Master Thesis Report

(1)

Master Thesis Report

The impact of metallic cranial implants on proton-beam

radiotherapy treatment plans for near implant located tumours

A phantom study on the physical effects and agreement between simulated treatment plans and the resulting treatment for near implant located cranial tumours

Adam Sjögren

June 21, 2018

Student Spring 2018

(2)

testing

Supervisor: Ulf Granlund (ulf.granlund@regionorebrolan.se), Department of medical physics, Örebro University Hospital

Examiner: Jonna Wilen (jonna.wilen@umu.se), Department of Radiation Sciences, Umeå University

Adam Sjögren (sjogren.adam@outlook.com):The impact of metallic cranial implants on proton- beam radiotherapy treatment plans for near implant located tumours, 2018.

Master’s Thesis in Engineering Physics & Medical Physics, 30.0 ECTS Examensarbete för civilingenjörsexamen i teknisk fysik, 30.0 hp

(3)

Abstract

Within the field of radiotherapy treatments of tumour diseases, the hunt for more accurate and effective treatment methods is a continuous process. For some years ion-beam based radiotherapy, especially the proton-beam based applications, has increased in popularity and availability.

The main reason behind this is the fact that ion-beam based applications make it possible to modulate the dose after the planning target volume (PTV) defined by the radiation oncologist.

This means that it becomes possible to spare tissue in another way, which might result in more effective treatments, especially in the vicinity of radio sensitive organs. Ion-beam based treatments are however more sensitive to uncertainties in PTV position and beam range as ion-beams have a fixed range depending on target media and initial energy, as opposed to the conventional x-ray beams that do not really have a defined range. Instead their intensity decreases exponentially at a rate dependent of the initial energy and target media. Therefore density heterogeneities result in uncertainties in the planned treatments. As the plans normally are created using a CT- images, for which metallic implants can yield increased heterogeneities both from the implants themselves and so called metal artifacts (distortions in the images caused by different processes as the X-rays used in image acquisition goes through metals). Metallic implants affects the accuracy of a treatment, and therefore also the related risks, so it is important to have an idea of the magnitude of the impact. Therefore the aim of this study is to estimate the impact on a proton-beam based treatment plan for six cranial implants. These were one Ti-mesh implant, one temporal plate implant, one burr-hole cover implant and three craniofix implants of different sizes, which all are commonly seen at the Skandion clinic. Also the ability of the treatment planning system (TPS), used at the clinic, to simulate the effects on the plans caused by the implants is to be studied. From this result it should be estimated if the margins and practices in place at the clinic, for when it is required to aim the beam through the implant, are sufficient or if they should be changed.

This study consisted of one test on the range shift effects and one test on the lateral dose distribution changes, with one preparational test in the form of a calibration of Gafchromic EBT3 films. The range shift test was performed on three of the implants, excluding the three craniofix implants using a water phantom and a treatment plan created to represent a standard treatment in the cranial area. The lateral dose distribution change test was performed as a solid phantom study using radiochromic film, for two treatment plans (one where the PTV was located 2 cm below surface, for all implants, and one where it was located at the surface, only for the Ti-mesh and the temporal plate). The results of both tests were compared to simulations performed in the Eclipse treatment planing system (TPS) available at Skandion.

The result of the range shift test showed a maximum range shift of−1.03(1) mm, for the burr-hole cover implant, and as the related Eclipse simulations showed a maximal shift of −0.17(1) mm there was a clear problem with the simulation. However, this might not be because of the TPS but due to errors in the CT-image reconstruction, such as, for example, geometrical errors in the representation of the implants. As the margin applied for a similar situation at the Skandion clinic (in order to correct for several uncertainty factors) is 4.2 mm there might be a need to increase this margin depending on the situation.

For the lateral distribution effects no definite results were found as the change varied in magnitude, even if it tended to manifest as a decreasing dose for the first plan and a increasing dose for the second. It was therefore concluded that further studies are needed before anything clear can be said.

iii

(4)

Sammanfattning på svenska

Inom cancervården går utvecklingen konstant framåt och sökandet efter nya, effektivare metoder inom området extern radioterapi är inget undantag. Inom extern radioterapi använder man sig utav joniserande strålar (strålar som genom att överföra energi till molekyler joniserar dem vilket bryter isär molekylen) som riktas genom en patient och koncentreras i ett lokalt target vid tumören. Genom att deposera tillräckligt mycket dos (ett mått av energi per massa) till tumören hoppas man, inom radioterapi döda samtliga cancerceller och på så vis lyckas behandla cancern, men man måste samtidigt vara försiktig med att inte deposera för mycket dos i frisk vävnad.

Det konventionella valet vid extern radioterapi är att använda sig av fotonstrålar inom röntgen- spektrat, med våglängder mellan 10 pm och 1 nm, men de senaste åren har metoder baserade på laddade partiklar, främst protoner, blivit alltmer tillgängliga. Den stora skillnaden mellan foton och partikel-baserade strålar är hur de deposerar dos. Medans fotonstrålar tränger igenom hela patientens kropp och deposerar avtagande mängd dos på väger tränger laddade partiklar bara in till ett specifikt djup, beroende av dess initiella energi, och deposerar majoriteten av dosen på slutet. Detta innebär att metoder baserade på strålar av laddade partiklar gör det möjligt att modulera behandlingen så att mer frisk vävnad blir helt skonad från deposerad dos, vilket gör dessa metoder extra lämpade för behandling av cancer i närheten av strålkänslig vävnad. Detta betende med ett specifikt djup gör samtidigt dessa metoder betydligt känsligare för variationer i patientens densitet, och framför allt om metalliska implantat finns i närheten av tumören. För att planera en behandling använder man sig av ett CT-underlag (en slags 3D röntgen-bild) och metall resulterar i vad man kallar för metallartefakter (en typ av störningar), dessa resulterar i att patientens anatomi, och metallens storlek, missvisas. Detta medför att det finns en risk att en del av tumören inte får tillräckligt mycket dos för att behandlingen skall lyckas eller att frisk närliggande vävnad får skadligt hög dos. Det är därför viktigt att ta hänsyn till detta då man är tvungen att rikta strålarna genom metall.

I denna studien har man kollat på hur 6 titan-baserade kraniella implantat påverkar protonbe- handlings planer vid Skandion-kliniken i Uppsala, den enda kliniken i sverige där protonbaserad radioterapi är möjlig. Detta har utförts genom en fantom-studie (studie där man använt olika material för att simulera en behandling) där man undersökt hur protonernas räckvidd påverkats av att implantaten placerats i fältet. Samt hur dosfördelningen i ett plan på olika djup påverkats för en plan där target låg på ett djup av 2 cm och en där target låg i ytan. Resultaten har även jämförts med simuleringar utförda i klinikens planerings programm för att studera hur väl de fysiska förändringarna har simulerats vid planeringen, vilket ger en indikation om huruvida marginalerna som redan läggs till vid kliniken är tillräckliga eller om de behöver ses över.

Resultaten visade att det var betydande skillnader mellan simuleringarna utav räckviddens påverkan och de fysiska mätningarna då simuleringarna till hög grad underskattade effekten.

Det noterades dock att detta kan ha grund hur väl CT-underlagen representerade verkligheten och att man bör studera detta närmare innan slutgiltig slutsats kan fastslås, men det kan, för vissa fall, vara aktuellt att se över marginalerna som appliceras vid kliniken. När det gällde dosfördelningen i de olika planen visade en dosminsknings trend för planen där target låg lite djupare och en dosöknings trend på den ytliga planen, men det gick inte att bestämma nån klar storlek på grund av varierande magnitud för de olika djupen och implantaten, så det fastslogs att vidare studier krävs innan slutgiltig slutsats kan fastslås.

iv

(5)

Acknowledgements

First I would like to direct the sincerest gratitude towards my supervisor, Ulf Granlund, at Uni- versitetssjukhuset Örebro who introduced me to the field and has done a great work guiding me on the path of this work.

A continued thank you to the personnel at the Skandion clinic, especially Karin Andersson, Christina Vallhagen Dahlgren and Liliana Stolarczyk for their great support with the practical work during the tests and for letting me take part of their wast expertise in the field.

A great deal of gratitude is directed towards Michael Gubanski and Petra Witt Nyström who supplied the study with the implants, without which this work would have been impossible to perform.

Thank you to the rest of the personnel at the department for medical physics at Universitetssjukhuset Örebro for taking me in and for the support. You all made this time great.

Further on I would like to thank my parents and grand parents, both paternal and maternal, for the different levels of support provided, while not understanding most of what I was talking about. And also thanks to my brothers for supporting me ”in their own special way”.

Lastly I wold like to thank my fellow students and friends both near and far who lightened up the days when the pressure of the studies grew a bit heavy. And that from time to time put some light in my days by, for example, trying to force me into saying something fun.

v

(6)

Nomenclature and abbreviations

A short description of some of the words and abbreviations used in this thesis.

CT Computed Tomography, a method used to creating 3D-

images using a x-ray setup that rotates around the object being imaged, usually a patient.

Radiotherapy A term used to group medical treatments in which ionizing radiation is used (not diagnostics).

Gy Gray, a unit used to describe the amount of energy absorbed in a tissue due to radiation, defined as one joule per kilogram of tissue.

DSB Double Strand Break, describes a damage to a DNA-

molecule where both strands of the DNA is damaged in a localized area, often resulting in a full break of the DNA- molecule. As it is difficult for the cell to repair a DSB it often results in the death of the cell.

CTV Clinical Target Volume, a volume deined within the patient, by a radiation oncologist, in which the tumour and possible microscopic cancerous spread is located.

PTV Planning Target Volume, defined by the radiation oncolo-

gist as the CTV + an extra margin in order to correct for uncertainties in tumour position due to, for example, errors in day to day setup and organ movements. The PTV is used during treatment planning in order to make sure that the tumour absorbs a high enough dose to guarantee successful treatment.

LET Linear Energy transfer, the local transfer of energy per unit track length traveled by a particle within a media.

BP Bragg peak, the distinctive peak which is found at the end of a Bragg curve, which in turn is the characteristic percentage depth-dose curve for protons and larger charged particles.

SOBP Spread Out Bragg Peak, a plateau received through the mod- ulation of particle energy resulting in a plateau of multiple Bragg peaks at different depths, making it possible to modulate the deposited dose after the PTV.

TPS Treatment Planning System, a software used to plan, and

simulate, a radiotherapeutic treatment.

MARS Metal Artifact Reduction Software, a type of software used to reduce image distortions in a CT image resulting from the presence of metals in the depicted patient.

iMAR Iterative Metal Artifact Reduction, the MARS supplied by Siemens Healthineers for their products.

vi

(7)

Nomenclature and abbreviations

WET Water Equivalent Thickness, the thickness of water that

has the same properties as a slab of a certain thickness and material from a radio-physical point of view.

Mass stopping power The amount of energy lost per unit length travelled by a certain particle with a certain initial energy through a certain material with a certain density due to interactions with matter.

R50 Range 50 %, the range from the surface into a target to

which a minimum of 50 % of the maximum dose has been absorbed. A value often used when comparing different treatment plans.

Gafchromic EBT3 A type of radiochromic film (radiation activated self developing film used as a radiation detector).

Plane-parallel ionization chamber

A radiation detector consisting of two conductive plates that measures dose through the collection and measurement of the charge created through ionization of air between the two plates within a waterproof shell.

vii

(8)

Contents

1 Introduction

The fact that cancer is a terrible diagnose for anyone to receive is widely known, but most people probably only have a minimal knowledge about how different forms of cancer are treated. Some of the most common treatments are the different forms of radiotherapy (the use of different forms of ionizing radiation) in which the external group (using beams of either photons or charged particles) are the most common one. The idea behind all radiotherapy can be described in the sentence ”The deliverance of the right dose to the right target volume”. This means that it is important that the right amount of energy is delivered to the cancer and as little as possible to the healthy tissue surrounding it as the radiation itself can result in severe side effects, such as radiation-induced cancer later in life or the loss of function in the irradiated organs. This means that it is highly important that the accuracy of the treatment is sufficient as the target is to be given a high enough dose to kill all the cancerous cells in order for the treatment to be successful.

The conventional method for external treatments are the use of photon-beams, where beams of photons are sent through the body with exponentially decreasing energy-deposition. Using several beams with different angles through the body the dose deposited can be modelled so that the target receives the prescribed dose and the healthy tissue a lower. Charged particle beam methods, like using proton beams, work a bit different as there are clear differences in the way that photons and charged particles deliver energy. Charged particles instead deliver a lower, rather constant, dose as it travels through the tissue, until it reaches a certain range, depending on its initial energy, where almost all is deposited to a single point. This means that while photons deliver dose through the whole body charged particles only deliver dose to a certain depth. The treatment plan is designed so that the initial energies of the charged particles results in that the target area receives the prescribed dose. There are, however, some increased risks with the use of charged particle beams though. The fact that these methods only deliver dose to a certain depth means that the treatments are sensitive to shifts in the range of the charged particles that can come from several sources. Generally a denser material allows for a shorter range as it is dependent on both the initial energy of the particle and the material properties. So if the body of the patient is not modelled sufficiently well by the CT image used in the planning of the treatment, the result might be that parts of the target is given too low dose and the cancer survive or the healthy tissue can be given doses that are so high that the severe side effects earlier mentioned might occur. This is a well known problem when the target is in the close vicinity of metallic implants as metals creates so called artifacts in the CT-images that misrepresents the size of the implants and the anatomical build of the patient.

There are different ways to tackle these problems and treatment-planning systems are designed to take it into consideration to some extent. But the effects of different implants and how the treatment plans mirrors the real outcome of the treatment are not studied sufficiently enough, so that is what this study looks into.

1.1 Purpose

The purpose of this study is to investigate how six of the cranial implants more commonly seen at the Skandion clinic in Uppsala, Sweden, effects proton-beam radiotherapy treatments of near implant located tumour diseases. The ability of the software, available at the clinic, to simulate these effects will also be studied in order to estimate if the margins and practices already at place are sufficient, or if they need to be changed for situations were it is required to aim the beams through the implants for some reason.

1.2 Disposition

In Sec. 2 and 3 the relevant theory and information about the equipment used in the study is presented.

Within Sec. 4 the methods of the study are described. The results are presented in Sec. 5 with additional

(11)

1. Introduction

information that can help with the understanding of the results presented within appendices A to D. In Sec. 6 a discussion regarding the results is presented as a base for the conclusion presented in Sec. 7.

(12)

2. Background and theory

2 Background and theory

2.1 Radiotherapy

Radiotherapy is a word used when speaking about medical treatments using radiation and the term includes several different methods used, both externally and internally. In internal radiotherapy methods the radiation is delivered by a source that is inserted into the patients body. This might be in the form of liquids, used in nuclear-medicine therapy applications, or solid sources, used in brachy-therapy. The external therapy methods are methods were beams of either photons or charged particles are used from the outside. The external methods are more common and the methods are constantly evolving.

2.1.1 The basics of radiotherapy

The aim with all forms of radiotherapy is to deposit the right dose to the right volume and in medical physics three quantities of dose are defined, absorbed dose, equivalent dose and effective dose. In radiotherapy one usually only use the absorbed dose, which is defined as

D = d ¯E

dm (1)

where d ¯E is the mean energy deposited to matter of mass dm [1] due to ionizing radiation. Absorbed dose is given in the unit gray (written as Gy) and one gray is defined as one joule per kilogram of tissue.

The main process searched after in successful radiotherapy treatments is the creation of so called double- strand breaks (DSB) which is a damage on the DNA sequence due to radiation where both strands are broken at a close distance due to the ionization of DNA-molecules by ionizing radiation [2]. DSB damage is hard for the cells to repair and the result is that the cells can not go through the mitosis cycle and eventually the cells dies. So the aim of radiotherapy treatments is to deposit enough dose to the so called clinical target volume (CTV). A CTV is defined as a volume containing the cancerous tumour with a margin containing possible microscopic cancer spread [3]. If the CTV does not receive enough dose some cancerous cells might survive, leading to a failed treatment as the cancer might return from just one cell.

In order to account for uncertainties in CTV position, for example due to daily setup uncertainty and organ movements, another volume is defined called planning target volume (PTV). The PTV is a volume containing the CTV and with a margin depending on the location of the tumour and the surrounding anatomy [3].

However, radiotherapy can also result in severe side effects when healthy tissues receives enough dose, for example radiation induced cancer up to several years later in life and decreased function of irradiated organs. This means that it is highly important that the treatment is given with a high enough accuracy so that the PTV gets enough dose and the surrounding, healthy tissue is spared as much as possible.

So one of the most important aspects in radiotherapy is the accuracy. This can be solved in several different ways. In internal radiotherapy this has been done by taking the radiation source to the tumour.

For nuclear-medicine radioactive isotopes of different substances that are naturally absorbed by the sick tissues are used. In brachy-therapy one inserts solid metallic sources in the direct vicinity of the tumour.

Through this only the area of the tumour absorbs dose. In external therapy beams of photons or charged particles are used and here the problem is solved through the use of several beam directions resulting in that beams are sent through the patient with different directions all aimed towards the PTV. This results in that the tissues on the beams way to and from the PTV will receive some dose but lower as the dose from the different beams are collected in the PTV, so when more beams are used a lower dose is absorbed by healthy tissue.

(13)

2.1.2 Photon- versus particle-beams

The conventional choice of beam type in external radiotherapy is photon beams, but particle based applications are increasing in popularity. Photon- and particle-beams have significantly different properties due to the nature of the processes through which they interact with matter, which results in that they deposit energy in different ways.

Photon-beams used in radiotherapy are in the X-ray range of the electromagnetic spectra and thus can travel through the human body. The beams intensity decreases exponentially as the distance traveled in a certain media increases. This can be seen in a percentage depth-dose curve, as in Fig. 1, which depict the amount of dose delivered by a certain ionizing beam at a certain energy as it goes through a specified media. However, this decrease in dose only takes place after a initial increase in the small build-up region, due to initial increase of photon-density as a result of scattering in the media. The dose from photon-beams are mainly deposited through the absorption of some or all of the energy of a single photon by an atom.

For charged particles the processes are quite different as the dose is instead deposited through different charge and particle interaction. Larger charged particles interact with more particles per unit distance traveled and also the charge and initial energy have a great impact on the dose deposition. For charged particles the concept of linear energy transfer (LET), defined as the local transfer of energy per unit track length, is applied. That is the average collision energy loss dE up to a maximum energy loss ∆E for a distance dl, defined as

LET = (_dE

dl )

∆E

. (2)

Basically, the lower the kinetic energy of a particle is the higher the probability of interaction is. This means that the LET increases with decreasing kinetic energy of the particle. Following this the percentage depth-dose curves of particle beams usually takes a rather distinct shape, especially for the heavy charged particles. For these heavy charged particles (protons and all charged particles larger than that) the percentage depth-dose curve is in the shape of a so called Bragg curve with the distinctive Bragg peak (BP). The distinctive shape of the BP is, as explained by Frank Herbert Attix in Introduction to Radiological Physics and Radiation Dosimetry [4], a consequence of the dependence of the kinetic energy for a heavy particle at lower energies when it comes to the probability of interactions. A heavy particle will deposit half its initial energy while traveling a distance x and the remaining half will then be deposited while traveling only a distance ∼= ^x₃. a typical example of this can be seen in Fig. 1. This means that a lower amount of dose is deposited until a certain, predictable, depth is reached where the majority of the dose is deposited. The BP behaviour of larger charged particles is a large part of the reason to why particle beams are increasing in popularity. The fact that dose is deposited on the way towards the target and in the target itself, but practically no dose after the target make it possible to spare more healthy tissue using charged particles instead of photons which deposit dose even after the target [5]. This is helpful in cases where radiation sensitive organs (clinically called organs at risk or OAR) are close to the PTV. Also, since the position range of the particles, and through that the position of the BP, are dependent of the initial energy of the particle-beam it is possible to shape a so called spread out Bragg peak (SOBP) by varying the initial energy, see Fig. 1. The use of a SOBP makes it possible to creat a plateau of a certain dose that can be modeled to cover the PTV, making it possible to have deposit a therapy dose to the PTV and only a relatively low dose to the healthy tissue on the way in and practically no dose to the rest of the tissue, except for a small amount of possibly scattered protons.

(14)

Figure 1 – A simple presentation of the depth-dose behaviour of photons and protons with the Bragg peak of a single-energy proton beam clearly visible at the end of the SOBP-plateau. Figure heavily based upon an image found in the publication Proton Beam Therapy for Non-Small Cell Lung Cancer: Current Clinical Evidence and Future Directions by Abigail T. Berman et al [6].

2.2 The problem with metallic implants in radiotherapy

The first part of any radiotherapy treatment is the creation of the treatment plans which are based on either a MR- (magnetic resonance) or, much more commonly, CT-scan (computed tomography) of the patient. This means that the quality of the treatment is highly dependent on the quality of the image- basis used in planning as the anatomical structure of the patient is used to calculate the deposited dose to different volumes. If the planing basis does not represent the patients anatomy well enough, the plan will contain larger uncertainties that needs to be corrected for.

It is well known that metallic objects are a common source of so called artifacts (distortions in the image) in CT- and MR-scans, but in this work focus lies upon the CT-based therapy. The reason behind the problems with metals is the fact that medias of higher atomic number are ”denser” when it comes to radiation and therefore the X-rays used in CT can not pass through metals as easily, and what builds up the CT-image is the amount of radiation that passes through the media to the detectors on the opposite side. The so artifacts resulting from metals are commonly called metal streak artifacts and are often seen as clear black and white streaks beaming out from the metallic structure, sometimes making it impossible to distinguish any anatomical structures. The intensity of a tissue in a CT-image is directly correlated to its radiation density and it is used later by the treatment planning system (TPS) to calculate and simulate the treatment plan, so it is clear that these effects have a great impact on the treatment and need to be accounted for. Different metal artifact reducing software’s are developed by the companies that produce CT-scanners, such as O-MAR (Metal Artifact Reduction for Orthopedic implants) from Philips [7] or iMAR (Iterative Metal Artifact Reduction) from Siemens Healthineers [8], that have been seen to improve the image quality (for most but not all cases). An example of the metal artifact streak problem can be seen in Fig. 2.

(15)

(a) (b)

Figure 2 – Images demonstrating the problems with metallic streak artifacts. In (a) the metallic streak artifacts resulting from amalgam dental repairs, as can be seen it is hard to distinguish any anatomical data of the jaw and teeth area. In (b) the same image is shown after being enhanced by a metal artifact reducing software. Images from University hospital of Örebro.

Even if metal artifact reducing softwares can increase the quality of a CT-image to a great degree there will still be uncertainties in the image that will need to be accounted for. There will be a certain uncertainty in the density of reconstructed tissue data and some smaller structures might be missing fully so the treatment plans might differ from the resulting treatment at a significant level. This means that the margin used to correct the plan for uncertainties from other sources such as intestinal movements and change of tumour size throughout the treatment might need to be increased depending on what kind of implant is located in the area. Also it is required to know how large the effects are in order to decide if it is possible to aim the beam through the implant or if it should be avoided.

There is however much more to be studied when it comes to the impact of metallic implants on treatment plans. While several studies have been performed regarding the impact of metallic objects on treatments, such as the publications by Isabelle Dietlicher et al[9], Yincui Jia et al [10] and Joost M. Verburg et al [11], only a few studies have been performed regarding the cranial area. One of those is the study The effects of titanium mesh on passive-scattering proton dose of Haibo Lin et al [5] where the impact of a 0.6 mm titanium mesh implant, with a metal-to-hole ratio of 0.85, on a plan in a certain patient case was studied. For the studied case their group saw dosimetric impacts below 1 % and a range reduction of less than 0.5 mm, however they noted that an similar analysis might bee needed for each individual case. As there are several other kinds of metallic cranial implants available, with varying thickness and structure this study will look upon some implants that are commonly encountered at the Skandion clinic in Uppsala, Sweden, in order to determine how treatment plans should be made when those implants are found in the vicinity of treatment areas.

The impact of an implant on a treatment plan is dependent of the material and structure of the implant.

It is critical that the implants surface is constructed by a biomedical material, which is a material that is wall suited functionally for the task and at the same time is as compatible as possible with human tissue. For many of the newer metallic implants titanium and titanium alloy based materials are the

(16)

material of choice due to the high bio-compatibility of those materials, and the fact that they produce relatively less artifacts, especially for smaller implants. Many of the cranial implants need to be able to protect the brain from impacts as they are often used to either used to replace pieces of cranial bone removed during neurosurgery or to hold those pieces in place after surgery. This means that metal is a well suited choice for implants due to their rigidity.

2.3 The connection between water and tissue from a radiophysical viewpoint

The human body mostly consists of water and about 50 %-60 % of an adults total weight is water, adding to that is the fact that most of the human organs consist of between 70 %-80 % water (notable exceptions to this is bone, 20 %, and fat, 10 %) [12]. This means that most of the human body can be approximated to water in most radiophysical situations making it easier to perform measurements and simulations using water instead of real tissues. This also makes it possible to produce sufficiently accurate calculations using tabulated values without the need to consider anatomical differences between individuals and different tissues.

It should however be noted that no tissue is exactly water equivalent. Which for most situations is not a critical problem, but in some cases there is a need to take the effects into consideration. In some cases it is impossible to perform certain measurements with regular water due to several possible reasons, for example when using solid phantoms or measurement equipment that can be damaged by water. In order to take this into consideration most phantom materials have a so called WET value, which stands for water equivalent thickness and describes the ratio between the thickness of the material and the thickness of a piece of liquid water that has the same properties in a radio-physical viewpoint. WET is defined (by Rui Zhang and Wayne D Newhauser [13]) as

W ET = tw≈ tm

ρmS¯m

ρwS¯w

, (3)

where t is the thickness, rho is the mass density and ¯S is the mean proton mass stopping power and the subindexes m and w denotes if the value is related to the relevant material (m) or water (w). ¯S in turn is defined as

S =¯

´

ESdE

´

EdE , (4)

where E is the energy of a proton and S is the proton mass stopping power related to a certain energy.

2.4 Range shift and R

₅₀

Due to the fact that a proton beam of a certain energy have a fixed range depending on the target medias density any heterogeneities introduced to the media will affect the range of the beam. This effect is simply called range shift and is for obvious reasons significant when it comes to proton beam based radiotherapy as it, at a great magnitude, affects the accuracy of the treatment, so it is important to correct for these effects in order to maximize the treatment outcome.

In the publication A semi-empirical model for the therapeutic range shift estimation caused by inhomo- geneities in proton beam therapy by Moskvin et al [14] a simple model is presented that makes it possible to calculate an estimation of the maximum range shift introduced by a object in the common proton beam energy range for radiotherapy purposes. This model, defined as

(17)

∆(tM, Z) = tM[ρM(1.192− 0.158ln(Z)) − 1] (5) calculates the water equivalent range shift ∆x coming from an object of thickness t_M and density ρ_M with atomic number Z.

In dosimetry radiotherapy it is often helpful to talk about well defined entities and as the total range often is hard to define the term RN is used, which is defined as the range into a target volume that receives at least N % of the maximum dose. In this study this range is defined as the range in the target that absorbs a dose equal to at least half the maximum absorbed dose, which is R50.

2.5 The Skandion clinic

The Skandion clinic in Uppsala, Sweden, is the first clinic dedicated towards proton-beam radiotherapy in Scandinavia. The clinic is a national project driven together by the 7 counties in Sweden that contains university classed hospitals, through komunalförbundet avancerad strålbehandling. The main equipment at the location is the model Proteus PLUS cyclotron (IBA worldwide) which supplies the clinic with proton beams of up to 230 MeV. In each treatment room the beam is distributed via a specialized nozzle which is moved around the patient by a so called gantry, which is not visible from the room itself, see Fig. 3. At this moment there are two active treatment rooms, with the space to install a third, and a room containing a nozzle without a gantry used purely for research.

The clinic is driven using the principle of distributed competence, the medical responsibility throughout the whole treatment is held by the patients home hospital and only the actual treatment is performed at Skandion. The patient is examined and prepared at the hospital where all fixation equipment and treatment plans are prepared. Only after all is prepared and a treatment plan has been approved by both the hospital and the group at Skandion the patient is sent to Skandion to receive treatment. After the treatment is done the follow up work and controls are performed in place at the hospital. From this follows that the results of this study is relevant to all hospitals connected to the Skandion clinc as the work with the plans and preparations is where one need to take the effects of the implants into consideration.

Figure 3 – One of the treatment rooms at The Skandion clinic with the nozzle clearly visible over the patient coach, the gantry is separated from the room via the wall but the size of it can be imagined due to the fact that the cylindrical area is within it.

(18)

3. Material and software used in the study

3 Material and software used in the study

Within this section the materials used within this study are presented.

3.1 The implants

What follows here is a presentation of the six implants used within the study.

3.1.1 Titanium mesh implants (implant 1 and 2)

In this study two types of titanium mesh implants were used, one 9 cm× 9 cm Ti-mesh implant (called implant 1 in this study) with a thickness of 0.3 mm and one temporal plate implant (called implant 2 in this study) with a thickness of 0.4 mm, both shown in Fig. 4 and 5. Titanium mesh implants are usually used to replace pieces of bone removed during brain surgery and are screwed in place over the hole in the cranium using small, self drilling, titanium screws, which means that it often might be hard to avoid beaming through the implant. However, as was seen in the work of Lin et al the effect of a 0.6 mm titanium mesh implant was quite small so it is expected that the effect of the thinner mesh is lower than that as the implant used in this study is only half as thick as the one used in their.

Figure 4 – The titanium mesh implant (implant 1) used in the tests

Figure 5 – The titanium temporal plate implant (implant 2), with a piece of tape used to mark position directors.

3.1.2 Burr-hole cover (implant 3)

The burr-hole cover implant used in this study (called implant 3) was a titanium based circular plate with a diameter of 17 mm and a thickness of 0.5 mm, see Fig. 6, which as the name implies is used to cover a burr-hole made in the cranial bone. The implant is fixated onto the cranium using small self drilling titanium screws of the same kind as the ones used for the titanium mesh implants, where the smaller peripheral rings on the implant is used as anchor points for the screws. There are also some models with a different structure specially designed to allow for a shunt to be inserted without the need for additional holes, but that type of implant will not be used in this study.

(19)

Figure 6 – The Burr-hole cover implant (implant 3), used in the study. In the upper right corner the kind of Ti screws, used to attach the burr-hole cover and Ti-mesh implants to the cranial bone, can be seen.

3.1.3 Craniofix (implant 4-6)

Craniofix implants are constructed as two parallel Ti-based plates connected in the center by a ribbed rod. They are used to re-fixate the pieces of cranial bone that are removed during surgery. One of the plates are positioned between the cranial bone and the brain and the pieces are fixated using a special tool that forces the other plate down over the ridged rod until the bone pieces are tightly fixated between the plates, the ridges on the rod then makes impossible for the titanium plates to separate unless significant amounts of force is applied, and therefore the rest of the rod can be cut of after fixation. In this study craniofix implants of three different sizes, one with a plate diameter of 20 mm (called implant 4 in this study), one with a plate diameter of 16 mm (called implant 5 in this study) and one with a plate diameter of 12 mm (called implant 6 in this study) were used, all shown in Fig. 7.

Figure 7 – The three different craniofix implants used, with plate diameter 20 mm (implant 4), 16 mm (implant 5) and 12 mm (implant 6)

(20)

3.2 Radiochromic film

The detector of choice for the measurements of the lateral measurements in this study is radiochromic films, a type of detector that has decreased in popularity due to the ever expanding availability of digital detectors. Radiochromic films however are still common within studies as they make it possible to store and re asses measurements at a later time, increasing the reproducibility of a study. For this study the Gafchromic EBT3 type film (Ashland inc.), were used and in this section some information about films in general and the EBT3 type itself will be presented.

3.2.1 General information regarding radiochromic films

One problem with radiochromic films when it comes to proton dosimetry is that they are known to be affected by a dose under-response in the presence of low energy protons, for example in the BP area, which generally needs to be accounted for. One of the possible explanations for this error is the local saturation of radiochromic films when exposed to radiation. Upon irradiation of radiochromic films the polymers in the active volume of the films are excited which results in polymerization reaction effectively changing the colour of the active volume of the film. When low energy protons goes through the active volume the high LET of the protons results in that the number of radiation activated polymerization sites are higher than for a proton of higher energy. The limited amount of polymerization sites in an area means that once all sites are activated there will be no more response to the radiation and an under-response is guaranteed.

Any available correction of the under-response requires defined and appropriate correction values related to a corresponding LET value. But as written in Investigation of EBT2 and EBT3 films for proton dosimetry in the 4-20 MeV energy range by S. Reinhardt et al [15] there are no available tabulated LET values in the vicinity of the BP region for protons. This due to the fact that the LET distribution increases when an ion penetrate into matter as the primary particle spread, due to scattering reactions, with increasing penetration depth, and then the secondary charged particles (additional ions formed in ionization due to the radiation) needs to be accounted for. To add to the problem secondary charged particles can have significantly greater LET values than the primary charged particles. This means that specially adapted softwares, normally Monte-Carlo based, have to be used in order to produce a more exact correction factor in order to handle the under-response. Although, depending on the situation it is possible to bypass these problems by not measuring in the regions where BPs are dominant, such as in the distal part of a SOBP. So as it was considered that it would take to much time to get accustomed to a Monte-Carlo software it was decided, after several discussions that we should try to measure no deeper than 50% of the SOBP.

One of the special things regarding the radiochromic films is the need to calibrate every single batch of films separately due to differences between batches in production, that means that any calibration is only valid for that batch alone. This calibration is performed by an irradiation of pieces of films with increasing, well known, dose. As the colour change is directly related to the absorbed dose it is possible to then create a calibration curve for that batch using the pixel values received after the films are scanned as the pixel values will vary depending on the decreasing transmission (of the light used to scan the film, due to the colour change) resulting in a relation between the absorbed dose and the transmission degree. Radiochromic films are self developing films, which means that there is no need for dark-rooms and chemicals in the developing process. The films just need time to stabilize, as the polymerization process continues for some time after irradiation, so one should wait at least 24 hours after irradiation before scanning the film. Some smaller changes might occur even after that but after about 48 hours the changes can be disregarded, and the film can be stored for a exceptionally long time, making it possible to archive measurement results. If the films after this time are stored correctly however no significant polymerization will occur and the films can be re scanned any time later on to verify results.

All radiochrimic films are sensitive, to a varying degree, to environmental factors, such as UV-light and temperature differences, thus the films should always be stored away from UV-light as much as possible

(21)

and as close to the temperature it is going to be used in as it is possible.

3.2.2 The Gafchromic EBT3 radiochromic film

The Gafchromic EBT3 type of radiochromic film is one of the most widely used today. It is well fitted for quality dosimetry even if the handling is more time consuming and require more work than more conventional digital equipment. The composition of the EBT3 type films was thoroughly described in the publication Under-response correction for EBT3 films in the presence of proton spread out Bragg peaks by Fiorini et al [16], from which the compositional data in Tab. 1 is collected. As can be seen in Fig. 8 the active volume of the film is layered between two layers of polyethylene terephtalate, commonly known as PET, plastic. The active volume contain a yellow dye, which includes lithium pentacosa-10,12- diynoate, or LiPCDA, molecules which upon radiation induced polymerization turns blue, which can be seen in Fig. 9. The outer coating layer contains silica particles in order to counteract so called Newton’s rings (a kind of image quality degrading artifacts formed during scanning). Due to the structure of the films, seen in Fig. 8, they should be handled with care as it is quite easy to separate the active layer from the PET layers through for example creasing, which will make the films useless.

Figure 8 – The cross-sectional structure of the EBT3, image based on the information in the work of Fiorini et al [16]

Figure 9 – Two pieces of EBT3 film used in the calibration that clearly shows the colour change due to radiation induced polymerization, the yellow on the left is a unirradiated reference piece showing the colour of the film and the blue on the right has received a dose of 1.9978 Gy

Table 1 – The chemical composition of the layers of the films, given as percentage of atoms. Data from the work of Fiorini et al [16]

EBT3 H Li C N O Na S Cl K Br

Active layer 58.2 0.8 29.2 0.1 10.7 0.1 0.1 0.9 - -

PET 36.4 - 45.5 - 18.2 - - - - -

PET + SiO2 PET = 99.986 SiO₂ = 0.014

The dose range of EBT3 films are depending on the colour channel used during scanning, the most commonly used is the red channel, which has a dose range up to 10 Gy, due to the fact that the film’s absorption is at maximum for the red area of the light spectra. Using the green channel will give a dose range exceeding 40 Gy, but for most dosimetric purposes this is way above the desired range, also the sensitivity would be to low for this study where a range up to about 2.5 Gy is desired.

(22)

3.3 Blue Phantom

²

One commonly used phantom type when it comes to measurements of depth doses or dose profiles in 1D is water phantoms which basically is a box of perspex that is filled with regular water. It also contains a mechanical rig to which a ionization chamber detector (usually of the so called plane-parallel type) can be attached, making it possible to move it with great precision within the water using a computer.

For the experiments regarding the range effects in this study the Blue Phantom² type water phantom (IBA Dosimetry) available at Skandion were used. The choice of detector was the ROOS type 34001 plane-parallel ionization chamber, which consists of two plates that collects the charge created by ions formed in air between the plates during interaction with the protons, all contained within a waterproof shell. A part of the phantom (as it was used in the tests) can be seen in Fig. 10.

Figure 10 – A part of the Blue Phantom² system as it was used within the tests. Centrally located at the front of the phantom are plates of a bone equivalent polymer (described in the next section). The ROOS chamber can be partially seen behind the upper right corner of the bone equivalent plates, also visible is the laser lights of the laser guidance system of the gantry used in the tests.

3.4 Solid phantom material

For the lateral profile measurements, where the radiochromic films were to be used, it was decided that solid phantoms would be used. In order to represent the cranial bone two plates of bone equivalent

(23)

phantom material, constructed of a polymer made by CIRS (Computerized Imaging Reference Systems, incorporated) to represent cortical bone with a WET value of 2.48 was acquired (seen in Fig. 11). In order to simulate the tissues in the brain, which mostly is brain tissue, the same plates of solid water (GAMMEX, Middleton, Wisconsin, USA) with a WET value of 1.024, seen in Fig. 12, usually used to simulate human tissue during QC (quality controls) of treatment plans at Skandion were used.

Figure 11 – The bone equivalent plates used in the study with visible markings added to be used as alignment guides during the tests.

Figure 12 – Solid water blocks of differ- ent sizes making it possible to modulate different depths.

3.5 Other equipment and software used

The type of scanner used to scan the films in this work was a Epson Expression 10000 XL flatbed scanner. The TPS of choice was the Eclipse treatment planning system (Varian Medical Systems) and much of the analytic work was performed in MATLAB R2017b (Mathworks) environment. Some of the simulations were to be made using a CT-image base that had been enhanced using a metal artifact reduction software and as the CT at Skandion was a SOMATOM Definition AS Open - RT Pro edition (Siemens Healthineers) the software of choice was iMAR.

(24)

4. Experimental methods

4 Experimental methods

In this study 3 different test will be performed, the first being the calibration of the EBT3 radiochromic films the second a test of the impact on the proton range and the third the effects on the lateral dose distribution. Before anything else was done, sets of CT scans were performed on the different setups in order to be used as a ground for the calculations, simulations and planning performed in the TPS.

4.1 Calibration of the EBT3 films

The calibration of each batch of film is a critically important step in the use of radiochromic film for dosimetry as any errors in this stage will move on to the measurements performed using the particular batch. The method used in the calibration in this study is based upon the method used by Krzempek et al in the publication Calibration of Gafchromic EBT3 film for dosimetry of scanning proton pencil beam (PBS) [17], however the method used is only based on their method and not exactly the same.

First a simplified ”treatment plan” consisting of a box shaped SOBP with a size of 10 cm× 10 cm × 10 cm located at a depth of 4 cm. 1 sheet of the film was cut into 16 5 cm× 5 cm pieces. These were then, one by one, placed in the centre of the SOBP, just on top of a ROOS type 34001 plane-parallel ionization chamber used to measure the exact dose, and irradiated with a dose increasing in 11 steps from 0 Gy to approximately 2.4 Gy. 4 additional films given doses within the range to be used as control films after the calibration has been performed. In Fig. 13 a representation of the calibration setup can be seen.

Figure 13 – A representation of the setup for the calibration of the EBT3 films. Note that the image isn’t proportional to the real case, as it would be impossible to distinguish certain details if that would have been the case. The red area represents the solid water phantom material. the black rectangle the ROOS type ionization chamber and the yellow area represents the film pieces.

The blue square represents the volume of the target to which the ”treatment plan” used was planned, which basically is the resulting SOBP.

After irradiation the films were left to stabilize for approximately 36 hours before being scanned with a Epson Expression 10000 XL scanner using the 48 bit channel with a resolution of 75 dpi. The scanned images were saved as TIFF-images and exported into an MATLAB program which performed the calibration. The program localized the corners of the films and cropped the images down to only contain

(25)

the film. It then again cropped the images about 1 cm extra from the edges as the data in that area can not be guaranteed to be valid due to damages inflicted on the film when cutting it and scattering effects due to the edges themselves. Then the program separated the RGB parts of the image as the calibration is performed on the separate channels one at a time. The pixel values, about 30 pixels in from the edges (to avoid edge effects distorting the data), of the red channel where then normalized. The mean value for each of the film pieces were plotted against the known doses , measured with the ROOS chamber, as- sociated with each individual film piece. Then a calibration curve was fitted to the measured points using

D(P Vk) = Ak· P Vk+ Bk

P Vk+ Ck

(6)

where D is the dose, P Vk is the normalized pixel value for channel k and Ak, Bk and Ck are fitting parameters related to channel k.

After the calibration curve had been found, the data from 4 extra control film pieces were put into the program in order to validate the curve. As a secondary control every single pixel value of the film that had been irradiated up to a dose close to what was expected to be relevant in the later measurement were recalculated, using the calibration, into dose values, resulting in a dose measurements. From the resulting dose image three crossline (horizontal) and three inline (vertical) linear dose-profiles were measured from which the maximum pixel to pixel variation was checked. If the variation was deemed to be to within reasonable limit, or if the profile still was deemed to fit well with the dose measured with the ROOS chamber, even if the variation was deemed a bit high, the calibration was deemed to be valid and applicable to the later tests.

4.2 Range effects

For the two remaining tests a new treatment plan more similar to what could be suspected to be seen in a real clinical case was needed, so a target volume of 8× 8 × 10cm³ were defined at a depth of 2 cm in a reference CT. To this a plan was produced to give 1.64 Gy to the whole target, and due to the restrictions from the cyclotron when it comes to the lowest available proton energies a so called range-shifter (a plate which decreases the travel length of protons with a certain distance) had to be applied. This plan could then be copied and simulated with other CT:s as basis, which makes it possible to see how the implants effects the same reference plan. Due to the size and structure of the different implants this test was only performable on implants 1-3 (however only the part of the test where the bone equivalent plates were applied due to the need to fixate it with a screw and the nature of how the implant usually is used).

In order to measure the range of the protons the Blue Phantom system was used. The patient bed were exchanged for the phantom base and the phantom was filled with water up to a certain level, which made it possible to do the measurements without any significant risk for unwanted effects due to interactions with the water surface. The ROOS chamber were placed in the holder and connected to the measurement system. The bone equivalent plates where taped onto the side of the phantom where the protons were to enter and the laser guidance light was used to make sure that the plates and the ROOS chamber were located in the center of the beam. The plan was then imported to the treatment station and run several times with the ROOS chamber located at different depths within the phantom until it would be possible to fit a curve of the protons depth dose with a few extra measurement points around the depth were 50 % of the maximum charge (directly related towards the absorbed dose) were collected.

This depth is called R50(which stands for Range 50 %) and is what is to be used in the comparison.

(26)

(a) (b)

Figure 14 – A representation of the setup used for the range effects test, using the Blue Phantom² system. (a) shows the setup which included the bone equivalent plates (seen as a grey box) and (b) the setup which excluded them. Again the black box represents the ROOS type ionization chamber, here placed in the mechanical structure so that it could be moved in the distal part of the SOBP (represented by the red rectangle).

The same procedure was applied to both the reference run (without implants) and runs where each of the implants (which this test could be performed on) were placed in the center of the beam on top of the bone equivalent plates. After the last measurements the data was put into MATLAB and a 5:th order polynomial curve was fitted to each of the series which made it possible to more correctly find R50 for each series and from that the relative range shift, compared to the reference, for each of the implants were calculated.

Then the bone equivalent plates where removed and all steps redone without them, except without implant 3, due to the fact that titanium mesh implants are used to replace bone and therefore there often is not any bone between implant and brain tissue in clinical cases. So implant 1 and implant 2 were tested both with and without the bone equivalent plates in order to get an idea of how both cases are affected.

As a control of the physical validity of the results of the physical measurements Eq. (5) was used to calculate the maximum water equivalent range shift possible for the implants (if they had been just solid pieces of Ti with the relevant thickness), which made it possible to estimate the quality of the measurements.

In order to check how well the Eclipse TPS simulates the range effects linear depth dose profiles were exported from the simulations and the relative range shifts were calculated and the results compared to the results found through the Blue Phantom measurements in order to estimate the accuracy of the simulations.

4.3 Lateral dose distribution

For this test the same plan was used as for the range tests but as this test involved the use of EBT3 films the procedure differed at some degree. First several sheets of film were cut into four pieces approximately 125 mm× 103 mm large. The phantom used in this test was build up with a base of two 50 mm thick slabs of solid water. On top of those a film piece were placed in the center of the beam path, followed by layers of solid water, with varying thickness in the order of 20 mm, 20 mm, 5 mm, 3 mm, 2 mm, 2 mm and finally the bone equivalent plates, with a film piece centered in the beam path between all layers. Also for the reference run and for implant 1 and implant 2 an eighth film piece was placed on top of the bone

(27)

equivalent plates in order to measure the effects directly below the implant, which can be seen in Fig. 15.

Figure 15 – A view of the setup for the lateral dose distribution tests with the bone equivalent plates and film pieces for the reference measurements visible on top of a stack solid water plates.

In Fig. 17 a more detailed schematic of the setup can be seen, however note that this is a image that is not proportional to the real situation and is just meant to give an idea of the setup.

(a)

(b)

Figure 16 – A representation of the setup used for the lateral dose distribution tests, using the Blue Phantom² system. (a) shows the setup which included the bone equivalent plates (seen as a grey box) and (b) the setup which excluded them. Again the yellow areas represents the film pieces and the red solid water plate material. The targets used to produce the SOBPs are represented by the blue rectangles. Note that no measurements are performed no deeper than 50% in the SOBP.

(28)

The treatment plan were then loaded into the treatment station and run first for the reference structure and then for every implant one by one. The films where then left alone to stabilize for approximately 30 hours before being scanned and exported to MATLAB where a program was used to first prepare the RGB TIFF-images in a way similar to the one used during calibration before converting the RGB images into DICOM images showing dose measurements directly. Here it was important to take day to day differences of the scanner into consideration so film pieces 1, 6, 9 and 15 were scanned again and the mean value of the drift between the films original pixel-values were used as a correction factor that was applied to the pixel values of the new films before the conversion into a dose image.

Then, using the simulations in Eclipse lateral dose profiles of the simulations were exported from the depths at which each film were located. These dose profiles, saved as images in DICOM-format, where later put into a program in MATLAB that cropped the corresponding Eclipse image and film image into the same size. From these images linear dose profiles in both crossline and inline direction were extracted and used in the following comparisons.

In order to figure out the physical impact the dose profile of each measurement were compared to the related measurement from the reference setup, which gave and idea of how the dose were affected at the different depths. A comparison were also made regarding the mean dose of the same measurements.

As it was also desired to see if the Eclipse TPS system simulated the effects in a satisfying manner the linear profiles from the measurements in the implant setups were also compared to corresponding profiles from the Eclipse simulations.

As it has been seen that it is somewhat common that Ti-mesh implants are in contact with the PTV it was requested to perform an additional test on implant 1 and implant 2 for a setup where the bone equivalent plates were removed and the PTV placed in contact with the implant. So a new plan were made and all the previous steps repeated, but with a new setup were the layers of solid water plates instead were ordered (counted from the base of 2 50 mm plates) as 20 cm, 10 mm, 3 mm, 5 mm, 3 mm, 2 mm and 2 mm plates.

(29)

5. Results

5 Results

As this study consists of 3 separate, yet connected, tests the results will be presented test for test and implant for implant in the following subsections.

5.1 Calibration

At the time of scanning it could be concluded that one of the 12 calibration films and one of the control films had been wrongly irradiated and that they therefore could not be used in the test. But the remaining films still proved to be enough to finish the test and perform the calibration. The three RGB channels all generated calibration curves that could be used in the later lateral dose distribution test, see ??????.

(a) (b)

(c)

Figure 17 – The calibration curves created from the red (a), blue (b) and green (c) RGB channel values using the modified version of the method used by Krzempek et al.

As the red channel resulted in the best agreement between the calibration curve and the 3 confirmation films it was deemed to be the most suited one for the study, which meant that the fitting parameters in Eq. (6) had the values Ak=−2.0523, Bk= 2.0597 and Ck= 0.8559. Regarding the secondary control

(30)

5. Results

the maximum pixel to pixel variation of the film piece irradiated with a dose of 1.594 Gy were 0.096 Gy, but as most of the profiles were clearly around the measured dose, the calibration was deemed valid, more details can be seen in appendix A.

5.2 The range effects

As this test were performed in two separate runs (excluding and including the bone equivalent plates) the results found will be presented separately. Using Eq. (5) the maximum water equivalent range shift for the three implants (seen as solid plates of the same thickness of the implants) were found to be 0.651 mm for implant 1, 0.868 mm for implant 2 and 1.085 mm for implant 3.

5.2.1 Including bone equivalent plates

In Fig. 18 the results of the Blue phantom measurements where the bone plates were included in the setup can be seen. Regarding the quality of the Eclipse simulations the results of the comparison of the depth-dose profiles exported from the simulations, which can be seen in appendix B for implant 1 and 2 and Fig. 19 for implant 3. The values for the relative R50shift are further seen within Tab. 2

Figure 18 – The results of the Blue phan- tom measurements for the measurements that included the bone equivalent plates.

Figure 19 – The depth profiles from which the simulated shift in R₅₀ for implant 3, with the setup that included the bone equivalent plates, were calculated.

Table 2 – The results of the range shift measurements for the setup that included the bone equivalent plates. The theoretical maximum range shift is calculated for a solid plate of Ti with the same thickness as the relevant implants using Eq. (5). The values from the Eclipse simulation has been recalculated as WET values.

Implant nr. Theoretical max shift [mm]

Measured shift [mm] Eclipse simulation (regular) [mm]

Eclipse simulation (iMAR) [mm]

1 -0.65 -0.20 -0.05 -0.15

2 -0.87 -0.53 -0.17 -0.15

3 -1.09 -1.03 -0.17 -0.15

(31)

5. Results

5.2.2 Excluding bone equivalent plates

The results of the Blue phantom measurements, for when the bone equivalent plates were excluded, can be seen in Fig. 20. For the quality of the Eclipse simulations the results of the comparison of the depth-dose profiles, exported from the simulations, can be seen in Fig. 21 for implant 1 and appendix C for implant 2. The values for the relative R50shift are further seen within Tab. 3

Figure 20 – The results of the Blue phantom measurements without the bone plates fixated.

Figure 21 – The depth profiles from which the simulated shift in R50 for implant 1, for the setup that excluded the bone plates, were calculated.

Table 3 – The results of the range shift measurements for the setup that excluded the bone equivalent plates. The theoretical maximum range shift is calculated for a solid plate of Ti with the same thickness as the relevant implants using Eq. (5). The values from the Eclipse simulation has been recalculated as WET values.

Implant nr. Theoretical max shift [mm]

Measured shift [mm] Eclipse simulation (regular) [mm]

Eclipse simulation (iMAR) [mm]

1 -0.65 -0.17 -0.00 -0.00

2 -0.87 -0.57 -0.17 -0.18