Interlaced proton grid therapy: development of aninnovative radiation treatment techniqueThomas Henry

Academic dissertation for the Degree of Doctor of Philosophy in Medical Radiation Physicsat Stockholm University to be publicly defended on Friday 9 November 2018 at 09.00 in CCKLecture Hall, Building R8, Karolinska University Hospital, Solna.

AbstractSpatially fractionated radiotherapy, also known as grid therapy (GRID), has been used for more than a century to try to treatseveral kinds of lesions. Yet, the grid technique remains a relatively unknown and uncommon treatment modality nowadays.Spatially fractionated beams, instead of conventional homogeneous fields, have been used to exploit the experimentalfinding that normal tissue can tolerate higher doses when smaller tissue volumes are irradiated. This increase in tolerancewith reducing beam size is known as the dose-volume effect. Despite the fact that targets were given inhomogeneous dosedistribution, sometimes with some volumes receiving close to no dose, good results in the form of shrinking of bulky tumorshave been observed in palliative treatments. The biological processes responsible for this effect are still under discussion,with several possible causes. However, numerous experiments on mice, rats and pigs have confirmed the existence ofsuch effect, which in turn motivates the present development of grid therapy.While mainly photons have been used in gridtherapy, proton and ion grid therapies are also emerging as potential alternatives. In this work, an innovative form of gridtherapy was proposed. Grids of proton beamlets were interlaced over a target volume with the intention of achieving twomain objectives: (1) to keep the grid pattern (made of adjacent high and low doses) from the skin up to the vicinity of thetarget while (2) delivering nearly homogeneous dose to the target volume. This interlaced proton grid therapy was exploredwith the use of different beam sizes, from conventional sizes deliverable at modern proton facilities, down to millimetersized beams. Other considerations that would prevent its clinical use, such as the variable relative biological effectivenessof protons or the use of cone beam computed tomography, were also evaluated. The overall aim was to assess if, and how,such treatment modality could be applied clinically, from a physics and dosimetry point of view. While it presented severaltheoretical advantages, its potential issues of concern and limitations were also evaluated.

Keywords: proton therapy, grid therapy, spatially fractionated therapy, interlacing.

Stockholm 2018http://urn.kb.se/resolve?urn=urn:nbn:se:su:diva-160427

ISBN 978-91-7797-442-0ISBN 978-91-7797-443-7

Department of Physics

Stockholm University, 106 91 Stockholm

INTERLACED PROTON GRID THERAPY: DEVELOPMENT OF ANINNOVATIVE RADIATION TREATMENT TECHNIQUE 

Thomas Henry

Interlaced proton grid therapy:development of an innovativeradiation treatment technique 

Thomas Henry

©Thomas Henry, Stockholm University 2018 ISBN print 978-91-7797-442-0ISBN PDF 978-91-7797-443-7 Cover picture by PixelAnarchy @ Pixabay, www.pixabay.com Printed in Sweden by Universitetsservice US-AB, Stockholm 2018Distributor: Department of Physics 

i

Abstract

Spatially fractionated radiotherapy, also known as grid therapy (GRID), has

been used for more than a century to try to treat several kinds of lesions. Yet,

the grid technique remains a relatively unknown and uncommon treatment

modality nowadays. Spatially fractionated beams, instead of conventional

homogeneous fields, have been used to exploit the experimental finding that

normal tissue can tolerate higher doses when smaller tissue volumes are

irradiated. This increase in tolerance with reducing beam size is known as

the dose-volume effect. Despite the fact that targets were given

inhomogeneous dose distribution, sometimes with some volumes receiving

close to no dose, good results in the form of shrinking of bulky tumors have

been observed in palliative treatments. The biological processes responsible

for this effect are still under discussion, with several possible causes.

However, numerous experiments on mice, rats and pigs have confirmed the

existence of such effect, which in turn motivates the present development of

grid therapy.

While mainly photons have been used in grid therapy, proton and ion

grid therapies are also emerging as potential alternatives.

In this work, an innovative form of grid therapy was proposed. Grids of

proton beamlets were interlaced over a target volume with the intention of

achieving two main objectives: (1) to keep the grid pattern (made of adjacent

high and low doses) from the skin up to the vicinity of the target while (2)

delivering nearly homogeneous dose to the target volume. This interlaced

proton grid therapy was explored with the use of different beam sizes, from

conventional sizes deliverable at modern proton facilities, down to

millimeter sized beams. Other considerations that would prevent its clinical

use, such as the variable relative biological effectiveness of protons or the

use of cone beam computed tomography, were also evaluated.

The overall aim was to assess if, and how, such treatment modality

could be applied clinically, from a physics and dosimetry point of view.

While it presented several theoretical advantages, its potential issues of

concern and limitations were also evaluated. The results from six studies are

here presented and summarized.

ii

iii

Sammanfattning

Oönskad bestrålning av friska vävnader har alltid varit ett bekymmer för

cancerbehandlingar med joniserande strålning då detta kan inducera

oönskade bieffekter. Redan 1909, bara några år efter att strålbehandling hade

börjat användas, var en tysk läkare med namnet Alban Köhler oroad över

strålningsinducerade hudskador som uppkommit i samband med

strålbehandling. Han hade då en ide att ersätta konventionell strålbehandling,

i vilken patienten bestrålas med homogena strålfält, med heterogena strålfält

där vissa regioner skulle få en hög stråldos och vissa regioner inte skulle få

någon stråldos alls. Med detta ville Köhler öka vävnandens tolerans för

joniserande strålning. Detta bestrålningsmönster har även kallats

"schackbräde" eller "rutnät" (grid). Grid-terapi föddes.

Efter att långsamt blivit bortglömt genom åren, återupptäcktes grid-

terapi på nittiotalet som en teknik för att bestråla skrymmande tumörer som

andra konventionella behandlingstekniker misslyckades med att behandla.

Observationer har visat att toleransen mot strålning hos vissa vävnader ökar

när små regioner av dessa vävnader inte bestrålas. I samband med att

strålbehandling med protoner eller tyngre joner blivit allt vanligare kan grid-

terapi utvecklas ytterligare.

Denna avhandling innehåller en omfattande redogörelse av grid-terapins

historia, från att konceptet introducerades 1909 till de senaste experimentella

studierna som använder väldigt finmaskiga bestrålningsmönster med väldigt

små strålar. Därefter beskrivs utvecklingen av en innovativ typ av grid-terapi

som använder sig av sammanflätade protonstrålar. Bakgrunden till behovet

av denna teknik och dess fördelar presenteras. Resultat från sex olika studier

sammanfattas och diskuteras.

iv

v

List of papers

PAPER I: Proton-grid therapy (PGT): a proof-of-concept study

T. Henry, A. Ureba, A. Valdman & A. Siegbahn,

Technology in Cancer Research & Treatment 16, 749–757

(2017).

PAPER II: Dosimetric Comparison of Plans for Photon- or

Proton-Beam Based Radiosurgery of Liver Metastases

G. Mondlane, M. Gubanski, P.A. Lind, T. Henry, A.

Ureba & A. Siegbahn, International Journal of Particle

Therapy 3, 277–284 (2016).

PAPER III: Robustness of interlaced proton grid therapy plans

against pencil-beam spot position variations

T. Henry & J. Öden, Manuscript

PAPER IV: Development of an interlaced-crossfiring geometry for

proton grid therapy

T. Henry, N. Bassler, A. Ureba, T. Tsubouchi, A.

Valdman & A. Siegbahn, Acta Oncologica 56, 1437–1443

(2017)

PAPER V: Interlaced proton grid therapy: linear energy transfer

and relative biological effectiveness distributions

T. Henry & J. Öden, Submitted to Physica Medica (2018)

PAPER VI: Organ doses from a proton gantry-mounted CBCT

system characterized with MCNP6 and GATE

O. Ardenfors, T. Henry, I. Gudowska, G. Poludniowski &

A. Dasu, Physica Medica 53, 56-61 (2018)

vi

Publication relevant to, but not included in the thesis:

PAPER VII: Quantitative evaluation of potential irradiation

geometries for carbon-ion beam grid therapy

T. Tsubouchi, T.Henry, A. Ureba, N. Bassler & A.

Siegbahn, Medical Physics 45, 1210–1221 (2018).

vii

Author’s contribution

PAPER I: I designed the study, produced and analyzed all the results

and was the main writer of the manuscript.

PAPER II: I took part in reviewing some of the plans included in this

study and helped revising the manuscript.

PAPER III: I designed the study, produced and analyzed all the results

and was the main writer of the manuscript.

PAPER IV: I designed the study, produced and analyzed all the results

and was the main writer of the manuscript.

PAPER V: The paper was a close collaboration between the co-author

and myself. Contribution to the design of the study,

production/analysis of the results and writing of the

manuscript was equally shared between the two of us.

PAPER VI: I took part in all the measurement sessions, scientific

discussions and in the design of the study, and was in charge

of all the GATE-based simulations. I took part in reviewing

the manuscript.

viii

ix

Outline of the thesis

This thesis focuses on exposing the train of thoughts that, over the past four

years, led to the development of the proposed interlaced proton grid

technique in its current state. As such, an extensive historical background

and state-of-the-art of grid therapy is first presented, giving the reader the

scientific bases on which the proposed proton grid therapy grew. In the

second part of this thesis, the studies performed in papers I-VI are

summarized to give an overview of the different aspects of the in-

development treatment modality.

Part of the text found in this thesis originates from my licentiate thesis

that was presented in June 2017. As such, the following chapters, sub-

sections and paragraphs were partially or fully reproduced from the said

licentiate to the present thesis:

The Abstract was partially reused and improved.

The Introduction was reused, improved and updated to provide a better

overview of the entire project.

Chapters 2 and 3 were originally included in the licentiate thesis. Apart from

small modifications, figures 2,8 and 10 were replaced with more instructive

figures and figure 11 was added.

Chapters 4 and 5 are based on sub-sections from the licentiate thesis. Some

parts of said sub-sections were reused, improved, complemented and made

up-to-date with the latest studies.

x

xi

Contents

Abstract ................................................................................................ i

Sammanfattning ................................................................................. iii

List of papers ....................................................................................... v

Author’s contribution ....................................................................... vii

Outline of the thesis ........................................................................... ix

Contents .............................................................................................. xi

Abbreviations ................................................................................... xiii

1. Introduction ............................................................................... 15

2. Grid Therapy ............................................................................. 19

2.1. The history of grid therapy ........................................................... 19

2.2. Radiobiological basis of grid therapy: the dose-volume effect ..... 21

2.3. Contemporary grid therapy ........................................................... 23

2.4. Experimental grid therapy with µm- and mm-wide beamlets ....... 25

3. Towards the development of an interlaced proton grid

therapy ............................................................................................... 29

3.1. Objectives and questions to be answered ...................................... 29

3.2. Grid geometries and irradiation setups ......................................... 30

4. Using conventional beam sizes: a first step ............................. 33

4.1. Establishing the potential with commercial treatment planning

systems ...................................................................................................... 33

4.2. Going further: planning clinically realistic grid plans .................. 34

4.3. Impact of spot position uncertainties ............................................ 36

4.4. Discussion ..................................................................................... 37

5. Proton grid therapy with small beamlets ................................ 41

5.1. Establishing the potential through Monte Carlo simulations ........ 41

xii

5.2. A motorized collimator design to produce millimeter-wide

beamlets .................................................................................................... 42

5.3. Discussion ..................................................................................... 45

6. Clinical challenges: RBE and CBCT imaging ........................ 47

6.1. RBE considerations ....................................................................... 47

6.2. The use of CBCT imaging ............................................................ 48

6.3. Discussion ..................................................................................... 50

7. General discussion and concluding remarks .......................... 53

8. Summary of papers ................................................................... 55

Acknowledgments ............................................................................. 57

References .......................................................................................... 59

xiii

Abbreviations

BEV Beam’s eye view

CBCT Cone beam computed tomography

C-T-C Center-to-center

DVH Dose-volume histogram

ED50 Effective dose 50

FWHM Full width at half maximum

IMPT Intensity-modulated proton therapy

IMRT Intensity-modulated radiation therapy

LET Linear energy transfer

MC Monte Carlo

MCS Multiple Coulomb scattering

MRT Microbeam radiation therapy

NTCP Normal tissue complication probability

OAR Organ at risk

PTV Planning target volume

PVDR Peak-to-valley dose ratio

QUANTEC Quantitative analyses of normal tissue effects in the clinic

RBE Relative biological effectiveness

SBRT Stereotactic body radiation therapy

SFUD Single field uniform dose

xiv

SOBP Spread-out Bragg peak

TCP Tumor control probability

TPS Treatment planning system

VMAT Volumetric modulated arc therapy

VPDR Valley-to-peak dose ratio

Chapter 1. Introduction

15

1. Introduction

For almost 120 years, from the discovery of x rays in 1895 by Wilhelm

Röntgen until now, ionizing radiations have been used as a regular modality

to treat cancerous tumors and other lesions. In search for the best treatment

outcome, new treatment modalities, machines and particle types have since

been introduced into the clinic.

As knowledge was gained over the years, more effort was put on

reducing the toxicity of radiotherapy treatments to the skin and healthy

tissues beneath it and around the lesion. Techniques using broad radiation

beams like intensity-modulated radiation therapy (IMRT), volumetric

modulated arc therapy (VMAT) with photons or intensity-modulated proton

therapy (IMPT) with protons, where the target is irradiated with beams

incident from several directions to lower the normal tissue dose, were

developed. In parallel, temporal dose fractionation i.e. whether the dose

should be given in several fractions over time or in a single fraction

(radiosurgery), has become the standard of care. Radiotherapy as a whole

has been greatly focusing on protecting the surrounding healthy tissues while

delivering a clinically relevant dose to the targeted tumor.

In proton therapy, the special dosimetric characteristics of charged

particles are exploited. As charged particles, protons interact through

numerous collisions with the atomic electrons of the medium in which they

are propagating, slowly reducing their speed. As a consequence, the slower

the particles become, the higher their probability of interaction with the

medium, increasing the energy locally deposited. This will result in a high

energy being deposited as the protons are completely stopped, known as the

Bragg peak. Furthermore, in addition to slowing down, protons experience

multiple elastic interactions with the positively charged atomic nuclei in the

medium, leading to these protons being deflected from their initial path and

scattered in the medium. This is known as multiple Coulomb scattering

(MCS). Finally, a proton can also undergo non-elastic collisions with the

said atomic nuclei, during which the proton is absorbed by the nucleus,

leading to the potential ejection of numerous particles: secondary protons

(which can have a small impact on the spatial dose distribution of the beam),

deuterons and heavier ions (which deposit all their energy locally due to

their small range and energy) and neutrons, which are very penetrating and

16

should therefore be given a special care as they can deposit energy far from

the interaction point.

In radiotherapy, the limited range of the protons (and more generally

heavy ions) is therefore used to improve dose conformity to the target and

consequently the outcome of radiotherapy treatments. From single field

uniform dose treatments (SFUD) and IMPT, where targets are given

homogeneous doses, to proton stereotactic body radiation therapy (proton

SBRT or proton radiosurgery) where high inhomogeneous doses are allowed

in the targeted volumes, proton therapy has greatly developed. It is worth

mentioning also that protons and other heavier charged particles have been

known for many years to have a different relative biological effectiveness

(RBE) compared to conventionally used photons (Durante and Paganetti,

2016). While a constant RBE value of 1.1 has been assumed for protons,

studies pointing towards the existence of a variable RBE have been

conducted in the past years (Paganetti, 2014).

One form of therapy that was developed to allow the further reduction

of the toxicity to normal tissue is the so-called grid therapy. In grid therapy,

narrow beams are arranged in a chess-board pattern leading to a spatial

fractionation of the beams, with parts of tissue left unirradiated in between

the beams. The idea of spatially fractionating the beams to increase the

healthy tissue tolerance to the treatment is not new (Köhler, 1909; Laissue et

al., 2012). The method has been abandoned and then rediscovered several

times since its introduction. It came back some years ago as an option to

treat bulky tumors for which there were no other treatment options available

(Huhn et al., 2006; Kaiser et al., 2013; Mohiuddin et al., 2015, 1999, 1990;

Neuner et al., 2012; Peñagarícano et al., 2010).

The motivation for using small beams in place of regular broad,

homogeneous fields directly comes from the experimental observation that

normal tissue can tolerate a higher dose if the radiation dose is deposited in a

small volume. This finding is known as the dose-volume effect and has been

mainly observed on skin and spinal cord in several animal experiments (Bijl

et al., 2002; Hopewell and Trott, 2000).

Historically, only photons have been used in grid therapy. The spatial

fractionation of the beams has been obtained with the use of perforated

collimators attached directly on the treatment head. However, the arrival of

more recent techniques such as proton and heavy-ion therapies opened a new

door for grid therapy. Because of the interesting depth-dose characteristics of

charged particles in the human body, a new proton or ion grid therapy has

emerged.

Chapter 1. Introduction

17

Until recently, most research and treatments using grid therapy had

been conducted using only one incident grid, itself consisting of numerous

micro-, mini- or centimeter-wide beams. Using such irradiation geometry,

not only the normal tissue but also the target receives a spatially fractionated

dose, i.e. some parts of the tumor/lesion receive very high doses (peak doses)

and some parts receive low doses (valley doses). Some attempts have been

made to deliver a nearly homogeneous dose to the target using only one

incident grid (Dilmanian et al., 2015; Zlobinskaya et al., 2013). This resulted

in the loss of the grid pattern of the dose distribution (made of consecutive

peaks and valleys of dose, theoretically responsible for the improved

tolerance of the tissue) already a few centimeters in depth beneath the skin,

i.e. a high and nearly homogeneous dose could already be observed several

centimeters before the target. This could be of importance since most

radiation necrosis in normal tissues seem to appear in the high dose regions,

near the target (Lawrence et al., 2010).

In this work an innovative grid method using proton beams was

proposed. The main objectives were the following: (1) to maintain the grid

pattern of the dose distribution, made of alternating high and low doses, in

the normal tissue up to the direct vicinity of the target, and (2) to deliver a

rather homogeneous dose to the target (with a high minimum dose). To do

so, grids of proton beamlets incident from several directions were interlaced

over the targeted volume.

This grid geometry, along with the use of proton beams, is anticipated

to produce desirable dosimetric results in terms of target volume dose

coverage and normal tissue sparing. In return, this would make the proposed

grid therapy a potential strong candidate as a new modality for curative

treatments, and not just palliative last-chance attempts.

This thesis summarizes the development of such interlaced proton grid

therapy, from the very first proof-of-concept study using conventional

proton beam sizes available at any modern proton facility, to the latest

advancements. Along the way, different beam sizes were explored, down to

millimeter sized beamlets, and more clinical concerns such as the impact of

variable RBE on the clinical validity of grid plans or the use of cone beam

computed tomography (CBCT) and its implication were considered. All of

which to achieve a unique goal: develop the proposed technique to,

eventually, make it closer to a clinical application and relevant as a stand-

alone treatment modality.

18

Chapter 2. Grid Therapy

19

2. Grid Therapy

2.1. The history of grid therapy

The history of grid therapy can be traced back to 1909 (Köhler, 1909;

Laissue et al., 2012), when Dr. Alban Köhler, a German physician, proposed

to attach an iron wire net with 1 mm-square holes to a radiation source,

pushed to the patient’s skin, with the aim of irradiating this latter with small,

separated beams. The suggested irradiation method was expected to increase

the maximum tolerated skin dose manyfold and therefore, achieve a better

skin sparing (Köhler, 1909). He tested this grid setup in 1912 on several

patients, irradiating the skin with up to approximately 60 Gy, and observed

the healing of skin necrosis in several weeks.

Figure 1. Schematic illustration of a typical grid with circular holes arranged in a

hexagonal pattern.

In 1933, Dr. Liberson, from the U.S Marine Hospital of New York City,

revived the concept of grid therapy and published a new study, without any

apparent knowledge of Köhler’s work (Liberson, 1933). In this paper,

Liberson explored different possible geometries for a perforated screen

(Figure 1), trying out squared, oval or round perforations, comprising one-

quarter, one-half or two-thirds of the total screen area. Similar to Köhler, his

main concern was the skin protection. Skin reactions were commonly

observed at the time of orthovoltage x ray therapy. He finally opted for a

1.5-2 mm-thick lead sheet, perforated with 2 to 8 mm circular holes in

diameter, comprising half of the total screen area. Liberson then applied his

grid technique on small animals, mice and rabbits, but also on a few human

2.1. The history of grid therapy

20

cases (mostly inoperable head & neck, lung, esophagus and bone cases) and

noticed an increased skin tolerance when using his perforated screen. He did

not however study other organs and focused on the skin toxicity.

One year later, in 1934, Dr. Wilhelm Haring performed his own study

of grid therapy (Haring, 1934). The grid he used consisted of a 25x25 cm2

large and 3 mm-thick lead sheet, perforated with 3 mm holes in diameter and

comprising half of the total screen area, much like Liberson’s perforated

screen. With the use of this grid, he irradiated a total of 20 patients, mainly

lung and stomach carcinomas and one case of pelvis sarcoma. He was able

to deliver approximately 5 Gy to the skin without noticing any important

unexpected effects. However, Haring remained relatively vague on the

outcome of the treated cases, providing only details on the evolution of the

skin. He simply mentioned that in the pelvis case, no further growth of the

tumor was observed after irradiation.

Dr. Hirsch Marks, in 1950, reported three cases of grid treatments

(Marks, 1950). The grids employed were made of 3 mm-thick lead-rubber

sheets, perforated with 0.25 cm2 to 4 cm2 squared holes (depending on the

grid used) comprising 40% of the total screen area. The three patients

presented with either vulva mass, suprapubic mass or tongue carcinoma.

After grid irradiations of 184, 210 and 212 Gy over 35, 18 and 20 treatment

days respectively, there was no clinical evidence of remaining disease for

any of the treated cases.

Finally, in 1953, in his book titled X-Ray Sieve Therapy in Cancer: A

Connective Tissue Problem, Dr. Benjamin Jolles provided an extensive study

of grid therapy, focusing on the radiobiology of the connective tissue, the

irradiated tissues and their neighborhood (Jolles, 1953). He notably

emphasized the protective role of the normal tissue and its part in the

repairing of irradiated adjacent tissues which could explain the results

witnessed with grid therapy. In the last chapter of his book, Jolles presented

detailed information on grid treatments for a total of 56 patients, which most

of the other authors failed to provide, focusing only on the evolution of the

skin after treatment and not on the treatment itself. The treated patients were

separated in two groups: the “accessible tumors” group, comprising 36

patients (12 skin, 3 anus, 5 vulva, 7 breast, 9 lymph nodes in neck and

groins) and the “deep-seated tumors” group containing 20 patients (8

carcinoma in uterus and ovary, 2 lung, 7 bowel, 3 miscellaneous). From

these study cases, Jolles reported remarkable results for many patients, some

with very advanced diseases that would have only been offered very limited

or no palliative care with conventional therapy at that time. In the first

patient group, the response to treatment for the skin cases was excellent

Chapter 2. Grid Therapy

21

while the treatments of the anus and vulva cancers provided a high degree of

palliation. However, the response of lymph nodes of neck and breast cancers

was poor. In the second group of patients (the deep-seated tumors), the

overall response was “fairly good”, with a degree of palliation way above

anything conventional therapy would have offered. For this second group,

Jolles emphasized that the curative value of the treatment had still to be

demonstrated, but concluded that grid therapy had proved to be useful in

some difficult cases which could be successfully irradiated without any risk

of skin necrosis.

2.2. Radiobiological basis of grid therapy: the dose-

volume effect

The rationale for using small and spatially-fractionated radiation beams to

achieve a higher normal tissue sparing is directly related to the dose-volume

effect, which has been historically described for single beam irradiations

(Hopewell and Trott, 2000).

Experiments on animals, including rats (Bijl et al., 2006, 2003, 2002,

Hopewell et al., 1987, 1986; Laissue et al., 2013; van der Kogel, 1993; Van

Luijk et al., 2005), pigs (Hopewell et al., 1986; van den Aardweg et al.,

1995; van der Kogel, 1993), monkeys (Schultheiss et al., 1994) and dogs

(Powers et al., 1998) have showed that the tolerance of the organs studied

(mainly skin and spinal cord) to radiation greatly increases with a decreasing

beam size (Figure 2). In the case of rat spinal cord irradiation, a steep rise in

the effective dose 50 (ED50), i.e. the dose required to expect a particular

endpoint to occur in 50% of the irradiated animals, was observed for

beamlets smaller than 8 mm when looking at the paralysis in fore or hind

limbs, or both, be it with photons (Hopewell et al., 1987; van der Kogel,

1993) or protons (Bijl et al., 2002, refered to as "Present study" in Figure 2).

In studies of the pig skin, similar results were observed for beamlets which

were 9 mm-wide and smaller when the acute reaction moist desquamation

and late dermal thinning endpoints were considered.

2.2. Radiobiological basis of grid therapy: the dose-volume effect

22

Figure 2. Effective dose 50 (ED50) (Gy) for the limb paralysis endpoint as a

function of irradiated spinal cord length in rats, using photons (Hopewell et al.,

1987; van der Kogel, 1993) or protons (Bijl et al., 2002, "present study") (reprinted

from Bijl et al., 2002 with permission from Elsevier).

The reason why the tolerance of the normal tissue increases when the

irradiated volume becomes smaller is still to-date not fully understood. But

several hypotheses have been proposed to explain the dose-volume effect.

Migration of unirradiated cells to irradiated volumes to repair the

radiation damage was suggested already in 1953 by Jolles as stated

previously (Jolles, 1953). Since then, other studies have explored this

process (Hopewell and Trott, 2000; Shirato et al., 1995; Withers et al.,

1988). The undamaged cells are believed to be able to migrate a short

distance (~2 mm) from the unirradiated to the irradiated volumes and

contribute to the repair of the locally damaged cells.

Furthermore, while a “classical” bystander effect has been observed and

described extensively, even for grid therapy (Asur et al., 2015), for which a

decrease in the survival of unexposed cells next to irradiated cells was

witnessed, some evidence seems to point towards the existence of another

type of bystander effect (described as type III in Mackonis et al., 2007). In

this case, communication between cells receiving a high dose and close cells

receiving a low dose is thought to lead to an increased survival of the heavily

irradiated cells. In grid therapy, this could occur when cells in the peak

region (high dose) are more efficiently repaired if the neighboring cells in

the valley region (low dose) are undamaged.

Chapter 2. Grid Therapy

23

2.3. Contemporary grid therapy

After the Second World War, the development of high-energy (MV) and

Cobalt-60 treatment units marked an important turn in the history of grid

therapy. With the new devices then available, the dose to the skin was

reduced due to the buildup effect of dose for high (1-20 MV) energy

photons. The concept of grid therapy was slowly abandoned and the number

of patients treated was greatly reduced.

Grid therapy has however resurfaced in the last three decades: it has

been occasionally performed at few clinics around the world as an attempt to

shrink bulky tumors that other treatment modalities failed to treat, with

impressive results (Huhn et al., 2006; Kaiser et al., 2013; Mohiuddin et al.,

2015, 1999, 1990; Neuner et al., 2012; Peñagarícano et al., 2010). It has

mainly been used for head-and-neck, thoracic and abdominal cancers, most

of the time in combination with other treatment modalities.

Figure 3. Photograph of a grid collimator used clinically (reprinted from Zwicker et

al., 2004 with permission from Elsevier).

Beam sizes of approximately 1 cm at the patient skin have been produced

with the help of collimators attached to a photon unit gantry head (Figure 3).

Only a single incident grid has normally been used, leading to an

inhomogeneous dose being delivered in the targeted volume, similar to what

can be observed in some brachytherapy treatments (Figure 4). But despite

the fact that some parts of the tumor are receiving very low doses, significant

reduction in the size of these bulky tumors has been observed (Figure 5).

2.3. Contemporary grid therapy

24

Figure 4. (a) Beam’s eye view (BEV) of one multileaf collimator strip over the

target breast in a patient with locally advanced breast cancer; (b) dosimetric profile

of spatially fractionated GRID radiotherapy beamlets in BEV. (reprinted from

Neuner et al., 2012 with permission from Elsevier)

Figure 5. Neck nodes from primary squamous cell cancer of the oropharynx. (Left)

Before and (Right) after a 15 Gy grid irradiation (reprinted from Mohiuddin et al.,

1999 with permission from Elsevier).

Chapter 2. Grid Therapy

25

In Figure 6, a typical grid lateral dose profile, with the associated grid-

specific parameters, is depicted. The characteristic succession of high doses

(peaks) and low doses (valleys) is shown, as well as the center-to-center (c-t-

c) distance separating each beam contained in the grid. This particular

example was Monte Carlo generated with 2 mm-wide proton beams, full

width at half maximum (FWHM), with a c-t-c distance of 4 mm.

Figure 6. Illustration of a typical grid lateral dose profile with 2 mm-wide (FWHM)

115 MeV monoenergetic proton beamlets and a c-t-c distance of 4 mm in water. The

profile was taken at the water tank entrance. (Monte Carlo simulation with the code

TOPAS).

2.4. Experimental grid therapy with µm- and mm-wide

beamlets

Another form of photon grid therapy which has been studied for the last

three decades is the so-called microbeam radiation therapy (MRT)

(Blattmann et al., 2005; Bouchet et al., 2013, 2010, Bräuer-Krisch et al.,

2010, 2005; Crosbie et al., 2010; Dilmanian et al., 2007, 2006, 2002; Grotzer

et al., 2015; Laissue et al., 2013; Serduc et al., 2010, 2009, 2008; Slatkin et

al., 1992; Smilowitz et al., 2006; Van Der Sanden et al., 2010). In MRT,

grids containing quasi-parallel photon beamlets produced by a synchrotron

accelerator with widths in the micrometer range (typically 25-100 µm) are

used. The beamlets comprised in the grid are usually spaced with distances

2.4. Experimental grid therapy with µm- and mm-wide beamlets

26

of 100-400 µm (Figure 7) and peak entrance doses up to 10 000 Gy have

been investigated. MRT is an extreme case where the dose-volume effect is

exploited maximally, with extremely small beamlets to greatly increase the

radiation tolerance of the underlying tissues.

Figure 7. Horizontal section of the cerebellum of a piglet 15 months after irradiation

with a skin entrance dose of 300 Gy. Some cells and their nuclei directly in the path

of microbeams were destroyed. There was no tissue destruction present, nor were

there signs of hemorrhage. The paths of the microbeams appear in the section as

thin, white horizontal parallel stripes, which are more easily visible in the insert.

Beam width 27 µm, spacing 210 µm. (reprinted from Blattmann et al., 2005 with

permission from Elsevier)

Even though only a small volume of the target receives high doses of

radiation, good tumor control has been reported after MRT (Dilmanian et al.,

2003; Laissue et al., 1998; Miura et al., 2006). Preferential damage seems to

occur to the tumor microvasculature when irradiating with microbeams. This

process has been shown mainly for rat brain tumors (Bouchet et al., 2013,

2010). In these studies, a differential effect of the microbeams on the tumor

vasculature has been observed, leading to increased blood vessel

permeability and to tumor hypoxia. However, no such effects were detected

in the normal brain vasculature.

Most MRT studies have been focusing on rat and mouse brain tissues,

with few additional data on the spinal cord and skin. Experiments on pigs are

currently ongoing, while the first clinical trials are in preparation (Bräuer-

Krisch et al., 2015; Schültke et al., 2017).

With the recent developments in proton and heavy-ion therapy a new

kind of grid therapy has been suggested. In proton and ion grid therapy, the

Chapter 2. Grid Therapy

27

special dose characteristics of charged particles are taken advantage of. With

their limited range, a conformal dose distribution on the distal edge of the

target volume can be produced. Furthermore, with the right c-t-c distance

between the beams and due to multiple Coulomb scattering of the charged

particles in the medium, the initially spatially fractionated beamlets will

overlap at a certain depth and produce a nearly homogeneous dose

distribution over the target as depicted in Figure 8. Proton and ion grid

therapies have mostly been explored with the use of minibeams, i.e.

beamlets with widths in the order of 0.3-0.7 mm (Dilmanian et al., 2015,

2012; Lee et al., 2016; Peucelle et al., 2015; Prezado et al., 2017; Prezado

and Fois, 2013) and sometimes microbeams (Kłodowska et al., 2015;

Zlobinskaya et al., 2013). A recent work also explored the use of

conventional, non-collimated proton beam sizes in a clinical grid application

(Gao et al., 2018). Overall, only few studies have been published so far, but

it is a growing field of research that has shown great potential for future

clinical applications.

Figure 8. Example of the dose pattern produced by a single grid irradiation in water

with 2 mm-wide (FWHM) 115 MeV monoenergetic proton beamlets and a c-t-c

distance of 4 mm (Monte Carlo simulation using the code TOPAS).

2.4. Experimental grid therapy with µm- and mm-wide beamlets

28

Chapter 3. Towards the development of an interlaced proton grid therapy

29

3. Towards the development of an

interlaced proton grid therapy

3.1. Objectives and questions to be answered

Several major differences in how grid therapy can be delivered can be

identified when studying retrospectively the history of grid therapy, from the

very first attempts in the beginning of the 20th century until today’s research

on this topic. With grid therapy, the dose distributions produced inside the

irradiated target volume can be of two different types:

Type 1: The beamlets are well separated throughout the irradiated

volume and the target, producing the grid pattern of the dose

distribution (“peaks” of high doses and “valleys” of lower doses) in

the normal tissue and in the targeted volume as well.

Type 2: The beamlets are well separated at the entrance but overlap

in depth due to scattering in the volume, creating a nearly

homogeneous dose distribution in the target but also several

centimeters in front of it.

In this project, these two approaches were merged to create an improved

approach for grid therapy. The objectives for the suggested approach were

the following:

1. To maintain the grid pattern of the dose distribution (peaks and

valleys) in the normal tissue from the entrance all the way to the

direct vicinity (< 1 cm) of the target, in order to keep the normal

tissue toxicity as low as possible for all organs, and not only for the

skin.

2. To deliver a nearly homogeneous dose to the targeted volume to

achieve a high tumor control probability (TCP).

To fulfill these objectives, an improved grid geometry was suggested.

By interlacing grids of proton beamlets over a target volume, a dose

distribution similar to what can be obtained with conventional proton

therapy inside a target volume can be achieved. Outside of the target volume

however, a well-defined grid pattern in the dose distribution can still be

preserved in the normal tissue down to the direct vicinity of the target. The

3.2. Grid geometries and irradiation setups

30

interlaced geometry has been used before (Dilmanian et al., 2012, 2006;

Lomax and Schaer, 2012; Prezado and Fois, 2013; Serduc et al., 2010) but

with different objectives and/or particle types. Prezado and Fois, 2013

presented results with proton beamlets. In their study, proton beamlets with

an energy of 1 GeV were used to irradiate a volume with the proposed

geometry but due to the high kinetic energy of the protons, the beamlets

propagated way beyond the target. The Bragg-peaks were therefore located

outside of the body. A similar grid irradiation with proton beams as used in

this thesis was presented in the abstract from Lomax and Schaer, 2012. Only

very limited information about the planning strategies being, however,

provided.

The work presented here aimed to use this improved grid geometry and

prove its potential usefullness for radiotherapy treatments. To do so, several

questions had to be answered:

1. Is it possible to deliver this kind of grid irradiation with currently

available proton beams at modern proton facilities with the help of a

commercial treatment planning system (TPS)? (Paper I)

2. How do these interlaced grid irradiations compare with conventional

photon and proton treatments? (Papers II & V)

3. How robust are these irradiations against spot position uncertainties,

considering the precision required for interlacing the beams? (Paper

III)

4. What kind of interlaced grid irradiations could be produced with

smaller beamlets, to exploit the dose-volume effect even further

(Paper IV), and how can one produce these small beamlets at

existing proton facilities?

5. What are the clinical implications of such irradiation modality with

focus on the RBE (Paper V) and the use of CBCT imaging? (Paper

VI)

3.2. Grid geometries and irradiation setups

Two types of grid were evaluated in this work:

A 1-D grid consisting of parallel planar beamlets as pictured in

Figure 9.a

A 2-D grid comprising parallel circular beamlets (Figure 9.b)

Chapter 3. Towards the development of an interlaced proton grid therapy

31

Figure 9. Schematic illustration of (a) a 1-D and (b) a 2-D grid. The center-to-center

(c-t-c) distance between the beamlets is shown for both cases.

Both of these grid types could be used depending on the

situation/objective/studied case. The 2-D grid type offers a higher normal

tissue protection, due to the spatial fractionation of the beamlets in two

dimensions, leaving a higher volume of non-irradiated tissue in-between the

beamlets. However, the 1-D grid type offers higher dose coverage of a

target, due to a larger volume being irradiated by one single grid.

To create a nearly homogeneous dose distribution inside the targeted

volume, the grids were interlaced as described previously and illustrated in

Figure 10. With this setup, one can arrange the beam grids to interlace over

the target. Due to multiple Coulomb scattering in the medium through which

the protons are transported, the elemental beamlets contained in the grids

will overlap in the targeted volume, which will produce a nearly

homogeneous dose.

Intrinsically, interlacing two opposing 1-D grids might be enough to

produce sufficient dose coverage inside a specific target volume. Adding two

additional opposed 1-D grids (2x2 opposing 1-D grids) will increase the total

target dose, while decreasing the peak and valley doses outside of the target.

However, when using 2-D grids, 2x2 opposing grids seem to be mandatory

to achieve reasonable dose coverage of the target volume.

Sometimes, the use of opposing grid angles is not possible, due to e.g.

the close proximity of a sensitive organ to the targeted volume. In that case,

another grid irradiation geometry can be used: 2 non-opposing 1-D grids, in

which the two grids are interlaced without the need to be opposing (Figure

11), allowing for a better sparing of the tissues located beyond the distal

edge of the different beams.

The c-t-c distance between the beamlets inside a single grid can also be

adapted. Intuitively, when the c-t-c distance is increased, a decrease in the

valley dose is obtained. However, obtaining good dose coverage of the

3.2. Grid geometries and irradiation setups

32

irradiated target with a large c-t-c distance will be challenging, due to a

reduced number of beamlets contributing to the total dose inside the target.

Figure 10. Illustration of the interlacing of 2x2 opposing 2-D grids inside a cubic

target volume, each grid containing 9 beamlets.

Figure 11. Illustration of the interlacing of 2 non-opposing 1-D grids.

Chapter 4. Using conventional beam sizes: a first step

33

4. Using conventional beam sizes: a first

step

While small beamlets in the µm- and mm-range seem to be more

advantageous for tissue sparing according to studies focusing on the dose-

volume effect (see 2.2), “larger” beamlets could still be beneficial compared

to the broad fields used in conventional proton therapy. Indeed, modern

photon grid therapy treatments have shown that good results in terms of

shrinking of bulky tumors and skin toxicity could still be achieved with

spatially fractionated beams with sizes of approximately 1 cm at the patient

skin (Huhn et al., 2006; Kaiser et al., 2013; Mohiuddin et al., 2015, 1999,

1990; Neuner et al., 2012; Peñagarícano et al., 2010).

At a modern proton facility center such as The Skandion Clinic in

Uppsala, proton beams with approximate widths of 7-15 mm FWHM,

depending on the energy, can be produced. For beamlets of these sizes, a

small but not insignificant dose-volume effect could be expected (as

observed in animal experiments, see 2.2).

In that regard, it appeared that exploring the interlaced proton grid

therapy presented in this work with these relatively large beamlets was an

evident first step serving as a proof-of-concept, before digging into the use

of mm-wide beamlets. A notable therapeutic advantage can still potentially

be achieved, and working with beams that do not require any kind of

collimation is a tremendous practical advantage, the spatial fractionation of

the beams being accomplished by pencil beam scanning.

4.1. Establishing the potential with commercial treatment

planning systems

The first study of this thesis work then focused on the possibility to carry out

interlaced grid therapy treatments using grids containing beamlets currently

available at a modern proton therapy center, with the grid geometries

described in 3.2 (Paper I). To do so, the TPS Eclipse Protons (Varian

Medical Systems, Palo Alto, USA) was used to replan two patients that were

previously treated with conventional photon therapy at Karolinska

4.2. Going further: planning clinically realistic grid plans

34

University Hospital in Stockholm for liver or rectal cancer. The patients

were replanned using the suggested interlaced proton grid therapy with

either 1-D or 2-D grids.

The results of this proof-of-concept study showed that it was possible to

irradiate the targeted disease (here the planning target volume, PTV) with a

high and nearly homogeneous dose, while preserving the grid pattern of the

dose distribution down to the vicinity of the target. When normalizing the

minimum target dose to 100%, the valley dose in the normal tissue was

approximately 5-20%, depending on the depth considered. The peak doses

ranged from approximately 50 to 120%. Even in the neighborhood of the

PTV (2 cm in front), a peak-to-valley dose ratio (PVDR) of 5 could still be

observed. The target dose ranged from 100% to 120-140% depending on the

grid geometry used.

This work demonstrated the potential of such proton grid therapy from a

dosimetric point of view. In this study, only already clinically available tools

were used, showing that the interlaced proton grid therapy technique

described in this paper is potentially achievable without overcoming

important technical issues.

As a comparison, more conventional proton therapy techniques (proton

radiosurgery) on the liver cancer case were investigated in Paper II. In this

study, ten patients diagnosed with liver metastasis and treated with SBRT at

Karolinska University Hospital were replanned using IMPT with the same

dose objectives set to the targets as in the photon SBRT plans. Patient no. 1

in this Paper II is the patient used for grid planning in Paper I. While similar

doses could be delivered to the PTV, a clear improvement regarding the

sparing of the organs at risk was noticed when using IMPT instead of photon

SBRT. Comparison of photon SBRT, proton SBRT and proton grid plans

can be found in Figure 12. With grid therapy, and based on the dose-volume

effect, our hypothesis is that an even greater normal tissue sparing could be

achieved compared to the studied IMPT radiosurgery.

4.2. Going further: planning clinically realistic grid plans

As stated, the work presented in Paper I served as a proof-of-concept for the

presented grid geometries, using regular proton beam sizes. In that sense, the

focus was put on fulfilling the objectives set previously, i.e. a nearly

homogeneous PTV dose in combination with a well-defined grid pattern in

the normal tissue. However, the different organs at risk (OARs) were given

no particular considerations, making these liver and rectal proton grid plans

probably clinically unrealistic.

Chapter 4. Using conventional beam sizes: a first step

35

Figure 12. Comparison of (a) conventional photon therapy (SBRT), (b) interlaced

proton grid and (c) proton radiosurgery (SBRT) plans on a liver cancer case.

This aspect was studied later on in the project, when looking at RBE for

interlaced proton grid therapy (Paper V). While a full report of the impact of

RBE on conventional and proton grid plans can be found in section 6.1, the

focus here will be on the comparison between conventional and grid plans

with a constant RBE of 1.1.

For this study, the TPS Raystation (RaySearch Laboratories AB,

Stockholm, Sweden) was used to replan five patients, each presenting tumor

volumes in a different location (esophagus, lung, liver, prostate and anus).

The same grid geometries as previously described were used. For

comparison, conventional IMPT plans were also produced. This resulted in a

total of five plans for each patient (one IMPT + four grid). To make both the

IMPT and grid plans clinically relevant, dose objectives to the different

OARs were set according to suitable sources (Emami, 2013; Marks et al.,

2010; Muirhead et al., 2016). Plan optimization was done on the PTV and

OARs.

An example of dose distributions for a conventional IMPT plan using

two fields (200° & 270°) and a grid plan (2x2 opposing 2-D grids, with grid

angles 20°, 110°, 200° & 290°) is presented in Figure 13. Of the 20 grid

plans computed, 18 fulfilled all the clinical goals set for the OARs and the

PTV. The failure of the 2 remaining grid plans was due to the close

proximity of the duodenum in the liver case that would not allow the use of

certain grid geometries. In terms of average doses to the OARs, most of the

passing grid plans were very comparable to the conventional IMPT plans.

4.3. Impact of spot position uncertainties

36

Figure 13. 2-D dose distributions of (a) a conventional two field proton plan, and (b) a 2x2

opposing 2-D proton grid plan, for a lung case.

4.3. Impact of spot position uncertainties

Eventually, this second study confirmed that producing interlaced proton

grid plans that are clinically relevant, with respect to clinical goals to the

OARs and the PTV, is possible with regular beam sizes available at modern

proton centers. However, there are still numerous limitations and questions

that need to be addressed.

The most evident one is the precision in spot delivery. Indeed, all the

presented plans, either with Eclipse or Raystation, assumed perfectly

positioned spots at the desired positions. However, it is commonly accepted

that during quality assurance of proton beams, a tolerance of < 1 mm

(sometimes < 2 mm) is allowed. What would then be the impact of small

mispositionings (within tolerance) of the spots for interlaced proton grid

plans? The interlacing of the beams is of course a central point of the

technique presented here. It is therefore crucial to understand the

consequences of an imperfect interlacing as a result of variability in spot

positioning. It will most likely give rise to local cold and hot spots in the

targeted volume due to a potential overlap of the beams in this latter.

To quantify these effects, a study was performed (Paper III). The idea

was to use the scripting capabilities of Raystation to introduce an uncertainty

in the spot positioning. The grid plans described in section 4.2 were in fact

created using a Python script in Raystation to position the spots. The same

patients and grid plans were reused. A script was created to include a

random error in the X and Y spot positions after the initial plan had been

optimized, based on a Gaussian probability distribution with σ= 0.5 mm and

cut-offs at ± 1 mm. Then, this new grid plan’s dose with slightly out-of-

position spots was recalculated without any optimization process. Dose

statistics and dose-volume histograms (DVHs) were recorded. This was done

Chapter 4. Using conventional beam sizes: a first step

37

30 times for each plan to study the average impact of such spot positioning

errors on all the 18 goal-fulfilling grid plans from the previous study.

Preliminary results have shown that, in the case of small variations in the

positions of the spots, the PTV dose coverage could be slightly altered and

doses could go below and above the clinically accepted dose thresholds

(96% of the PTV should receive 95-107% of the prescribed dose), while

OARs doses would remain almost identical and below limits for most cases.

The observed trend was that the introduction of an imperfect interlacing led

to the occurrence of localized hot and cold spots inside and in the close

vicinity of the PTV. Of the 540 plans computed with spot position

variations, 341 failed the clinical goals set for the PTV. An underdosage of

this latter was observed in 216 plans while an overdosage was seen in 33

plans. The remaining 92 plans presented both an underdosage and an

overdosage of the PTV. However, while the recorded doses of interest were

indeed out of clinically accepted boundaries, they remained very close to

these clinical goals, opening the door for potential robust planning. Also, the

grid geometry played an important role in the success of the plans, with the

2x2 opposing 1-D grid geometry delivering the best PTV dose homogeneity

and highest passing rate of all geometries, while the 2x2 opposing 2-D grid

geometry was the less efficient in terms of dose coverage and passing rate.

4.4. Discussion

The studies discussed in this chapter present a proof-of-feasibility of

interlaced proton grid therapy using beam sizes producible by any modern

proton beam line. The validity of such treatment modality was explored with

the use of treatment planning systems only, and has to be further explored in

a more applied way. Indeed, delivering interlaced proton grid plans will give

rise to additional issues and limitations that have not been evaluated yet in

the TPS.

The concern of spot mispositioning has already been touched on in 4.3,

but could now be assessed for real irradiations. The idea would be to deliver

actual proton grid plans in a phantom at The Skandion Clinic and, similarly,

to recalculate the plan’s dose distribution using the spot positions recorded

after irradiation in the system’s log files. This could be done on a daily basis

to evaluate the average deviation to the original plan, and would provide a

preliminary indication whether or not interlaced proton grid treatments could

be safely delivered at this specific center, on a dosimetric level exclusively.

4.4. Discussion

38

Another obvious issue one may encounter when delivering proton grid

plans is organ motion. For most grid geometries used throughout this entire

thesis (namely the 2 opposing & non-opposing 1-D and 2x2 opposing 2-D

grids), some volumes within the PTV are mainly irradiated by only one

beam (see Figure 10). In the case of organ motion, it is then very likely that

one or more beams could end up being misplaced compared to the planned

spot positions, leading to local hot and cold spots, or even completely

unirradiated small volumes in an extreme case. What would then be the

consequence of such altered dose distributions on the validity of the

treatment plan, for the OARs of course, but more specifically its curative

value? As it was previously observed (section 2.3), current clinical grid

treatments using photons are purposely delivering very heterogeneous doses

in the targeted volume, with some parts receiving very little or close to no

dose. But these treatments are usually followed by more conventional

irradiations. Therefore, if interlaced proton grid therapy is to be evaluated as

a stand-alone treatment modality, as it is the long-term objective here, the

organ motion problem must be addressed. It is uncertain what the best

solution would be, but it definitely requires additional studies. One

workaround would be to first limit proton grid experiments to brain-type

irradiations. There, organ motion seems to be minimal (Enzmann and Pelc,

1992). Moreover, with the help of high precision stereotactic frames, which

can be found in Gamma Knife treatments for example (Heck et al., 2007),

good accuracy in the positioning, within < 1 mm, could potentially be

achieved.

Another concern for interlaced proton grid therapy is range uncertainty.

However, this concern is not expected to be more significant for proton grid

than for, e.g., IMPT. Indeed, the exact same beamline and beam data are

used in the two modalities, the only difference being the beam arrangement.

Therefore, range uncertainty should be dealt with in proton grid as it is in

IMPT. Furthermore, taking advantage of the robust optimization approaches

developed for IMPT, one could spend extra effort on assessing the variable

spot position issue discussed previously. One could think of a special grid

robust planning where tighter PTV doses objectives would be set in order to

prevent any failure of the plans. Also, as the grid geometry used seems to

have a significant impact on the overall passing rate of the plans, extra

attention could be given to choosing the right geometry for the right

situation. Sometimes, as it was the case with the liver patient studied here,

the patient’s geometry and locations of certain OARs prevent the use of

some grid angles/geometries, otherwise exceeding the dose constraints set

for said OARs. A compromise might have to be found in search for the best

Chapter 4. Using conventional beam sizes: a first step

39

option; unless, of course, a small dose-volume effect can still be proved for

these rather large pencil beams, in which case the number of possible grid

geometries and angles for such difficult cases might not be decreased.

Finally, few words should be said about dosimetry. If experimental

interlaced proton grid plans are to be delivered in a phantom as suggested

earlier, they will have to be somehow evaluated. Checking the spot positions

should be fairly easy with existing tools: log files from the system, film

measurements, 2D scintillators etc. However, dose homogeneity in the PTV

has to be verified as well, and one should not only rely on the correct spot

positions to assume that interlacing was done properly. This could prove to

be rather cumbersome. 3D detectors like gels or organic scintillators could

be used but these detectors are prone to quenching, which would add another

issue to solve. 2D arrays of ionization chambers are also an option but they

provide only limited spatial information due to their quite low resolution. A

good compromise/solution has to be found and will be explored in future

works.

4.4. Discussion

40

Chapter 5. Proton grid therapy with small beamlets

41

5. Proton grid therapy with small beamlets

The use of small proton beamlets would theoretically enable a better

exploitation of the dose-volume effect, and might therefore prove to be a

better choice to achieve a higher protection of the sensitive tissues. However,

there are additional concerns with smaller beamlets that need to be

evaluated, such as altered dosimetric properties compared with broad beams,

and the need for extra collimation to create these small beamlets.

5.1. Establishing the potential through Monte Carlo

simulations

To determine if what was presented with “large” beams would be possible to

achieve with beamlets which are smaller than those used clinically today, i.e.

beamlets with widths in the range 1-3 mm FWHM, a Monte Carlo study was

performed (Paper IV).

The aim of this specific study was to perform a simple interlaced proton

grid irradiation of a small cubic target at the center of a water tank, with the

same grid geometries proposed in section 3.2, and the same dose objectives

as previously stated: well defined grid pattern outside of the irradiated

volume and homogeneous dose distribution inside. An extensive focus was

put on assessing the impact of the beamlet width and the c-t-c distance on

the final dose distribution. The Geant4-based MC code TOPAS was used

(Perl et al., 2012).

Also, efforts were made in identifying and establishing the additional

parameters required to characterize grid therapy. There were four parameters

in total: the c-t-c distance, the initial beamlet width and the peak dose, all of

which were discussed previously, but more importantly the valley-to-peak

dose ratio (VPDR). Until now, all grid studies reported the “quality” of the

grid as the peak-to-valley dose ratio. However, this ratio appears to be

numerically unstable and divergent with very low valley doses. Hence, from

this point onwards, the use of the VPDR, that equals 0 for a perfect grid with

no valley dose, and 1 for homogeneously distributed dose, was preferred and

seems to have been adopted by other research groups (Gao et al., 2018).

The results of this study showed that with the optimal c-t-c distances, a

high and nearly homogeneous target dose could be achieved, with a standard

5.2. A motorized collimator design to produce millimeter-wide beamlets

42

deviation of less than ± 5% in all cases (Figure 14). Furthermore, outside of

the target, the grid pattern of the dose distribution could still be maintained

in the normal tissue, with peak doses more than twice as high as the valley

doses in the direct vicinity of the target (1 cm).

Hence, the use of small proton beamlets in the millimeter range is

compatible with the dosimetric objectives set in section 3.1. What was

observed with larger proton beams seems to be reproducible with smaller

beamlets, for a very simple case. The dose distributions obtained with these

beam grids, containing mm-wide beamlets, however need to be further

evaluated for clinically more realistic and complex cases. Tools are under

development with this aim, and are further discussed in section 5.3.

Figure 14. Example of 2x2 opposing 1-D grids interlacing in a cubic target with 2

mm-wide beams and a c-t-c distance of 7 mm (Monte Carlo simulation with the

code TOPAS).

5.2. A motorized collimator design to produce

millimeter-wide beamlets

The production of mm-wide proton beamlets at current proton facilities

could be cumbersome. Such centers are usually equipped with commercial

Chapter 5. Proton grid therapy with small beamlets

43

proton treatment solutions that are designed to better fit clinical

requirements. Unfortunately, the use of small beamlets is not a priority.

Therefore, collimation of the already producible beams seems mandatory.

That poses an extra challenge when working with proton beams, and

more specifically proton beams delivered with pencil beam scanning. Since

the beam can move in the two dimensions perpendicular to the beam

propagation direction, three possible ways of collimating such beams were

identified:

1. The easy way is to make the beam fixed, and move the table

instead to create the spatial fractionation. Then, a collimator

with a simple hole in the middle and with the desired size and

shape can be inserted in the nozzle’s holder (Figure 15 left).

2. A more advanced solution would be to attach a collimator to the

nozzle’s holder with holes already arranged in a grid pattern,

with the desired size and c-t-c distance between them. (Figure

15 middle).

3. Finally, the most advanced solution would be to have a small

collimator that can move in two dimensions, accordingly to the

beam’s position. (Figure 15 right).

Figure 15. Different grid-oriented collimation options for proton pencil beam

scanning.

In these three options the holes could of course be of any shape: circular for

example, to obtain a 2-D grid, or elongated, to have a 1-D grid.

While it is definitely the easiest solution for research purposes, option 1

is not really clinically applicable as it would require numerous small table

movements with high precision. Option 2 is the go-to solution for the vast

majority of grid-focused research groups nowadays. It ensures a good

positioning of the beams thanks to high precision capabilities in industrial

5.2. A motorized collimator design to produce millimeter-wide beamlets

44

manufacturing, and remains quite easy to use. However, it lacks

customization possibilities: the c-t-c distance, the size and position of the

holes and the total number of holes are fixed. Thus, even the slightest change

of one of these parameters requires the manufacturing of a new similar

collimator. The price of such pieces to have some flexibility in the grid

geometry could then add up quickly. Also, since the collimator as a whole is

fixed in the nozzle’s holder, the interlacing between e.g. two opposing

beams would have to be done using small table movements.

It was consequently decided that option 3 was the most suitable for the

current needs and applications. A motorized collimator, capable of moving

in two dimensions, was therefore designed.

The system is made of the following parts (Figure 16):

- Two pairs of motor + linear guide, each producing linear motion in

one dimension (1)

- One linear support to distribute the weight over several points (2)

- A cylindrical brass collimator (3)

- A mounting frame to hold all the different pieces together and

designed to mount the system on the Skandion Clinic’s IBA nozzle

(4)

Figure 16. 3D model of the designed motorized collimator.

This collimator system was recently manufactured, and preliminary

studies evaluating the induced activity in the collimator after proton

irradiations were performed. Testing and irradiation measurements at The

Chapter 5. Proton grid therapy with small beamlets

45

Skandion Clinic are planned for the near future to prove the concept and

assess the efficiency of such system.

5.3. Discussion

Determining what would be possible to do with smaller beamlets than those

currently delivered at modern proton centers was the natural evolution of this

project. Indeed, all observations of the dose-volume effect point towards the

increase of tolerance of tissues to radiation beams with decreasing size of

said beams (see 2.2).

While studies on micro- and minibeams are currently being conducted

by several research groups (see 2.4), it appeared unreasonable to evaluate if

interlacing such beams would result in advantageous dose statistics, in

regards to the objectives set for this work (see chapter 3). The technical

requirements to properly interlace such small beams seemed too

cumbersome to imagine an easy and efficient clinical application. It was

therefore decided that beamlets in the order of 1-3 mm FWHM were more

suitable, as they could still be considered as small beams compared to

regular clinical proton beams, while being technically less difficult to

produce.

In the first study, Monte Carlo simulations were used to assess if the

grid geometries and objectives used in the previous evaluation with “broad”

beams could still be achieved with these millimeter beamlets. While it

provided encouraging results, valuable information and permitted the

establishment of a few grid-specific but important parameters, this study

remained however fairly simple with respect to the geometry - the irradiation

of a cubic target in a cubic water phantom. Unfortunately, studying more

complex irradiation cases, for example on patient CT data instead of simple

Monte Carlo geometries, requires additional parameters and greatly

increases the computation time. One cannot easily and quickly find the

optimal spot positions/spot weights for a complex case, and blindly tuning

these parameters by hand until a satisfying solution is found is not an

efficient way to proceed. Therefore, an in-house Monte Carlo-based and

grid-oriented optimizer tool is currently under development within our

research group and will be presented in a near future. With this tool,

simulating grid irradiations with Monte Carlo-generated beamlets on any

kind of geometry, including patient CT data, will be possible. This will be of

great help to evaluate the potential applications on a whole variety of clinical

cases of interlaced proton grid therapy using small beams.

5.3. Discussion

46

The issue of correct spot positioning has been discussed in the previous

chapter, but it will definitely be different when working with small beamlets.

If the described collimator is to be used to produce these beamlets then the

accuracy in the spot positioning will no longer depend on the accuracy in the

beam delivery position, but rather on the correct positioning of the

collimator itself. Thanks to the linear guide units used in the presented

prototype, this can theoretically be achieved with a fairly good accuracy

(0.04 mm) and repeatability (± 0.01 mm). Therefore, with a correct

calibration of the system (i.e. an accurate determination of the isocenter

position for the collimator), good accuracy in the collimator’s movements

should supposedly be reached. However, if this system is to be used on a

rotational proton gantry as it is currently planned, one will also have to

evaluate the variation of mechanical isocenter position with different gantry

angles. Indeed, a slight variation of e.g. 1 mm between two opposing gantry

angles could have a considerable impact on the overall quality of the

interlacing. While this mechanical variation is usually compensated for the

proton beam with the magnets, one will also have to think about calibrating

the collimator position to match the beam’s isocenter.

Nevertheless, one question remains: what are the biological

implications of using spatially fractionated beams? The lack of data for an

extended range of parameters (beam size, c-t-c distance, peak dose, valley

dose, tissue type, fractionation scheme) is preventing a clear understanding

of all the consequences of using such unconventional beam arrangements.

As it was seen with the use of small beamlets (Paper IV), the peak dose in

the normal tissue is sometimes higher than the actual target dose, depending

on the grid type, beam size and c-t-c distance. What would then be the

outcome for the normal tissue of such high but very localized doses? One

could be tempted to generalize observations of the dose-volume effect for

certain tissue types in animals (Figure 2) to all tissue types in humans,

concluding that the tolerance of the healthy tissue will be tremendously

increased considering the beams sizes used (1-3 mm). This, of course,

represents a most-likely incorrect over-simplification. Clinical reports from

simple grid irradiations from the past three decades seem to be pointing

towards the existence of a dose-volume effect in some human tissues as well.

But substantial efforts still have to be made on the radiobiological side to

fully understand the extent of this effect, and how to use it to our advantage.

Chapter 6. Clinical challenges: RBE and CBCT imaging

47

6. Clinical challenges: RBE and CBCT

imaging

The suggested interlaced proton grid technique is still at a very early stage of

development, and it is of course highly unclear how it could be transferred to

the clinic. However, to ensure that the best treatment option is proposed, one

has to think beforehand of all the underlying implications of using the

interlaced grid geometry.

6.1. RBE considerations

One issue that has to be evaluated before clinical application is directly

linked to the use of proton beams. Indeed, the use of charged particles adds

several physical and biological challenges as compared to radiotherapy with

photon beams. The correct management of the different relative biological

effectiveness (RBE) is one of them.

So far, conventional clinical proton treatments have been planned using

a constant RBE of 1.1. However, a large number of in vitro studies are

pointing towards the existence of a variable RBE, depending on the dose,

linear energy transfer (LET), tissue type and more (Paganetti, 2014).

Variable RBE is therefore an issue that can lead to increased normal tissue

dose and, as such, has to be addressed. This is of course true for regular

IMPT plans, but for proton grid as well, where the fractionation of the beams

usually leads to high doses being delivered in each single beam, and could

therefore result in an even higher increase in the normal tissue dose. Hence,

a study of the differences between dose, LET and RBE-weighted dose

distributions and their consequences for conventional and grid plans was

required.

This was performed in Paper V. The method used in this study has

already been partially described in section 4.2, when comparing clinically

relevant conventional and grid plans. The same set of patients and plans

were considered here, i.e. five patients, one IMPT plan and four grid plans

per patient.

RBE-weighted dose distributions for these plans were calculated using

the method described previously by Ödén et al. (Ödén et al., 2017b, 2017a).

6.2. The use of CBCT imaging

48

In summary, dose-averaged LET was computed on voxel level for the 20

plans (assuming a constant 1.1 RBE) using RayStation v4.6 experimental

Monte Carlo engine. Then, RBE-weighted dose distributions were calculated

following two variable RBE models (McNamara et al., 2015; Wedenberg et

al., 2013). The newly created plans taking into account variable RBE were

then compared, examining fulfillment of clinical goals for the PTV and

different OARs.

Of the 50 variable RBE plans evaluated (five patients, five plans for

each patient and two RBE models), only 15 passed the clinical goals set for

the different structures: 2 IMPT plans and 13 grid plans. Most passing plans

were for a single patient, the esophagus case, where both IMPT plans with

the two models and all the 8 grid plans were accepted. One can also note that

the five remaining passing plans were achieved with grid irradiations for the

liver and prostate case, while the respective IMPT plans failed for both

models considered.

Interestingly enough, despite the different beam arrangements between

IMPT and grid plans, results when considering variable RBE models were

very similar between the two techniques. Most plans were rejected due to the

failure of the same clinical goals: genitalia DRBE,5% for the anus case, bladder

V80 Gy (RBE) for the prostate case, duodenum DRBE,0.1cc for the liver case and

esophagus DRBE,0.1cc for the lung case.

Overall, this study showed that compared to IMPT, the use of spatially

fractionated beams does not seem to add extra difficulties when considering

variable RBE models. However, this does not mean that it is not

problematic, as it remains an issue that has to be addressed, just as it should

be in regular proton therapy.

6.2. The use of CBCT imaging

Another issue that has to be evaluated is the use of CBCT imaging. Indeed,

interlacing proton beams will most likely require high precision in motion

management (as touched briefly in section 4.4) and in positioning of the

patient. For this latter, the use of imaging will most probably be mandatory.

Modern proton therapy facilities can be equipped with CBCT systems

that are often custom made, and which could benefit the suggested proton

grid therapy for positioning purposes.

The aim in proton grid therapy is to give as low dose as possible to the

healthy tissues in between the beams (valley dose, see chapter 2). But

performing CBCT could result in a bath of potentially non-negligible dose

Chapter 6. Clinical challenges: RBE and CBCT imaging

49

over an important volume, therefore increasing the valley dose (and by

extension the valley-to-peak dose ratio), and consequently hypothetically

reducing the benefit expected from the dose-volume effect in terms of tissue

tolerance. But to which extent? Also, due to the lack of radiobiological data,

it is very uncertain what kind of fractionation regimen would the proposed

interlaced proton grid therapy benefit the most. As a consequence, one has to

be prepared to any possibility, including treatment courses spanning over

several weeks. In this case, the use of daily CBCT does not seem implausible

and poses extra challenges due to the increased dose burden of using such

imaging modality on a daily basis. This issue then required an extensive

study of the doses delivered by such CBCT system. This was done in Paper

VI.

In this work, doses to specific organs in an anthropomorphic phantom

delivered by the proton gantry-mounted CBCT system at The Skandion

Clinic in Uppsala were evaluated by means of two Monte Carlo codes,

MCNP6 (Goorley et al., 2016) and Geant4-based GATE (Jan et al., 2004).

To do so, the CBCT system was first modelled in the MC codes using

measurements of depth-dose profiles in water and lateral profiles in air.

Then, those two models were validated against absolute dose measurements

in a CTDI phantom. Finally, the validated models were used to evaluate

doses delivered by this specific CBCT system to organs in an

anthropomorphic phantom (PBU-60, Kyoto Kagaku, Kyoto, Japan),

following irradiation from three protocols (head, thorax (Figure 17) and

pelvis) and different scan modes (full 360° scan, anterior 190° scan,

posterior 190° scan). The cumulative doses from weekly and daily imaging

were also calculated.

Figure 17. Example of simulated energy deposition in the anthropomorphic

phantom for a 360° scan of the thorax (GATE).

6.3. Discussion

50

The average doses from a 360° scan to the organs located inside the

irradiation field for the three protocols (head, thorax, pelvis) ranged between

6-8 mGy, 15-17 mGy and 24-54 mGy, respectively. For daily imaging,

based on a 30-day treatment course, this resulted in doses ranging between

0.2-1.6 Gy, depending on the organ and protocol. The highest organ dose for

daily imaging was achieved on the femoral heads (1.6 Gy).

In comparison, with the same total mAs, anterior 190° scans delivered

doses 24% higher in average, while the posterior 190° scans showed a 37%

decrease. This could be due, as a first rough explanation, to the locations of

the organs evaluated, most of them sitting on the anterior side of the body

(eyes, thyroid, breasts, heart). For more centrally located organs (brain,

lungs, prostate) the difference was negligible. The trend was even inverted

for posterior organs (rectum).

This study would then suggest the use of posterior 190° scans for

clinical routine if, and only if, image quality can be preserved. The impact of

those scans on proton grid plans will be discussed in the following section.

6.3. Discussion

As it was shown in chapter 4, interlaced proton grid therapy is potentially

already applicable using conventional beam sizes. Nonetheless, one has to be

careful when using protons and spatially fractionated beams. The two studies

summarized in this chapter do not aim to blindly support an implementation

of the interlaced proton grid technique, but represent more of an attempt to

anticipate any issue that would come with its clinical application, if/when

this point is reached in the future.

In that regard, assessing the impact of RBE on the quality of the grid

plans, as compared to IMPT plans, was an important step. As it was

observed, considering variable RBE in proton grid plans did not seem to

introduce additional difficulties in comparison with the reference IMPT

plans, despite very different LET distributions. However, one should be

aware of several limitations of such studies.

First of all, results presented here and in Paper V are based on a couple

variable RBE models. As the name suggests, models use observations to

develop a relation between input parameters and observed effects. In the

case of interest for us here, the RBE models are based on in vitro data, as

they attempt to establish a connection between quantities such as LET, the

sensitivity of the tissue to fractionation expressed as the ratio of the

parameters of the LQ model for cell survival - the α/β ratio, dose and more,

Chapter 6. Clinical challenges: RBE and CBCT imaging

51

and the observed enhanced biological effect of charged particles. Can it be

assumed that these in vitro data-based models are a perfect representation of

in vivo effects? Probably not, but they nonetheless certainly provide an

interesting first guess, and should be dealt with as such: an approximation.

In that regard, some results should be handled carefully. For example, a

result in Paper V states that for the liver case, only a grid plan passed all the

clinical goals for the OARs when considering both variable RBE models,

while the IMPT and other grid plans failed. However, a closer look at the

numbers reveals that the IMPT plan was rejected due to the failure of

DRBE,0.1cc ≤ 24 Gy for the duodenum with a value of 24.5 Gy (RBE), while

the passing grid plan was accepted with a value of 23.6 Gy (RBE), both with

the McNamara model. Considering the model-based approach, is such a

minor difference really significant and enough to draw a pass/fail

conclusion?

Another question also remains: what could be done? Taking into

account variable RBE seems to lead to the same observations, independently

of whether grid or conventional plans are considered. However, it does not

mean that this impact is inexistent or negligible. If the variable RBE issue

proves to be as problematic as it seems according to the models used, it has

to be addressed. And that is where grid and conventional proton modalities

could be set apart. For IMPT plans, one solution already explored (Ödén et

al., 2017a) could be to re-optimize plans based on the LET distribution, and

aim for homogeneous RBE in the PTV instead of homogeneous physical

dose. But for grid it might be different. If radiobiological data support the

claim that using spatially fractionated beams allows for a better tolerance of

the normal tissue to radiation, one might seriously consider to take

advantage of this enhanced protection of healthy tissues to increase the target

dose. This would in turn lead to increased doses per fraction, or even very

high doses in a single fraction. Yet, some variable RBE models predict a

decreased RBE with increasing dose (McNamara et al., 2015; Wedenberg et

al., 2013). Would the variable RBE issue still be as significant then? Or

would the observed trend be inversed? Optimizing grid plans to better fit the

RBE variations could prove to be quite different from IMPT plans. Needless

to say, additional radiobiological data for grid and the dose-volume effect

are needed to better understand all the implications of using spatially

fractionated beams.

Regarding the CBCT-related dose considerations, the application will

most probably be more straightforward, although it will also require

additional data.

6.3. Discussion

52

Results presented in Paper VI showed that the use of CBCT imaging

could result in significant doses delivered to large volumes, especially in the

case of a daily use. For interlaced proton grid therapy, this would then result

in a slight increase of the peak and valley doses. Will it have any impact on

the overall protection of normal tissue? Since in a given volume/organ, the

irradiation from such imaging modality is relatively homogeneous, the

valley-to-peak dose ratio should theoretically change when including CBCT

imaging. Additionally, in the attempt to understand the origin of the dose-

volume effect, it has been hypothesized that the valley dose could have a

large impact on the observed effect, larger than the influence of the actual

ratio between valley and peak doses (see 2.2). In that case, increasing the

valley dose through CBCT imaging might reduce the benefits obtained from

the dose-volume effect.

Looking at an example from Paper VI, with the prostate case and the 2

opposing 1-D grid geometry, the valley dose in the femoral heads is around

10 Gy, for a 39 fractions/78 Gy treatment. However, the study in Paper VI

showed that the average dose to these organs after a single CBCT irradiation

could be above 53 mGy. For a 39 fractions treatment, that would result in a

total average dose to the femoral heads of above 2 Gy. For the valley dose,

this would then correspond to a 20% increase, potentially worsening the

outcome of said grid treatment for this organ. Unfortunately, the full extent

of such outcome cannot be quantified with our current knowledge of the

dose-volume effect. The lack of radiobiological data is, again, preventing the

proposed interlaced proton grid therapy to advance a step further and at this

point, one can only guess the consequences that a higher valley dose could

have.

In conclusion, using CBCT imaging for interlaced proton grid therapy

might prove to be complex in certain situations. Of course one could debate

if a treatment course spanning over 30-40 days is really realistic for this type

of treatment, but only further studies will be able to answer these questions.

In the meantime, one should be aware of the potential consequences of such

imaging modality and its use on a daily basis, and be prepared to adjust the

way interlaced proton grid plans are delivered to reach an optimal treatment

solution.

Chapter 7. General discussion and concluding remarks

53

7. General discussion and concluding

remarks

Despite the fact that grid therapy has been proposed in the last century and

used for decades, it still remains unknown to the vast majority of

physicians/physicists working in cancer-related facilities. Important

differences with conventional treatment modalities in terms of dose

distributions (dose coverage of the PTV) or technical applicability might be

reasons why it still appears as peculiar. While modern radiotherapy usually

dictates homogeneous irradiation of the targeted volume to achieve the

highest tumor control probability (TCP) while keeping the normal tissue

complication probability (NTCP) as low as possible, the lack of

experimental data does not allow a clear understanding of what would be the

TCP of a grid treatment delivering very heterogeneous dose in a PTV for

example.

Yet, the grid technique presents numerous potential advantages that

cannot be overlooked, especially in terms of normal tissue sparing.

Consequently, the interest in grid therapy has tremendously grown in the

past few years. However, most research in the field still focus on delivering

very heterogeneous doses to the targeted volume, sometimes with setups that

would require important additional efforts to be clinically achievable.

In this work a different approach to grid therapy was therefore

proposed. Delivering doses to the PTV that are similar to conventional

treatments in terms of coverage to achieve similar TCP is a key point to help

bringing grid therapy closer to the clinic and change the way it is regarded

by clinicians. To have a chance of application, grid therapy needs to be

developed to the point where it would be seen as a solid and safe alternative

to conventional modalities, with undeniable advantages.

This thesis aimed to do just that: develop another form of grid therapy

to get closer to a realistic approach that could be accepted in the clinic. This

was achieved by exploring the use of interlaced proton grids. The interlacing

permits homogeneous irradiation of a specific volume, in the case of an

ideal, or close to ideal, interlacing. Also, while it would have been attractive

to use micrometer beams, it was shown that very interesting results could be

achieved using usual beam sizes, without unsurmountable technical

difficulties. Smaller beam sizes, and how to produce them, were also studied,

to take advantage of the discussed dose-volume effect.

54

But, this thesis should also be taken as it really is: the first steps of an

unconventional modality’s development, almost from square one. In that

sense, while some aspects like physics and dosimetry have been partially

covered or discussed, the amount of remaining unknown variables is large.

Radiobiological data for different beam sizes, different c-t-c distances, with

different valley doses, different peak doses, different tissue types and more

are still very much needed to understand the limits of such technique, and

what would be a safe but efficient application. Some technical difficulties

also need to be addressed and, as it was seen with the RBE considerations,

going further into the development of this interlaced proton grid therapy

usually brings additional issues and questions to be answered.

In addition, it should not be forgotten that the majority of the results

presented in this work were produced using tools such as Monte Carlo codes

and treatment planning systems. These tools are based on models developed

to quickly mimic and approximate realistic interactions/observations. While

they provide valuable results in first-approach studies as it was conducted

here, one could argue about their validity for more sensitive applications.

One evident example would be the production of interlaced proton grid plans

intended for actual treatments. To provide the best grid plans possible extra

optimization objectives might be needed on top of regular PTV dose

coverage and OARs dose constraints: the minimization of the valley dose

and of the VPDR in healthy tissues. However, they directly rely on the

correct modeling of the beam’s lateral scattering in a patient volume, and are

therefore dependent on the model used in the TPS. In that case, are the

current models sufficient for this application or does interlaced proton grid

therapy require additional models? This will have to be evaluated as well.

But, one should not be afraid of the difficulties to come, as this

innovative modality has the potential, on paper, to improve the quality of

certain treatments compared to what can be done currently. Its concept was

proven, and now needs to be further developed. This has to be done on a

scientific level of course, but spreading the word on this research area and

making it widely available is just as important to insure a growing interest

and a potential increased brain-power dedicated to this topic.

Chapter 8. Summary of papers

55

8. Summary of papers

In Paper I, a study of what could be currently done with available beam sizes

at a modern proton center facility was conducted. Two patients previously

treated with photon therapy at Karolinska University Hospital for a liver and

a rectum irradiation were replanned using a commercial proton treatment

planning system with the suggested grid therapy method. Doses inside and

outside the target were studied as a proof-of-concept study, to evaluate if

grid therapy could be technically feasible today, using beams in the

centimeter range currently employed at The Skandion Clinic for example.

Data for one of the two patients, using more conventional radiation therapy

techniques, can be found in Paper II for comparison. In this paper, 10

patients with liver metastasis treated with conventional photon therapy were

replanned using broad-beam proton therapy with the same dose objectives.

In Paper III, the impact of spot position uncertainty on the quality of the

interlaced proton grid plans was assessed. A robustness analysis was

performed in a treatment planning system where numerous proton grid plans

were produced assuming an imperfect spot positioning, with realistic

variations. The effect on target dose coverage and organs at risk constraints

fulfilments were then assessed.

In Paper IV, a Monte Carlo study was performed to obtain dosimetric

characteristics of small proton beamlets in the 0.5-3 millimeters (FWHM)

range and determine which beam width could potentially be the most

advantageous for grid therapy. The doses produced by such beamlets were

then used to produce a virtual grid irradiation were a cubic target at the

center of a water tank was irradiated with interlaced proton grids, using

different grid parameters.

In Paper V, a comparison between interlaced proton grid and IMPT plans

was performed, taking into account a fixed RBE of 1.1 and two variable

RBE models. Dose, LET and RBE-weighted dose distributions were

evaluated, as well as dose-volume histograms.

In Paper VI, a study of the organ doses induced by a proton-gantry based

CBCT system was achieved. The CBCT system was modelled using two

56

Monte Carlo codes. Those two models were then used to evaluate doses

delivered to an anthropomorphic phantom using different scanning protocols

(Head/Thorax/Pelvis, Full/Short scan). Doses delivered in case of

single/weekly or daily imaging were also calculated.

Acknowledgments

57

Acknowledgments

First and foremost, I would like to thank my supervisors, Irena Gudowska

and Iuliana Toma-Dasu for all their help, scientific input, guidance and

support over the years. Your words and advices have really helped me to

always look forward in my projects and become a better scientist. Thank you

also for giving me the opportunity to teach and supervise research projects.

These experiences made me realize that my will to teach and help others was

stronger than I first imagined.

To all my co-authors: thank you for your help, feedback and all the

interesting discussions. Albert Siegbahn is acknowledged for initiating this

project. I would like to extend special thanks to Ana Ureba, Jakob Öden,

Oscar Ardenfors and Niels Bassler. To Ana, thanks for all your very

valuable help and all the discussions that we have had. You made this

project grow with your advices and your never-ending joy and motivation. It

was always a pleasure to have a talk with you, be it scientific or not. To

Jakob, thanks for accepting to get on board of the grid train with me. I am

really proud of the two papers we have achieved together and I would not

have been able to do it without you. I hope there will be me more. Thanks

for being a great office mate and the best “Vindicated” singer in the world.

You are always welcome for an apéro in France with Oscar. To Oscar,

world record holder for fastest accepted publication, thanks for your help

and the discussions from the very beginning. I am proud of what we have

accomplished with our paper, and hopefully there will be more to come.

With all the problems that we encountered, all the sweat, I learned a lot.

Finally playing “Simply the best” was definitely rewarding. Thank you also

to you and Jakob for the help and support you gave me during the less

enjoyable moments of these four years. To Niels, here’s to all the small

discussions that we have had and that taught me a lot. You definitely made

me a better researcher, and I hope that we will be able to continue working

together in the future.

To my mentor Bo Nilsson: thank you for your help and guidance. Our

discussions always helped me go forward.

To Emely “Mormor” Kjellsson Lindblom: thank you for everything, the

laughter therapy, all the help, support, advices and all the things that you did

58

for me and more. I do not know what the outcome of all this would have

been if you had not been there for me. It was always reassuring to know I

had someone I could rely on, no matter the problem or the time. Thanks

again for everything, and I hope you will make good use of all the extra

years of lifetime you deservedly earned!

To all my friends at MSF: Ana, Emely, Gracinda, Hamza, Helena, Jakob,

Marta, Oscar & Toshiro - thanks for all the discussions, laughs, help, good

vibes, dinners, nights outs and all the traveling around the world, karaokes

and more. These four years as a PhD student have been amazing thanks to all

of you.

To Mona and Mariann: thanks for all the administrative and overall help

for a lot of things!

To all my friends in Sweden, France, all over Europe, Canada, USA, Japan

and everybody I had the chance to meet through conferences, workshops,

projects, seminars, study visits and more: thanks for all the good memories.

All of you have become very dear friends. We met through science all over

the world but you have become a lot more than coworkers or colleagues to

me.

Finally, to my family: thank you for the unconditional support you have

given me throughout the years, no matter where my studies took me –

another city, the USA or Sweden. Although I would not have expected 10

years ago to achieve a PhD, you have made it possible.

References

59

References

Asur, R., Butterworth, K.T., Penagaricano, J.A., Prise, K.M., Griffin, R.J., 2015.

High dose bystander effects in spatially fractionated radiation therapy. Cancer

Lett.

Bijl, H.P., Van Luijk, P., Coppes, R.P., Schippers, J.M., Konings, A.W.T., Van der

Kogel, A.J., 2002. Dose-volume effects in the rat cervical spinal cord after

proton irradiation. Int. J. Radiat. Oncol. Biol. Phys. 52, 205–211.

Bijl, H.P., Van Luijk, P., Coppes, R.P., Schippers, J.M., Konings, A.W.T., Van Der

Kogel, A.J., 2003. Unexpected changes of rat cervical spinal cord tolerance

caused by inhomogeneous dose distributions. Int. J. Radiat. Oncol. Biol. Phys.

57, 274–281.

Bijl, H.P., Van Luijk, P., Coppes, R.P., Schippers, J.M., Konings, A.W.T., Van Der

Kogel, A.J., 2006. Influence of adjacent low-dose fields on tolerance to high

doses of protons in rat cervical spinal cord. Int. J. Radiat. Oncol. Biol. Phys.

64, 1204–1210.

Blattmann, H., Gebbers, J.O., Bräuer-Krisch, E., Bravin, A., Le Duc, G., Burkard,

W., Di Michiel, M., Djonov, V., Slatkin, D.N., Stepanek, J., Laissue, J.A.,

2005. Applications of synchrotron X-rays to radiotherapy. In: Nuclear

Instruments and Methods in Physics Research, Section A: Accelerators,

Spectrometers, Detectors and Associated Equipment. pp. 17–22.

Bouchet, A., Lemasson, B., Christen, T., Potez, M., Rome, C., Coquery, N., Le

Clec’H, C., Moisan, A., Bräuer-Krisch, E., Leduc, G., Rémy, C., Laissue, J.A.,

Barbier, E.L., Brun, E., Serduc, R., 2013. Synchrotron microbeam radiation

therapy induces hypoxia in intracerebral gliosarcoma but not in the normal

brain. Radiother. Oncol. 108, 143–148.

Bouchet, A., Lemasson, B., Le Duc, G., Maisin, C., Bräuer-Krisch, E., Siegbahn,

E.A., Renaud, L., Khalil, E., Rémy, C., Poillot, C., Bravin, A., Laissue, J.A.,

Barbier, E.L., Serduc, R., 2010. Preferential effect of synchrotron microbeam

radiation therapy on intracerebral 9l gliosarcoma vascular networks. Int. J.

Radiat. Oncol. Biol. Phys. 78, 1503–1512.

Bräuer-Krisch, E., Adam, J.F., Alagoz, E., Bartzsch, S., Crosbie, J., DeWagter, C.,

Dipuglia, A., Donzelli, M., Doran, S., Fournier, P., Kalef-Ezra, J., Kock, A.,

Lerch, M., McErlean, C., Oelfke, U., Olko, P., Petasecca, M., Povoli, M.,

Rosenfeld, A., Siegbahn, E.A., Sporea, D., Stugu, B., 2015. Medical physics

aspects of the synchrotron radiation therapies: Microbeam radiation therapy

(MRT) and synchrotron stereotactic radiotherapy (SSRT). Phys. Medica 31,

568–583.

Bräuer-Krisch, E., Requardt, H., Régnard, P., Corde, S., Siegbahn, E., LeDuc, G.,

Brochard, T., Blattmann, H., Laissue, J., Bravin, A., 2005. New irradiation

geometry for microbeam radiation therapy. Phys. Med. Biol. 50, 3103–3111.

60

Bräuer-Krisch, E., Serduc, R., Siegbahn, E.A., Le Duc, G., Prezado, Y., Bravin, A.,

Blattmann, H., Laissue, J.A., 2010. Effects of pulsed, spatially fractionated,

microscopic synchrotron X-ray beams on normal and tumoral brain tissue.

Mutat. Res. Mutat. Res. 704, 160–166.

Crosbie, J.C., Anderson, R.L., Rothkamm, K., Restall, C.M., Cann, L., Ruwanpura,

S., Meachem, S., Yagi, N., Svalbe, I., Lewis, R.A., Williams, B.R.G., Rogers,

P.A.W., 2010. Tumor Cell Response to Synchrotron Microbeam Radiation

Therapy Differs Markedly From Cells in Normal Tissues. Int. J. Radiat.

Oncol. Biol. Phys. 77, 886–894.

Dilmanian, F.A., Button, T.M., Le Duc, G., Zhong, N., Peña, L.A., Smith, J.A.L.,

Martinez, S.R., Bacarian, T., Tammam, J., Ren, B., Farmer, P.M., Kalef-Ezra,

J., Micca, P.L., Nawrocky, M.M., Niederer, J.A., Recksiek, F.P., Fuchs, A.,

Rosen, E.M., 2002. Response of rat intracranial 9L gliosarcoma to microbeam

radiation therapy. Neuro. Oncol. 4, 26–38.

Dilmanian, F.A., Eley, J.G., Krishnan, S., 2015. Minibeam therapy with protons and

light ions: physical feasibility and potential to reduce radiation side effects and

to facilitate hypofractionation. Int. J. Radiat. Oncol. Biol. Phys. 92, 469–474.

Dilmanian, F.A., Morris, G.M., Zhong, N., Bacarian, T., Hainfeld, J.F., Kalef-Ezra,

J., Brewington, L.J., Tammam, J., Rosen, E.M., 2003. Murine EMT-6

carcinoma: high therapeutic efficacy of microbeam radiation therapy. Radiat.

Res. 159, 632–641.

Dilmanian, F.A., Qu, Y., Feinendegen, L.E., Peña, L.A., Bacarian, T., Henn, F.A.,

Kalef-Ezra, J., Liu, S., Zhong, Z., McDonald, J.W., 2007. Tissue-sparing

effect of x-ray microplanar beams particularly in the CNS: Is a bystander

effect involved? Exp. Hematol. 35, 69–77.

Dilmanian, F.A., Rusek, A., Fois, G.R., Olschowka, J., Desnoyers, N.R., Park, J.Y.,

Dioszegi, I., Dane, B., Wang, R., Tomasi, D., Lee, H., Hurley, S.D., Coyle,

P.K., Meek, A.G., O’Banion, M.K., 2012. Interleaved carbon minibeams: An

experimental radiosurgery method with clinical potential. Int. J. Radiat. Oncol.

Biol. Phys. 84, 514–519.

Dilmanian, F.A., Zhong, Z., Bacarian, T., Benveniste, H., Romanelli, P., Wang, R.,

Welwart, J., Yuasa, T., Rosen, E.M., Anschel, D.J., 2006. Interlaced x-ray

microplanar beams: a radiosurgery approach with clinical potential. Proc. Natl.

Acad. Sci. U. S. A. 103, 9709–9714.

Durante, M., Paganetti, H., 2016. Nuclear physics in particle therapy: A review.

Reports Prog. Phys. 79, 096702.

Emami, B., 2013. Tolerance of Normal Tissue to Therapeutic Radiation. Reports

Radiother. Oncol. 1, 36–48.

Enzmann, D.R., Pelc, N.J., 1992. Brain motion: measurement with phase-contrast

MR imaging. Radiology 185, 653–660.

References

61

Gao, M., Mohiuddin, M.M., Hartsell, W.F., Pankuch, M., 2018. Spatially

fractionated (GRID) radiation therapy using proton pencil beam scanning

(PBS): Feasibility study and clinical implementation. Med. Phys. 45, 1645–

1653.

Goorley, T., James, M., Booth, T., Brown, F., Bull, J., Cox, L.J., Durkee, J., Elson,

J., Fensin, M., Forster, R.A., Hendricks, J., Hughes, H.G., Johns, R.,

Kiedrowski, B., Martz, R., Mashnik, S., McKinney, G., Pelowitz, D., Prael,

R., Sweezy, J., Waters, L., Wilcox, T., Zukaitis, T., 2016. Features of MCNP6.

Ann. Nucl. Energy 87, 772–783.

Grotzer, M.A., Schültke, E., Bräuer-Krisch, E., Laissue, J.A., 2015. Microbeam

radiation therapy: Clinical perspectives. Phys. Medica 31, 564–567.

Haring, W., 1934. Siebstrahlung. Strahlentherapie 51, 154–163.

Heck, B., Jess-Hempen, A., Kreiner, H.J., Schöpgens, H., Mack, A., 2007. Accuracy

and stability of positioning in radiosurgery: Long term results of the Gamma

Knife system. Med. Phys. 34, 1487–1495.

Hopewell, J.W., Coggle, J.E., Wells, J., Hamlet, R., Williams, J.P., Charles, M.W.,

1986. The acute effects of different energy beta-emitters on pig and mouse

skin. Br. J. Radiol. Suppl. 19, 47–51.

Hopewell, J.W., Morris, A.D., Dixon-Brown, A., 1987. The influence of field size

on the late tolerance of the rat spinal cord to single doses of X rays. Br. J.

Radiol. 60, 1099–1108.

Hopewell, J.W., Trott, K.R., 2000. Volume effects in radiobiology as applied to

radiotherapy. Radiother. Oncol. 56, 283–288.

Huhn, J.L., Regine, W.F., Valentino, J.P., Meigooni, A.S., Kudrimoti, M.,

Mohiuddin, M., 2006. Spatially Fractionated GRID Radiation Treatment of

Advanced Neck Disease Associated with Head and Neck Cancer. Technol.

Cancer Res. Treat. 5, 607–612.

Jan, S., Santin, G., Strul, D., Staelens, S., Assié, K., Autret, D., Avner, S., Barbier,

R., Bardiès, M., Bloomfield, P.M., Brasse, D., Breton, V., Bruyndonckx, P.,

Buvat, I., Chatziioannou, A.F., Choi, Y., Chung, Y.H., Comtat, C.,

Donnarieix, D., Ferrer, L., Glick, S.J., Groiselle, C.J., Guez, D., Honore, P.-F.,

Kerhoas-Cavata, S., Kirov, A.S., Kohli, V., Koole, M., Krieguer, M., Laan,

D.J. van der, Lamare, F., Largeron, G., Lartizien, C., Lazaro, D., Maas, M.C.,

Maigne, L., Mayet, F., Melot, F., Merheb, C., Pennacchio, E., Perez, J.,

Pietrzyk, U., Rannou, F.R., Rey, M., Schaart, D.R., Schmidtlein, C.R., Simon,

L., Song, T.Y., Vieira, J.-M., Visvikis, D., Walle, R. Van de, Wieërs, E.,

Morel, C., 2004. GATE: a simulation toolkit for PET and SPECT. Phys. Med.

Biol. 49, 4543–4561.

Jolles, B., 1953. X-ray sieve therapy in cancer : a connective tissue problem. Boston:

Little, Brown and Company.

62

Kaiser, A., Mohiuddin, M.M., Jackson, G.L., 2013. Dramatic response from

neoadjuvant, spatially fractionated GRID radiotherapy (SFGRT) for large,

high-grade extremity sarcoma. J. Radiat. Oncol. 2, 103–106.

Kłodowska, M., Olko, P., Waligórski, M.P.R., 2015. Proton microbeam

radiotherapy with scanned pencil-beams - Monte Carlo simulations. Phys.

Medica 31, 621–626.

Köhler, A., 1909. Une nouvelle méthode permettant de faire agir, dans la profondeur

des tissus, de hautes doses de rayons Roentgen et un moyen nouveau de

protection contre les radiodermites. Ann. d’Electrobiologie Radiol. 10, 661–

664.

Laissue, J.A., Bartzsch, S., Blattmann, H., Bräuer-Krisch, E., Bravin, A., Dalléry,

D., Djonov, V., Hanson, A.L., Hopewell, J.W., Kaser-Hotz, B., Keyriläinen,

J., Laissue, P.P., Miura, M., Serduc, R., Siegbahn, A.E., Slatkin, D.N., 2013.

Response of the rat spinal cord to X-ray microbeams. Radiother. Oncol. 106,

106–111.

Laissue, J.A., Blattmann, H., Slatkin, D.N., 2012. Alban Köhler (1874-1947):

Erfinder der Gittertherapie. Z. Med. Phys. 22, 90–99.

Laissue, J.A., Geiser, G., Spanne, P.O., Dilmanian, F.A., Gebbers, J.O., Geiser, M.,

Wu, X.Y., Makar, M.S., Mica, P.L., Nawrocky, M.M., Joel, D.D., Slatkin,

D.N., 1998. Neuropathology of ablation of rat gliosarcomas and contiguous

brain tissues using a microplanar beam of synchrotron-wiggler-generated X

rays. Int. J. Cancer 78, 654–660.

Lawrence, Y.R., Allen Li, X., el Naqa, I., Hahn, C.A., Marks, L.B., Merchant, T.E.,

Dicker, A.P., 2010. Radiation dose-volume effects in the brain. Int J Radiat

Oncol Biol Phys 76, 20–27.

Lee, E., Meyer, J., Sandison, G., 2016. Collimator design for spatially-fractionated

proton beams for radiobiology research. Phys. Med. Biol. 61, 5378–5389.

Liberson, F., 1933. The Value of a Multi-perforated Screen in Deep X-ray Therapy.

Radiology 20, 186–195.

Lomax, A.J., Schaer, M., 2012. 322 Intensity Modulated ‘Grid’ Proton Therapy.

Trying To Exploit ‘Spatial Fractionation’ With Protons. Radiother. Oncol.

102, S171–S173.

Mackonis, E.C., Suchowerska, N., Zhang, M., Ebert, M., McKenzie, D.R., Jackson,

M., 2007. Cellular response to modulated radiation fields. Phys. Med. Biol.

52, 5469–5482.

Marks, H., 1950. A new approach to the roentgen therapy of cancer with the use of a

grid. J. Mt. Sinai Hosp. N. Y. 17, 46–48.

Marks, L.B., Yorke, E.D., Jackson, A., Ten Haken, R.K., Constine, L.S., Eisbruch,

A., Bentzen, S.M., Nam, J., Deasy, J.O., 2010. Use of Normal Tissue

Complication Probability Models in the Clinic. Int. J. Radiat. Oncol. 76, S10–

S19.

References

63

McNamara, A.L., Schuemann, J., Paganetti, H., 2015. A phenomenological relative

biological effectiveness (RBE) model for proton therapy based on all

published in vitro cell survival data. Phys. Med. Biol. 60, 8399–8416.

Miura, M., Blattmann, H., Bräuer-Krisch, E., Bravin, A., Hanson, A.L., Nawrocky,

M.M., Micca, P.L., Slatkin, D.N., Laissue, J.A., 2006. Radiosurgical palliation

of aggressive murine SCCVII squamous cell carcinomas using synchrotron-

generated X-ray microbeams. Br. J. Radiol. 79, 71–75.

Mohiuddin, M., Curtis, D.L., Grizos, W.T., Komarnicky, L., 1990. Palliative

treatment of advanced cancer using multiple nonconfluent pencil beam

radiation: A pilot study. Cancer 66, 114–118.

Mohiuddin, M., Fujita, M., Regine, W.F., Megooni, A.S., Ibbott, G.S., Ahmed,

M.M., 1999. High-dose spatially-fractionated radiation (GRID): A new

paradigm in the management of advanced cancers. Int. J. Radiat. Oncol. Biol.

Phys. 45, 721–727.

Mohiuddin, M., Park, H., Hallmeyer, S., Richards, J., 2015. High-Dose Radiation as

a Dramatic, Immunological Primer in Locally Advanced Melanoma. Cureus 7,

e417.

Muirhead, R., Adams, R., Gilbert, D., Harrison, M., Glynne-Jones, R., Sebag-

Montefiore, D., Hawkins, M., 2016. National guidance for IMRT in anal

cancer.

Neuner, G., Mohiuddin, M.M., Vander Walde, N., Goloubeva, O., Ha, J., Yu, C.X.,

Regine, W.F., 2012. High-dose Spatially Fractionated GRID Radiation

Therapy (SFGRT): A comparison of treatment outcomes with Cerrobend vs.

MLC SFGRT. Int. J. Radiat. Oncol. Biol. Phys. 82, 1642–1649.

Ödén, J., Eriksson, K., Toma-Dasu, I., 2017a. Inclusion of a variable RBE into

proton and photon plan comparison for various fractionation schedules in

prostate radiation therapy. Med. Phys. 44, 810–822.

Ödén, J., Toma-Dasu, I., Eriksson, K., Flejmer, A.M., Dasu, A., 2017b. The

influence of breathing motion and a variable relative biological effectiveness

in proton therapy of left-sided breast cancer. Acta Oncol. (Madr). 56, 1428–

1436.

Paganetti, H., 2014. Relative biological effectiveness (RBE) values for proton beam

therapy. Variations as a function of biological endpoint, dose, and linear

energy transfer. Phys. Med. Biol. 59, R419–R472.

Peñagarícano, J.A., Moros, E.G., Ratanatharathorn, V., Yan, Y., Corry, P., 2010.

Evaluation of Spatially Fractionated Radiotherapy (GRID) and Definitive

Chemoradiotherapy With Curative Intent for Locally Advanced Squamous

Cell Carcinoma of the Head and Neck: Initial Response Rates and Toxicity.

Int. J. Radiat. Oncol. Biol. Phys. 76, 1369–1375.

Perl, J., Shin, J., Schümann, J., Faddegon, B., Paganetti, H., 2012. TOPAS: An

innovative proton Monte Carlo platform for research and clinical applications.

Med. Phys. 39, 6818–6837.

64

Peucelle, C., Nauraye, C., Patriarca, A., Hierso, E., Fournier-Bidoz, N., Mart�nez-

Rovira, I., Prezado, Y., 2015. Proton minibeam radiation therapy:

Experimental dosimetry evaluation. Med. Phys. 42, 7108–7113.

Powers, B.E., Thames, H.D., Gillette, S.M., Smith, C., Beck, E.R., Gillette, E.L.,

1998. Volume effects in the irradiated canine spinal cord: Do they exist when

the probability of injury is low? Radiother. Oncol. 46, 297–306.

Prezado, Y., Fois, G.R., 2013. Proton-minibeam radiation therapy: A proof of

concept. Med. Phys. 40, 031712.

Prezado, Y., Jouvion, G., Hardy, D., Patriarca, A., Nauraye, C., Bergs, J., González,

W., Guardiola, C., Juchaux, M., Labiod, D., Dendale, R., Jourdain, L., Sebrie,

C., Pouzoulet, F., 2017. Proton minibeam radiation therapy spares normal rat

brain: Long-Term Clinical, Radiological and Histopathological Analysis. Sci.

Rep. 7, 14403.

Schultheiss, T.E., Stephens, L.C., Ang, K.K., Price, R.E., Peters, L.J., 1994. Volume

effects in rhesus monkey spinal cord. Int. J. Radiat. Oncol. Biol. Phys. 29, 67–

72.

Schültke, E., Balosso, J., Breslin, T., Cavaletti, G., Djonov, V., Esteve, F., Grotzer,

M., Hildebrandt, G., Valdman, A., Laissue, J., 2017. Microbeam radiation

therapy — grid therapy and beyond: a clinical perspective. Br. J. Radiol. 90,

20170073.

Serduc, R., Bouchet, A., Bräuer-Krisch, E., Laissue, J.A., Spiga, J., Sarun, S.,

Bravin, A., Fonta, C., Renaud, L., Boutonnat, J., Siegbahn, E.A., Estève, F.,

Le Duc, G., 2009. Synchrotron microbeam radiation therapy for rat brain

tumor palliation - Influence of the microbeam width at constant valley dose.

Phys. Med. Biol. 54, 6711–6724.

Serduc, R., Bräuer-Krisch, E., Siegbahn, E.A., Bouchet, A., Pouyatos, B., Carron,

R., Pannetier, N., Renaud, L., Berruyer, G., Nemoz, C., Brochard, T., Rémy,

C., Barbier, E.L., Bravin, A., Le Duc, G., Depaulis, A., Estève, F., Laissue,

J.A., 2010. High-precision radiosurgical dose delivery by interlaced

microbeam arrays of high-flux low-energy synchrotron x-rays. PLoS One 5,

e9028.

Serduc, R., Van De Looij, Y., Francony, G., Verdonck, O., Van Der Sanden, B.,

Laissue, J., Farion, R., Bräuer-Krisch, E., Siegbahn, E.A., Bravin, A., Prezado,

Y., Segebarth, C., Rémy, C., Lahrech, H., 2008. Characterization and

quantification of cerebral edema induced by synchrotron x-ray microbeam

radiation therapy. Phys. Med. Biol. 53, 1153–1166.

Shirato, H., Mizuta, M., Miyasaka, K., 1995. A mathematical model of the volume

effect which postulates cell migration from unirradiated tissues. Radiother.

Oncol. 35, 227–231.

Slatkin, D.N., Spanne, P., Dilmanian, F.A., Sandbora, M., 1992. Microbeam

radiation therapy. Med. Phys. 19, 1395–1400.

References

65

Smilowitz, H.M., Blattmann, H., Bräuer-Krisch, E., Bravin, A., Di Michiel, M.,

Gebbers, J.O., Hanson, A.L., Lyubimova, N., Slatkin, D.N., Stepanek, J.,

Laissue, J.A., 2006. Synergy of gene-mediated immunoprophylaxis and

microbeam radiation therapy for advanced intracerebral rat 9L gliosarcomas.

J. Neurooncol. 78, 135–143.

van den Aardweg, G.J.M.J., Hopewell, J.W., Whitehouse, E.M., 1995. The radiation

response of the cervical spinal cord of the pig: Effects of changing the

irradiated volume. Int. J. Radiat. Oncol. Biol. Phys. 31, 51–55.

van der Kogel, A.J., 1993. Dose-volume effects in the spinal cord. Radiother. Oncol.

29, 105–109.

Van Der Sanden, B., Bräuer-Krisch, E., Siegbahn, E.A., Ricard, C., Vial, J.C.,

Laissue, J., 2010. Tolerance of arteries to microplanar X-ray beams. Int. J.

Radiat. Oncol. Biol. Phys. 77, 1545–1552.

Van Luijk, P., Bijl, H.P., Konings, A.W.T., Van der Kogel, A.J., Schippers, J.M.,

2005. Data on dose-volume effects in the rat spinal cord do not support

existing NTCP models. Int. J. Radiat. Oncol. Biol. Phys. 61, 892–900.

Wedenberg, M., Lind, B.K., Hårdemark, B., 2013. A model for the relative

biological effectiveness of protons: The tissue specific parameter α/β of

photons is a predictor for the sensitivity to LET changes. Acta Oncol. (Madr).

52, 580–588.

Withers, H.R., Taylor, J.M.G., Maciejewski, B., 1988. Treatment volume and tissue

tolerance. Int. J. Radiat. Oncol. Biol. Phys. 14, 751–759.

Zlobinskaya, O., Girst, S., Greubel, C., Hable, V., Siebenwirth, C., Walsh, D.W.M.,

Multhoff, G., Wilkens, J.J., Schmid, T.E., Dollinger, G., 2013. Reduced side

effects by proton microchannel radiotherapy: Study in a human skin model.

Radiat. Environ. Biophys. 52, 123–133.

Zwicker, R.D., Meigooni, A., Mohiuddin, M., 2004. Therapeutic advantage of grid

irradiation for large single fractions. Int. J. Radiat. Oncol. Biol. Phys. 58,

1309–1315.