Bouchard et al. (2015) also 1.2 mm and 1.55 mm were studied. (Reproduced from Underwood, T.
emphasise the major role played by volume averaging in small S., Thompson, J., Bird, L., Scott, A. J., Patmore, P., Winter, H. C. et al.,
fields (see P vol in Section 5.8.3.1 and k vol in Section 19.6.2). Phys. Med. Biol., 60 (7), 2939–2953, 2015. With permission.)

FIGURE 5.38 For 6 MV beams: the impact of the modifications to the diode detector on the correction factor kQf clin , f10×10
clin , Q 10 ×10
for square fields defined
by the linac jaws. Error bars correspond to k = 1 statistical uncertainties. (a) Monte-Carlo simulations, (b) experiments using radiochromic EBT3
film (see Section 18.3). (Reproduced from Underwood, T. S., Thompson, J., Bird, L., Scott, A. J., Patmore, P., Winter, H. C. et al., Phys. Med.
Biol., 60 (7), 2939–2953, 2015. With permission.)
Part A References

Alfonso, R., Andreo, P., Capote, R., Huq, M. S., Kilby, W., Berger, M. J., Hubbell, J. H., Seltzer, S. M., Chang, J., Coursey, J. S.,
Kjäll, P. et al. A new formalism for reference dosimetry of Sukumar, R. et al. XCOM: Photon Cross Sections Database.
small and nonstandard fields. Med. Phys. 35 (11):5179– NIST Standard Reference Database 8 (XGAM) NIST, PML,
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

5186, 2008. doi:10.1118/1.3005481 Radiation Physics Division, Washington, DC, 2010.


Andreo, P. The interaction of electrons with matter: II. Bichsel, H. Charged-particle interactions. In Radiation
Scattering. In The Computation of Dose Distributions in Dosimetry. Vol. I, edited by F. H. Attix and W. C. Roesch,
Electron Beam Radiotherapy, edited by A. E. Nahum, pp. pp. 157–228. New York: Academic Press, 1968.
56–71. Madison, WI: Medical Physics Publishing, 1985. BIPM (Bureau International des Poids et Mesures). The
Andreo, P. Stopping-power ratios for dosimetry. In Monte- International System of Units (SI). 9th edition. Sèvres,
Carlo Transport of Electrons and Photons, edited by T. M. France: Organisation Intergouvernementale de la
Jenkins, W. R. Nelson, A. Rindi, A. E. Nahum and D. W. Convention du Mètre, 2019.
O. Rogers, pp. 485–501. New York: Plenum, 1988. Blanc, D. and Portal, G. Précis de Physique Nucléaire. 2nd edi-
Andreo, P. and Nahum, A. E. Stopping-power ratio for a photon tion. Paris: Dunod, 1999.
spectrum as a weighted sum of the values for monoener- Bohr, N. On the theory of the decrease of velocity of moving
getic photon beams. Phys. Med. Biol. 30 (10):1055–1065, electrified particles on passing through matter: Phil. Mag.
1985. doi:10.1088/1361-6560/aa562e 25 (1913). Phil. Mag. 25:10–30, 1913. https://archive.
Andreo, P. and Benmakhlouf, H. Role of the density, density org/details/londonedinburg6251913lond
effect and mean excitation energy in solid-state detectors Bohr, N. The penetration of atomic particles through mat-
for small photon fields. Phys. Med. Biol. 62 (4):1518–1532, ter, Mathematisk-fysiske Meddelelser XVIII, 8: Det Kgl.
2017. doi:10.1088/1361-6560/aa562e Danske Videnskabernes Selskab, 1948.
Andreo, P., Burns, D. T., Nahum, E. E., Seuntjens, J. and Attix, Borg, J., Kawrakow, I., Rogers, D. W. and Seuntjens, J. P. Monte
F. H. Fundamentals of Ionizing Radiation Dosimetry. Carlo study of correction factors for Spencer-Attix cavity
Weinheim: Wiley-VCH, 2017. theory at photon energies at or above 100 keV. Med. Phys.
Aspradakis, M. M., Byrne, J. P., Palmans, H., Duane, S., 27 (8):1804–1813, 2000. doi:10.1118/1.1287054
Conway, J., Warrington, A. P. et al. IPEM Report 103. Bortfeld, T. An analytical approximation of the Bragg curve for
Small Field MV Photon Dosimetry. York: IPEM, 2010. therapeutic proton beams. Med. Phys. 24 (12):2024–2033,
Attix, F. H. Basic gamma-ray dosimetry. Health Phys. 15 (1):49– 1997. doi:10.1118/1.598116
56, 1968. Bouchard, H. A theoretical re-examination of Spencer-Attix
Attix, F. H. The partition of kerma to account for bremsstrah- cavity theory. Phys. Med. Biol. 57 (11):3333–3358, 2012.
lung. Health Phys. 36 (3):347–354, 1979. www.ncbi.nlm. doi:10.1088/0031-9155/57/11/3333
nih.gov/pubmed/489286 Bouchard, H., Seuntjens, J., Carrier, J. F. and Kawrakow, I.
Attix, F. H. Introduction to Radiological Physics and Radiation Ionization chamber gradient effects in nonstandard beam
Dosimetry. New York: Wiley, 1986. configurations. Med. Phys. 36 (10):4654–4663, 2009.
Attix, F. H., De La Vergne, L. and Ritz, V. H. Cavity ionization doi:10.1118/1.3213518
as a function of wall material. J. Res. Natl. Bur. Std. 60 Bouchard, H., Seuntjens, J., Duane, S., Kamio, Y. and Palmans,
(3):235–243, 1958. doi:10.6028/jres.060.028 H. Detector dose response in megavoltage small photon
Benmakhlouf, H., Sempau, J. and Andreo, P. Output correction beams. I. Theoretical concepts. Med. Phys. 42 (10):6033–
factors for nine small field detectors in 6 MV radiation 6047, 2015. doi:10.1118/1.4930053
therapy photon beams: A PENELOPE Monte Carlo study. Bragg, W. H. Studies in Radioactivity. New York: Macmillan,
Med. Phys. 41 (4):041711, 2014. doi:10.1118/1.4868695 1912.
Benmakhlouf, H. and Andreo, P. Spectral distribution of par- Buckley, L. A., Kawrakow, I. and Rogers, D. W. An EGSnrc
ticle fluence in small field detectors and its implication on investigation of cavity theory for ion chambers mea-
small field dosimetry. Med. Phys. 44 (2):713–724, 2017. suring air kerma. Med. Phys. 30 (6):1211–1218, 2003.
doi:10.1002/mp.12042 doi:10.1118/1.1573891
Berger, M. J. and Seltzer, S. M. Tables of Energy Losses and Buckley, L. A. and Rogers, D. W. Wall correction factors, P wall,
Ranges of Electrons and Positrons. Washington, DC: for thimble ionization chambers. Med. Phys. 33 (2):455–
National Aeronautics and Space Administration, 1964. 464, 2006a. doi:10.1118/1.2161403
Berger, M. J. and Hubbell, J. H. XCOM: Photon Cross Sections Buckley, L. A. and Rogers, D. W. Wall correction factors, P wall,
on a Personal Computer 87-3597. Washington, DC: NBS, for parallel-plate ionization chambers. Med. Phys. 33
1987. (6):1788–1796, 2006b. doi:10.1118/1.2199988
Berger, M. J. and Wang, R. Multiple-scattering angular deflec- Burlin, T. E. An experimental examination of theories relat-
tions and energy-loss straggling. In Monte-Carlo Transport ing the absorption of gamma-ray energy in a medium
of Electrons and Photons, edited by T. M. Jenkins, W. R. to the ionization produced in a cavity. Phys. Med.
Nelson, A. Rindi, A. E. Nahum and D. W. O. Rogers, pp. Biol. 6:33–53, 1961. https://www.ncbi.nlm.nih.gov/
21–56. New York: Plenum, 1988. pubmed/13689113

87
88 Part A: Fundamentals

Burlin, T. E. The limits of validity of cavity ionization theory. Fenwick, J. D., Georgiou, G., Rowbottom, C. G., Underwood,
Br. J. Radiol. 35:343–348, 1962. doi:10.1259/0007- T. S. A., Kumar, S. and Nahum, A. E. Origins of the
1285-35-413-343 changing detector response in small megavoltage photon
Burlin, T. E. A general theory of cavity ionisa- radiation fields. Phys. Med. Biol. 63 (12):125003, 2018.
tion. Br. J. Radiol. 39 (466):727–734, 1966. doi:10.1088/1361-6560/aac478
doi:10.1259/0007-1285-39-466-727 Fermi, E. The ionization loss of energy in gases and in condensed
Burlin, T. E. Cavity-chamber theory. In Radiation Dosimetry. materials. Phys. Rev. 57:485–493, 1940. physics.princeton.
Vol. I, edited by F. H. Attix and W. C. Roesch, pp: 331- edu/~mcdonald/examples/EP/fermi_pr_57_485_40.pdf
392 New York: Academic Press, 1968. Fernández-Varea, J. M., Mayol, R., Salvat, F. and Baro, J. On
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

Burlin, T. E. and Chan, F. K. The effect of the wall on the Fricke the theory and simulation of multiple elastic scattering of
dosemeter. Int. J. Appl. Radiat. Isot. 20 (11):767–775, electrons. Nucl. Instrum. Methods Phys. Res., Sect. B 73
1969. doi:10.1016/0020-708X(69)90040-4 (4):447–473, 1993. doi:10.1016/0168-583X(93)95827-R
Bushberg, J. T., Seibert, J. A., Leidholt, E. M., Jr. and Boone, J. Francescon, P., Cora, S. and Satariano, N. Calculation of
M. The Essential Physics of Medical Imaging. 3rd edition. kQf clin , f msr
clin ,Q msr for several small detectors and for two linear
Baltimore: Lippincott, William and Wilkins, 2012. accelerators using Monte Carlo simulations. Med. Phys. 38
Charles, P. H., Crowe, S. B., Kairn, T., Knight, R. T., Hill, B., (12):6513–6527, 2011. doi:10.1118/1.3660770
Kenny, J. et al. Monte Carlo-based diode design for correc- Gambini, D. J. and Granier, R. Manuel Pratique de
tion-less small field dosimetry. Phys. Med. Biol. 58 (13):4501– Radioprotection, Technique et Documentation Collection.
4512, 2013. doi:10.1088/0031-9155/58/13/4501 Cachan, Paris: Lavoisier Editions, 1997.
Charles, P. H., Cranmer-Sargison, G., Thwaites, D. I., Crowe, Goodhead, D. T. Relationship of microdosimetric techniques
S. B., Kairn, T., Knight, R. T. et al. A practical and theo- to applications in biological systems. In The Dosimetry
retical definition of very small field size for radiotherapy of Ionising Radiation. Vol. 2, edited by K. R. Kase, B.
output factor measurements. Med. Phys. 41 (4):041707, E. Bjarngard and F. H. Attix, pp. 1–89. Orlando, FL:
2014. doi:10.1118/1.4868461 Academic Press, 1987.
Cherry, S. R., Sorenson, J. and Phelps, M. P. Physics in Nuclear Goodwin, P. N. and Rao, D. V. An Introduction to the Physics
Medicine. 4th edition. St Louis: W. B. Saunders-Elsevier, of Nuclear Medicine. Springfield, IL: Charles C. Thomas,
2012. 1977.
Chilton, A. B. A note on the fluence concept. Health Phys. Gottschalk, B., Koehler, A. M., Schneider, R. J., Sisterson, J.
34 (6):715–716, 1978. www.ncbi.nlm.nih.gov/ M. and Wagner, M. S. Multiple Coulomb scattering of
pubmed/730526 160 MeV protons. Nucl. Instrum. Methods Phys. Res.,
Christen, T. Radiometry. Arch. Roentgen Ray 19:201–219, Sect. B 74:467–490, 1993. doi:10.1016/0168-583X(93)
1914. 95944-Z
Czarnecki, D. and Zink, K. Monte Carlo calculated correc- Gray, L. H. The absorption of penetrating radiation. Proc. R.
tion factors for diodes and ion chambers in small pho- Soc. A 122:647–668, 1929. doi:10.1098/rspa.1929.0050
ton fields. Phys. Med. Biol. 58 (8):2431–2444, 2013. Gray, L. H. An ionisation method for the absolute measurement
doi:10.1088/0031-9155/58/8/2431 of gamma-ray energy. Proc. R. Soc. A 156:578–596, 1936.
Das, I. J., Ding, G. X. and Ahnesjö, A. Small fields: doi:10.1098/rspa.1936.0169
Nonequilibrium radiation dosimetry. Med. Phys. 35 Greening, J. R. An experimental examination of theories of
(1):206–215, 2008. doi:10.1118/1.2815356 cavity ionization. Br. J. Radiol. 30 (353):254–262, 1957.
Ding, G. X. and Ding, F. Beam characteristics and stop- doi:10.1259/0007-1285-30-353-254
ping-power ratios of small radiosurgery photon Greening, J. R. Fundamentals of Radiation Dosimetry. Bristol:
beams. Phys. Med. Biol. 57 (17):5509–5521, 2012. Adam Hilger, 1981.
doi:10.1088/0031-9155/57/17/5509 Halliday, D., Resnick, R. and Walker, J. Fundamental of Physics.
Dutreix, J., Dutreix, A. and Tubiana, M. Electronic equilibrium 10th edition. New York: Wiley, 2014.
and transition stages. Phys. Med. Biol. 10:177–190, 1965. Harder, D. Fano’s theorem and the multiple scattering correc-
doi:10.1088/0031-9155/10/2/302 tion. In Proc. 4th Symposium Microdosimetry, Verbania
Dutreix, J., Desgrez, A., Bok, B. and Vinot, J. H. Biophysique Pallanza, Italy, September 1973, edited by H. G. E. J.
des Radiations et Imagerie Médicale. 3rd edition. Paris: Booz, R. Eickel and A. Walker, pp. 677–693. Luxembourg:
Masson, 1993. Commission of the European Communities, 1974. core.
Evans, R. D. The Atomic Nucleus. New York: McGraw Hill, ac.uk/download/pdf/12168944.pdf
1955. Harder, D., Harigel, G. and Schultze, K. Tracks of fast elec-
Fano, U. Note on the Bragg-Gray cavity principle for measur- trons; pictures of electron and roentgen irradiation of a
ing energy dissipation. Radiat. Res. 1 (3):237–240, 1954. 35-MeV betatron with a propane-filled bubble chamber.
doi:10.2307/3570368 Strahlentherapie 115:1–21, 1961.
Fenwick, J. D., Kumar, S., Scott, A. J. and Nahum, A. E. Hendee, W. R. Medical Radiation Physics. 2nd edition. Chicago,
Using cavity theory to describe the dependence on detec- IL: Year Book Medical Publishers, 1979.
tor density of dosimeter response in non-equilibrium Highland, V. L. Some practical remarks on multiple scatter-
small fields. Phys. Med. Biol. 58 (9):2901–2923, 2013. ing. Nucl. Instrum. Methods 129 (2):497–499, 1975.
doi:10.1088/0031-9155/58/9/2901 doi:10.1016/0029-554X(75)90743-0
Part A References 89

Hogstrom, K. R. and Almond, P. R. The effect of multiple ICRU. Report 90. Key Data for Ionizing-Radiation Dosimetry:
scattering on dose measured in non-water phantoms. Measurement Standards and Applications. J. ICRU 14 (1),
Med. Phys. 9:607, 1982. doi:10.1002/j.2473-4209.1982. 2016. doi:10.1093/jicru/ndw043
tb36173.x ICRU. Report 91. Prescribing, recording, and reporting of ste-
Horowitz, Y. S. and Dubi, A. A proposed modification of Burlin’s reotactic treatments with small photon beams. J. ICRU 14
general cavity theory for photons. Phys. Med. Biol. 27 (2), 2017. doi:10.1093/jicru/ndx017
(6):867–870, 1982. doi:10.1088/0031-9155/27/6/008 IPEM (Institute of Physics and Engineering in Medicine).
Hubbell, J. H. Review of photon interaction cross section data Report 103. Small Field MV photon dosimetry – see
in the medical and biological context. Phys. Med. Biol. 44 Aspradakis et al., 2010.
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

(1):R1–R22, 1999. doi:10.1088/0031-9155/44/1/001 Janssens, A. A proposed modification of Burlin’s general cavity


Hubbell, J. H., Veigele, W. J., Briggs, E. A., Brown, R. T., theory for Photons. Phys. Med. Biol. 28 (6):745–747, 1983.
Cromer, D. T. and Howerton, R. J. Atomic form factors, doi:10.1088/0031-9155/28/6/015
incoherent scattering functions, and photon scattering Janssens, A., Eggermont, G., Jacobs, R. and Thielens, G.
cross sections. J. Phys. Chem. Ref. Data 4 (3):471–538, Spectrum perturbation and energy deposition mod-
1975. doi:10.1063/1.555523 els for stopping power ratio calculations in general cav-
Hubbell, J. H. and Øverbø, I. Relativistic atomic form factors and ity theory. Phys. Med. Biol. 19 (5):619–630, 1974.
photon coherent scattering cross sections. J. Phys. Chem. doi:10.1088/0031-9155/19/5/003
Ref. Data 8 (1):69–105, 1979. doi:10.1063/1.555593 Johns, H. E. and Cunningham, J. R. The Physics of Radiology.
Hubbell, J. H., Trehan, P. N., Singh, N., Chand, B., Mehta, 4th edition. Springfield, IL: Charles C. Thomas, 1983.
D., Garg, M. L. et al. A review, bibliography, and tab- Kase, K. R. and Nelson, W. R. Concepts of Radiation Dosimetry.
ulation of K, L, and higher atomic shell X-ray fluores- Oxford: Pergamon Press, 1978.
cence yields. J. Phys. Chem. Ref. Data 23 (2):339, 1994. Kearsley,E.Anewgeneralcavitytheory.Phys.Med.Biol.29(10):1179–
doi:10.1063/1.555955 1187, 1984. doi:10.1088/0031-9155/29/10/001
Hubbell, J. H. and Seltzer, S. M. Tables of X-ray Mass Kellerer, A. M., Hahn, K. and Rossi, H. H. Intermediate dosi-
Attenuation Coefficients and Mass Energy-absorption metric quantities. Radiat. Res. 130 (1):15–25, 1992.
Coefficients 1 keV to 20 MeV for Elements ZZ1 to 92 doi:10.2307/3578474
and 48 Additional Substances of Dosimetric Interest. US Kumar, S., Deshpande, D. D. and Nahum, A. E. Monte-Carlo-
Department of Commerce, Gaithersburg, MD, 1995. derived insights into dose-kerma-collision kerma inter-
IAEA (International Atomic Energy Agency). Absorbed Dose relationships for 50 keV-25 MeV photon beams in water,
Determination in Photon and Electron Beams: An aluminum and copper. Phys. Med. Biol. 60 (2):501–519,
International Code of Practice. Technical Report Series 2015a. doi:10.1088/0031-9155/60/2/501
No. 277. Revised version. Vienna: IAEA, 1997. Kumar, S., Fenwick, J. D., Underwood, T. S., Deshpande, D.
IAEA. Dosimetry of Small Static Fields Used in External Beam D., Scott, A. J. and Nahum, A. E. Breakdown of Bragg-
Radiotherapy. Technical Report Series No. 483, Vienna: Gray behaviour for low-density detectors under electronic
IAEA, 2017. www-pub.iaea.org/MTCD/Publications/ disequilibrium conditions in small megavoltage pho-
PDF/D483_web.pdf ton fields. Phys. Med. Biol. 60 (20):8187–8212, 2015b.
ICRU (International Commission on Radiation Units and doi:10.1088/0031-9155/60/20/8187
Measurements). Recommendations of the International Kumar, S. and Nahum, A. E. Secondary bremsstrahlung and
Commission on Radiological Units, 1950. In the energy-conservation aspects of kerma in photon-irra-
Recommendations of the International Commission on diated media. Phys. Med. Biol. 61 (3):1389–1402, 2016.
Radiological Protection and of the International Commission doi:10.1088/0031-9155/61/3/1389
on Radiological Units, 1950, pp. 19–29. Washington: National Kumar, S., Deshpande, D. D. and Nahum, A. E. Dosimetric
Bureau of Standards, 1951. https://babel.hathitrust.org/ response of variable-size cavities in photon-irradiated
cgi/pt?id=mdp.39015095063999;view=1up;seq=3 media and the behaviour of the Spencer-Attix cavity inte-
ICRU. Report 16. Linear Energy Transfer. Bethesda, MD: gral with increasing Delta. Phys. Med. Biol. 61 (7):2680–
ICRU, 1970. doi:10.1093/jicru/os9.1.Report16 2704, 2016. doi:10.1088/0031-9155/61/7/2680
ICRU. Report 33. Radiation Quantities and Units. Bethesda, Li, X. A. and Rogers, D. W. Electron mass scattering powers:
MD: ICRU, 1980. doi:10.1093/jicru/os17.2.Report33 Monte Carlo and analytical calculations. Med. Phys. 22
ICRU. Report 36. Microdosimetry. Bethesda, MD: ICRU, (5):531–541, 1995. doi:10.1118/1.597582
1983. doi:10.1093/jicru/os19.1.Report36 Loevinger, R. The dosimetry of beta sources in tissue; the
ICRU. Report 35. Radiation Dosimetry: Electron Beams with point-source function. Radiology 66 (1):55–62, 1956.
Energies between 1 and 50 MeV. Bethesda, MD: ICRU, doi:10.1148/66.1.55
1984a. doi:10.1093/jicru/os18.2.Report35 Ma, C.-M. and Nahum, A. E. Bragg-Gray theory and ion
ICRU. Report 37. Stopping Powers for Electrons and Positrons. chamber dosimetry for photon beams. Phys. Med. Biol. 36
Bethesda, MD: ICRU, 1984b. doi: 10.1093/jicru/ (4):413–428, 1991. doi:10.1088/0031-9155/36/4/001
os19.2.Report37 McKerracher, C. and Thwaites, D. I. Assessment of new small-
ICRU. Report 60. Fundamental Quantities and Units for field detectors against standard-field detectors for practi-
Ionizing Radiation. Bethesda, MD: ICRU, 1998. cal stereotactic beam data acquisition. Phys. Med. Biol. 44
doi:10.1093/jicru/os31.1.Report60 (9):2143–2160, 1999. doi:10.1088/0031-9155/44/9/303
ICRU. Report 85a. Fundamental quantities and units for ioniz- Metcalfe, P., Kron, T. and Hoban, P. The Physics of Radiotherapy
ing radiation (revised). J. ICRU 11 (1), 2011. doi:10.1093/ X-Rays from Linear Accelerators. Madison, WI: Medical
jicru/ndr012 Physics Publishing, 1997.
90 Part A: Fundamentals

Mobit, P. N., Nahum, A. E. and Mayles, P. An EGS4 Monte Carlo Rajan, G. Advanced Medical Radiation Dosimetry. New Delhi:
examination of general cavity theory. Phys. Med. Biol. 42 Prentice-Hall of India, 1992.
(7):1319–1334, 1997. doi:10.1088/0031-9155/42/7/007 Rogers, D. W. O. and Townson, R. W. On calculating kerma,
Mobit, P. N., Sandison, G. A. and Nahum, A. E. Photon fluence collision kerma and radiative yields. Med. Phys. 46
perturbation correction factors for solid state detectors (11):5173–5184, 2019. doi:10.1002/mp.13744
irradiated in kilovoltage photon beams. Phys. Med. Biol. 45 Rossi, B. High-Energy Particles. Englewood Cliffs, NJ: Prentice-
(2):267–277, 2000. doi:10.1088/0031-9155/45/2/302 Hall, 1952.
Molière, G. Theorie der Streuung schneller geladener Teilchen Rossi, H. H. Microscopic energy distribution in irradiated mat-
I. Einzelstreuung am abgeschirmten Coulomb-Feld. ter. In Radiation Dosimetry. Vol. I, edited by F. H. Attix
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

Zeitschrift für Naturforschung A 2 (3):133–145, 1947. and W. C. Roesch, pp. 43–92. New York: Academic Press,
doi:10.1515/zna-1947-0302 1968.
Møller, C. Zur Theorie des Durchgangs schneller Elektronen Sánchez-Doblado, F., Hartmann, G. H., Pena, J., Roselló, J. V.,
durch Materie. Annalen der Physik 406 (5):531–585, Russiello, G. and Gonzalez-Castaño, D. M. A new method
1932. doi:10.1002/andp.19324060506 for output factor determination in MLC shaped narrow
Nahum, A. E. Calculations of Electron Flux Spectra in beams. Phys. Med. 23 (2):58–66, 2007. doi:10.1016/j.
Water Irradiated with Megavoltage Electron and ejmp.2007.03.002
Photon Beams with Applications to Dosimetry. PhD, Scott, A. J., Nahum, A. E. and Fenwick, J. D. Using a Monte
University of Edinburgh, 1976. www.era.lib.ed.ac.uk/ Carlo model to predict dosimetric properties of small
handle/1842/17774 radiotherapy photon fields. Med. Phys. 35 (10):4671–
Nahum, A. E. Water/air mass stopping power ratios for mega- 4684, 2008. doi:10.1118/1.2975223
voltage photon and electron beams. Phys. Med. Biol. 23 Scott, A. J., Nahum, A. E. and Fenwick, J. D. Monte Carlo
(1):24–38, 1978. doi:10.1088/0031-9155/23/1/002 modeling of small photon fields: Quantifying the impact
Nahum, A. E. “Stopping Powers and Dosimetry.” Annual of focal spot size on source occlusion and output fac-
meeting of the Canadian Association of Physicists, tors, and exploring miniphantom design for small-field
Quebec, 18-23 June 1983 (available as NRCC report no. measurements. Med. Phys. 36 (7):3132–3144, 2009.
PXNR-2653) doi:10.1118/1.3152866
Nahum, A. E. The interactions of electrons with matter, Vol. Scott, A. J., Kumar, S., Nahum, A. E. and Fenwick, J. D.
I. Energy losses, stopping power and range. In The Characterizing the influence of detector density on
Computation of Dose Distributions in Electron Beam dosimeter response in non-equilibrium small pho-
Radiotherapy, edited by A. E. Nahum, pp. 27–55. ton fields. Phys. Med. Biol. 57 (14):4461–4476, 2012.
Madison, WI: Medical Physics Publishing, 1985. doi:10.1088/0031-9155/57/14/4461
Nahum, A. E. Perturbation effects in dosimetry: Part I. Seltzer, S. M. Calculation of photon mass energy-transfer and
Kilovoltage x-rays and electrons. Phys. Med. Biol. 41 mass energy-absorption coefficients. Radiat. Res. 136
(9):1531–1580, 1996. doi:10.1088/0031-9155/41/9/001 (2):147–170, 1993. doi:10.2307/3578607
Nahum, A. E. Cavity theory, stopping power ratios, correc- Seltzer, S. M., Hubbell, J. H. and Berger, M. J. Some theoretical
tion factors. In Clinical Dosimetry Measurements in aspects of electron and photon dosimetry. In National and
Radiotherapy, AAPM Summer School, edited by D.W.O. International Standardization of Radiation Dosimetry,
Rogers and J. Cygler pp. 91–136. Madison, WI: Medical Vol. 2. IAEA-SN-222/05, pp. 3–43. Vienna: IAEA, 1978.
Physics Publishing, 2009. Spencer, L. V. and Attix, F. H. A theory of cavity ionization.
Nahum, A. E. and Brahme, A. Electron depth-dose distributions Radiat. Res. 3 (3):239–254, 1955. doi:10.2307/3570326
in uniform and nonuniform media. In The Computation of Sternheimer, R. M. Interaction of radiation with matter. In
Dose Distributions in Electron Beam Radiotherapy, edited Methods of Experimental Physics. Vol. 5A, edited by L. C.
by A. E. Nahum, pp. 98–127. Madison, WI: Medical L. Yuan and C. S. Wu, pp. 1–89. New York: Academic
Physics Publishing, 1985. Press, 1961.
NCRP (National Council on Radiation Protection and Storm, E. and Israel, H. I. Photon Cross Sections from 1 keV to
Measurements). Report 51. Radiation protection design 100 MeV for Elements Z=1 to Z=100. Los Alamos Scientific
guidelines for 0.1 to 100 MeV particle accelerator facili- Laboratory, New Mexico 1970. www.ge.infn.it/geant4/
ties. Washington, DC: NCRP, 1977. temp/saracco/cor/Storm_israel_photon_pub_1970.pdf
Nedunchezhian, K., Aswath, N., Thiruppathy, M. and Taylor, M., Kairn, T., Kron, T., Dunn, L., Johnston, P. N. and
Thirugnanamurthy, S. Boron neutron capture therapy – a Franich, R. D. The influence of field size on stopping-
literature review. J. Clin. Diagn. Res 10 (12):ZE01–ZE04, power ratios in- and out-of-field: Quantitative data for
2016. doi:10.7860/JCDR/2016/19890.9024 the BrainLAB m3 micro-multileaf collimator. J. Appl.
Oliver, P. A. K. and Thomson, R. M. Cavity theory applica- Clin. Med. Phys. 13 (6):4019, 2012. doi:10.1120/jacmp.
tions for kilovoltage cellular dosimetry. Phys. Med. Biol. 62 v13i6.4019
(11):4440–4459, 2017. doi:10.1088/1361-6560/aa6a42 Underwood, T. S., Winter, H. C., Hill, M. A. and Fenwick, J.
Papiez, L. and Battista, J. J. Radiance and particle flu- D. Modifying detector designs for small field dosimetry.
ence. Phys. Med. Biol. 39 (6):1053–1062, 1994. Radiother. Oncol. 103 (Suppl 1 ESTRO 31 Barcelona):S206,
doi:10.1088/0031-9155/39/6/011 2012. doi:10.1016/S0167-8140(12)70851-4
Part A References 91

Underwood, T. S., Winter, H. C., Hill, M. A. and Fenwick, Underwood, T. S., Thompson, J., Bird, L., Scott, A. J., Patmore,
J. D. Mass-density compensation can improve the P., Winter, H. C. et al. Validation of a prototype DiodeAir
performance of a range of different detectors under for small field dosimetry. Phys. Med. Biol. 60 (7):2939–
non-equilibrium conditions. Phys. Med. Biol. 58 2953, 2015. doi:10.1088/0031-9155/60/7/2939
(23):8295–8310, 2013. doi:10.1088/0031-9155/58/23/ Whyte, G. N. Principles of Radiation Dosimetry. New York:
8295 Wiley, 1959.
Downloaded By: 10.3.97.143 At: 20:07 15 Nov 2023; For: 9780429201493, chapter5, 10.1201/9780429201493-7

You might also like