• No results found

Measurement of absorbed dose to the skin and its relation with microcircular changes in breast cancer radiotherapy

N/A
N/A
Protected

Academic year: 2021

Share "Measurement of absorbed dose to the skin and its relation with microcircular changes in breast cancer radiotherapy"

Copied!
45
0
0

Loading.... (view fulltext now)

Full text

(1)

changes in breast cancer radiotherapy

Master of Science thesis in Medical Radiation Physics

Stockholm University Link¨oping University

(May 27, 2016)

Author:

Chahed Yacoub

Supervisor:

Erik Tesselaar - PhD, Medical Radiation Physics, Department of Medical and Health Sciences, Link¨oping University

Co-supervisors:

(2)

Abstract

Radiation therapy has been shown to increase local and regional control as well as overall survival with breast cancer, but the vast majority of patients develop acute skin reactions, which are in part related to microvascular changes. These reactions vary between different skin sites. The aim of this work is to determine the absorbed dose to the skin by measure-ments and investigate if there is a correlation between the absorbed dose at different areas of the breast and the local changes in microcirculation in the skin after breast cancer radiother-apy. The study includes characterisation of the Gafchromic EBT3 film and Epson Perfection V600 Photo scanner which are used for absorbed dose determination. The measurements were done both on an anthropomorphic female phantom and on a patient undergoing breast cancer radiotherapy. Twenty-one pieces of film (2×1 cm2) were placed on the surface of

the breast (both for the phantom and patient) and irradiated with a prescribed dose to the target of 2.66 Gy with two opposed fields using 6 MV beam.

(3)

Contents

1 Introduction 4

2 Background 4

2.1 Skin reactions . . . 4

2.2 Dose distribution in the build-up region . . . 6

2.2.1 Obliquity . . . 6

2.2.2 Exit dose . . . 7

2.3 Detectors for surface dose determination . . . 7

3 Materials and Methods 8 3.1 Gafchromic EBT Film . . . 8

3.1.1 Energy dependence . . . 10

3.1.2 Variability with film sheet . . . 10

3.2 Scanner . . . 10

3.2.1 Temporal variability . . . 10

3.2.2 Variability with ROI size . . . 11

3.2.3 Spatial variability . . . 11

3.3 Calibration . . . 12

3.4 Film handling for phantom- and patient study . . . 13

3.5 Anthropomorphic phantom study . . . 14

3.5.1 Treatment planning . . . 14

3.5.2 Phantom irradiation . . . 15

3.5.3 Entrance and exit dose . . . 16

3.6 Patient study . . . 16

3.6.1 Treatment of patient . . . 16

3.6.2 Laser Speckle Contrast Imaging . . . 16

3.6.3 Polarised Light Spectroscopy . . . 17

3.6.4 Skin microcirculation measurements . . . 18

4 Results 19 4.1 Gafchromic EBT Film . . . 19

4.1.1 Energy dependence . . . 19

4.1.2 Variability with film sheet . . . 19

4.2 Scanner . . . 19

4.2.1 Temporal variability . . . 19

4.2.2 Variability with ROI size . . . 20

4.2.3 Spatial variability . . . 25

4.3 Calibration . . . 25

4.4 Anthropomorphic phantom study . . . 27

4.4.1 Entrance and exit dose . . . 29

4.5 Patient study . . . 30

4.5.1 Pearson’s correlation test . . . 33

5 Discussion 34 5.1 Gafchromic EBT Film . . . 34

5.2 Scanner . . . 35

5.3 Calibration . . . 36

(4)

5.4.1 Entrance and exit dose . . . 37 5.5 Patient study . . . 38 6 Conclusions 39 7 Acknowledgement 40 8 Bibliography 40 9 Appendix 44 List of acronyms

CPE Charge Particle Equilibrium

CT Computed Tomography

ICRP International Commission on

Radiological Protection

ICRU International Commission on Radiation Units and Measurements

LRA Lateral Response Artefacts

LSCI Lateral Speckle Contrast Imaging

MPV Mean Pixel Value

MU Monitor Unit

MV Megavoltage

OBI On-Board Imaging

PDD Percentage Depth Dose

PLSI Polarised Light Spectroscopy Imaging

PTV Planning Target Volume

PU Perfusion Units

RBCC Red Blood Cell Concentration

ROI Region of Interest

RTOG/EORTC The Radiation Therapy Oncology Group European Organisation for Research and Treatment of Cancer

SD Standard Deviation

SSD Source-Surface Distance

TB Tangential Beam

(5)

1

Introduction

The aim of radiotherapy is to eradicate the tumour while sparing the normal tissue as much as possible. Unfortunately, the number of new cancer cases increases from year to year [1] and the importance to fulfil the aim of the radiotherapy becomes more important since the radiation not only affects the mitosis and causes cell death to the tumour cells, but also the normal cells are affected as an inevitable consequence. According to some studies, 90-95 % of breast cancer patients treated with external radiotherapy will to some extent develop some kind of skin reaction of different grades due to the radiosensitivity of the skin [2],[3]. Porock et al. suggest that only 4-8 % of the women completed the treatment without any reactions [3]. To study the effect of the dose to the skin, there is a need of appropriate techniques that have the ability to measure the energy deposition of the ionisation radiation at shallow depth and techniques that have the ability to verify the changes in the microcirculation quantitatively either by looking at the blood flow or the concentration of red blood cells.

Previous studies have concentrated on estimating the absorbed dose at superficial depth of the skin [4], [5], [6] and the skin reactions due to the radiation have been investigated by subjective methods. Nowadays the possibility to objectively examine the changes in microcirculation of the skin has arisen [7], [8]. It is therefore of interest to determine the absorbed dose at depths which also include the upper dermis, where the skin microcirculation is predominately located. The aim is to determine the absorbed dose to the skin for an anthropomorphic female phantom and a female breast cancer patient and to investigate if there is a correlation between the absorbed dose and the changes in microcirculation in the skin of the breast in breast cancer radiotherapy.

2

Background

2.1 Skin reactions

(6)

Figure 1: The main skin structures of epidermis and dermis. (Adapted from M. Well et al. [13])

Archambeau et al. suggested that the critical prescribed dose to the skin threshold (beginning of the loss of basal cells) is 20-25 Gy and the prescribed dose at which the basal cells are completely damaged is 50 Gy [14]. Considering these values, the skin problems thus start approximately after 10 days of the treatment assuming 2.66 Gy per fraction and increases with each fraction. Observation from clinical practice shows that the maximum acute skin reactions are observed one week after the end of the treatment. However, no studies have been found that prove this statement for energies and doses that are of interest in this thesis.

The acute effects of radiation in the skin are observed visually by increased redness of the skin during the treatment which is due to increased blood flow in the sub-papillary vascular plexus and often referred to as erythema. The late toxicity is due to the fact that the microcirculation and connective tissue are disturbed by the ionising radiation. The effect appears months or years later and may in some cases be permanent [15].

Some areas on the body are more likely to develop skin reactions. Skin folds under the breast and the head and neck have been found to be particularly sensitive [16], [17], [18], [19]. Those areas which contain skin folds are in general more moist. Except the other factors, e.g. irradiation time, that affect the skin reaction severity, also the age, health and ethnicity play a role [3], [20].

Several treatment techniques have been developed during the last years to ensure a better sparing of the normal tissue, including the skin and to get a more conformal dose to the target. However, some techniques involved in the treatment still increase the dose to the skin. These include the use of a bolus to cover a scar when the dose is built up and the use of tangential fields in the treatment of breast cancer. A particular problem with radiation therapy for breast cancer is the sloping form of the breast which makes the probability higher for some thin areas to receive a higher dose [13].

(7)

desqua-mation and faint erythema but instead they are scored the same. According to the patients, these reactions do not have the same severity grade. Secondly, only the subjective opinion of the observers are considered and the results are often underrated in severity if compared to the opinion of the patients [13]. Despite this, this grading score does distinguish for example between faint erythema and tender/bright erythema which makes it useful.

Table 1: RTOG/EORTC acute scoring criteria - skin [21]

0 1 2 3 4

No change Follicular, faint or Tender or bright Confluent, moist Ulceration, of baseline dull erythema; erythema, patchy desquamation haemorrhage,

epilation; dry moist desquamation; other than skinfolds, necrosis desquamation; moderate oedema pitting oedema

decreased sweating

2.2 Dose distribution in the build-up region

When the megavoltage (MV) photon beam hits the patient or phantom, secondary charged par-ticles are released and deposit their kinetic energy in the media through Coulomb interactions. The range of these secondary particles corresponds to the dose build-up region and refers to the region between the surface (z=0) and the depth dose maximum z=zmax. In this region, there is lack of electronic equilibrium, i.e. charge particle equilibrium (CPE) and the dose gra-dient is steep. According to Metcalfe et al., the relative dose in the first millimetres using a 6 MV photon beam and a field size of 10×10 cm2 increases from 14 % to 43 % [22]. However, this example refers to one measurement and varying results may exist. Nevertheless, the steep dose gradient at shallow depth makes the determination of the surface dose a challenge since a small variation in the depth may result in large dose alteration. This issue becomes even more emphasised for curved structures such as the breast.

The factors that contribute to the surface dose are the photons scattered from the collimator, flattening filter and air, the backscattered photons from the patient/phantom and the electrons with high energies produced in the air or shielding in the vicinity of the patient/phantom by photon interactions. However, the largest contribution to the surface dose is due to the scattered electrons and not due to the scattered photons. For higher energy photon beams, there is less contribution to the surface dose. For normal incident beams, approximately 15 % of the dose is deposited at the surface for 6 MV beam while for the 18 MV beam 10 % is deposited for a field size of 10×10 cm2 [23]. The surface dose is also dependent on the field size [24] and according to Kry et al., the surface dose given as percentage of dose maximum (entrance dose) increases linearly with field size [11].

2.2.1 Obliquity

(8)

ionisation chamber was placed in the central axis of the beam. Thus, due to the inverse square law and geometry of the beam, a large part of the beam will hit the surface of the phantom far from the central axis, leading to that the ionisation chamber will not be able to detect all the electrons ejected in the phantom [26]. It was concluded that the largest increase in absorbed dose to the surface is observed at the first 3 mm of tissue for 6 MV x-ray beams.

Gerbi et al. also studied the variation of surface dose and oblique angles of incident (0◦-84◦) depending on the beam quality (6-, 10-, 18- and 24 MV) using a plane parallel ionisation chamber in a polystyrene phantom. He observed that for angles above 45◦, the absorbed doses compared to absorbed dose for normally incident beam, increase markedly for 10-, 18- and 24 MV beams compared to 6 MV beam. It was explained by the fact that photons of higher energy will travel further into the body before depositing their energy while low energy photons will have a much shorter penetration depth in the body and deposit most of their energy closer to the surface [26]. However, it is of importance to notice that Gerbi used a flat phantom for these measurements and the results will vary if one consider a curved surface instead.

For smaller angles (< 40◦) the effect on the surface dose is small. At ∼55◦, the dose increases with 50 % compared to normal incident beam [27]. For 6 MV with a field size of 10×10 cm2, the relative surface dose is 58-65 % (of Dmax) at a beam angulation of 90◦ [28] [29].

2.2.2 Exit dose

Exit dose refers to the dose given to the phantom/patient at beam exit area. The dose curve is somewhat more bended downwards towards the exit point than what is expected from the extrapolation curve. The reason for this observation is loss of scattering contribution beyond the dose exit point but still this effect of a more bended dose curve than expected by the extrapolated curve is considered to be negligible [23]. The exit dose is a contributing factor to the skin dose and must be considered for megavoltage treatment. Measurements have shown that the contribution to the skin dose from the exit dose is less than predicted by the percentage depth dose (PDD). The main reason is the lack of material in the vicinity of the patient that backscatter the charged particles. For a 6 MV beam, the relative difference is approximately 15 % [29], [30], [31].

2.3 Detectors for surface dose determination

An appropriate detector for surface dose determination in this project would be a detector that has a high sensitivity in the expected dose range (0.1-3 Gy), which is made of a near tissue equivalent material and that measures the ionisation at a depth closely corresponding to the depth of the upper dermis. It should also be possible to place the detector at different locations on the curved surface of the breast in order to make regional comparisons of the absorbed dose.

(9)

when used at a surface of a phantom. Thus, making this type of chamber unsuitable for this study [36]. Furthermore, for measurements on anthropomorphic phantom/patients with curved surface, plane-parallel chamber would not be the detector to choose. Other detectors with ap-propriate properties are the MOSFET/MOSkin detectors that have a linear response and are reproducible in the dose range 0.05 Gy to 3 Gy [5].

Another detector type that has been used in skin dose estimations is thin TLDs with a thickness of approximately 130 µm and it overestimates the surface dose by less than 10 % over a beam quality range that includes60Co, 6 MV and 21 MV [12]. Lastly, the use of radiochromic film for surface dose measurements has many advantages, due to its minimal energy dependence, tissue equivalent properties and allows for high spatial resolution at shallow depths [GafchromicT M EBT3, Scan handling guide, P/N 828533 12/14, Rev. 1], [37], [38], [39]. The recommended dose range for these films is 0.01-10 Gy and measures at a depth between 100-128 µm. All the other mentioned detectors measure at a certain point while the Gafchromic films measures over an area. In this study, it is thus more convenient to obtain the dose over an area than a certain point. Also, the method for measuring would require longer time if the detector has to be moved around to measure regional changes of the dose and it would thus not be possible to perform patient measurements. It is therefore concluded that the Gafchromic EBT3 film is the ultimate choice for this study. An overview of the characteristics of detectors suitable for measurements of surface dose and in the build-up region is done by Shea et al. [35].

3

Materials and Methods

3.1 Gafchromic EBT Film

In this study, the dosimetry device of choice is the Gafchromicr EBT3 film (Ashland Specialty Ingredients, Bridgewater). The photon energy dependence of a Gafchromic EBT film dosimeter has been investigated and showed to be minimal by several studies [40], [41]. Rink et al. concluded that the mean change in optical density is 3 % compared to that in60Co beam over an energy range of 75 kVp-18 MV [41].

A study by Nakano et. al confirmed the suitability and accurateness of the Gafchromic film for surface dosimetry as well as the suggestion of using this detector for in vivo measurements [42]. These conclusions are based on the fact that the film material has similar interaction properties as the skin and allows for high spatial resolution for dose measurements at submillimetre depth [43].

(10)

proportional to the dose. According to the manufacturer, the EBT3 films have a recommended dose range of 0.01-10 Gy.

Figure 2: Configuration of GafchromicT M EBT3 Dosimetry Film.

The procedures describing how to handle the film are recommended by the AAPM Report No. 63 [45] and followed in a large extent in this study to ensure high accuracy of the readout. The scan response of the EBT3 film have been observed to be sensitive to how the user place the film on the scanner. The film is not dependent on the obliquity of the beam which is an advantages when it is used as a dose verification tool during tangential beam (TB) irradiations and for modalities such as VMAT [43].

(11)

3.1.1 Energy dependence

According to the manufacturer of the EBT3 film, these films should be energy independent. This was checked for the two beam qualities that are common in breast cancer therapy: 6 MV and 15 MV. The procedure was done by using one sheet of film that was cut into two smaller pieces with an area of 5×6 cm2 each and marked to keep track of their orientation relative to the original film and assure that the film pieces are scanned in the same orientation. The films were consequentially placed in a PMMA slab at 3 cm depth with a backscatter material of 10 cm using a SSD of 100 cm. The films were irradiated with approximately 1 Gy for both beam qualities. The same procedure was done for both energies.

3.1.2 Variability with film sheet

As stated by the manufacturer and confirmed by ˚Astrand [48], the different film sheets from the same batch should be identical and thus it should be sufficient to do the calibration for only one film sheet per batch of films. This was verified in this study by irradiation of three pieces of films (5×6 cm2) from three different film sheets. Each piece of film was irradiated separately. The film set-up used is the same as in section 3.1.1 and all films were irradiated with approximately 1 Gy using 6 MV beam.

3.2 Scanner

To scan a film, the flatbed Epson Perfection V600 Photo scanner (model: J252A, serial no.: *LU5W000052*) is used. This type of scanner uses the triplet-channel method and can collect images at a depth of 16 bits per colour channel. The spatial resolution of such scanner is 6400×9600 dpi (optical density).

The program EPSON Scan (Ver. 3.9.2.0SV) was used for digitalising of the films. The settings could not be chosen exactly as proposed by the EBT3 manufacturer due to an older version of the program. Thus, the films were scanned in professional mode and reflective document type with 48-bit colour and a spatial resolution of 75 dpi. As recommended, all image adjustment features were turned off.

To enable the calculation of the mean pixel value (MPV) by choosing a ROI, a program named ImageJ [49] was used. The digitised images were opened with ImageJ and a ROI was chosen. The program measured the area of a ROI (number of pixels), the MPV and the standard deviation for every MPV.

To minimize the source of error, the variation of the MPV over time, with the ROI size and the position of the film on the scanner was investigated. For all films, the scanning was done in the landscape orientation and the orientation was marked on the films with a pen.

3.2.1 Temporal variability

(12)

3.2.2 Variability with ROI size

An unexposed film sheet was placed in the centre of the scanner and scanned 30 times without removing it from the scanner. Using the program ImageJ, three different sizes of ROIs were chosen to see if there is an effect which needs to be considered. The small ROI has an area of 0.31 cm2 and consists of 256 pixels. The medium ROI covers 9600 pixels and has an area of 11.73 cm2. The largest ROI consists of 81 472 pixels and has an area of 99.54 cm2.

The results show a pronounced difference in MPVs between the medium ROI and large ROI. Therefore, one of the images obtained from 30 scans was alternated in terms of contrast (bright-ness) by changing the window settings in ImageJ to see more clearly the variation of optical density. Furthermore, paper sheets were scanned to see if the variation in optical density is due to the scanner background or the LED-lamp in the scanner. Lastly, a film sheet was irradiated to 200 MUs using a 6 MV beam was scanned after 24 hours and the contrast was altered to conclude whether the variation of the optical density over the scan area will affect the irradiated films further on in the project.

3.2.3 Spatial variability

An unexposed film was cut into a piece of 5×5 cm2. The film piece was marked with a small arrow to keep track of the orientation relative to the original sheet and thus allow for correct placement on the scanner, see Figure 3. The varied variables in this experiment are x and y, where x is the distance between the right edge of the scan area and the centre of the film area. The variable y defines the distance between the nether end of the scan area and the centre of the film area. By varying x and y according to Table 2, a result of different MPVs is obtained. The ROI was chosen in the centre of the film and included 5621 pixels.

Table 2: The six different placements of the film on the scanner area. Scan number Distance x cm Distance y cm

(13)

Figure 3: a) Scan area (22×30 cm2), b) Film area, c) ROI. The film is handled as recommended by AAPM [45].

3.3 Calibration

All irradiations in this thesis are done using a Clinac IX (Varian Medical Systems, installed in 2011 at Link¨oping university hospital, Sweden). The electrometer UNIDOSwebline (PTW, Germany) was turned on one hour before the irradiation to warm up and exclude possible errors connected with the electrometer readout. An electric voltage of -400 V was applied before the electrometer was connected to the Farmer ionisation chamber type 2571 (serial number 1921). This ionisation chamber has a sensitive volume of 0.69 cm3 and thin walls consisting of high purity graphite. The useful energy range is between 50 keV and 35 MeV [50]. According to the specification of this chamber it is recommended to expose the chamber to 2 Gy prior the start of measurements. Furthermore, it is of importance to correct for influence factors (for example pressure and temperature) for each measurement. The chamber is cross calibrated to an identical chamber which has been calibrated at the Swedish secondary standard dosimetry laboratory at Swedish Radiation Safety Authority once every other year using60Co-source.

(14)

Figure 4: Illustration of the set-ups using an ionisation chamber in one case and a film in the other case. Phantom material is PMMA and a SSD of 100 cm is used.

Twelve pieces of film with an area of 5×6 cm2each were placed separately in a PMMA phantom at 3 cm depth with a backscatter material of 10 cm using a SSD of 100 cm, as illustrated in Figure 4. A field size of 10×10 cm2 was used and the film was placed in the centre of the field. All 12 films originate from the same sheet of film and were marked with number of MU together with the direction relative to the original film sheet. One of the films was kept unirradiated (0 MU) while the other remaining films were irradiated to 10, 20, 35, 50, 75, 100, 125, 150, 175 and 200 MU. The scanning of the films was done twice: 24 hours and 48 hours after irradiation to see if there are significant differences in the readout depending on the time after exposure. According to AAPM, it is recommended to read the films at least 24 hours after exposure, but to wait 48 hours is preferred [45].

Two films were irradiated by an assistant to ”unknown” doses to further verify the uncertainties in the calibration curve. The readouts of the ionisation chamber (type 2571) were noted by the assistant and the determined absorbed doses were to be compared with the doses obtained from the calibration fitted curve. All absorbed dose determinations in this study are done by following the IAEA TRS-398 [51] and using the appropriate correction factors. The beam correction factors used are valid for 6 MV photon beam considering reference condition in a water phantom (SSD of 100 cm, 10 cm depth and 10×10 cm2 field). The films were scanned 24 hours and 48 hours after exposure.

3.4 Film handling for phantom- and patient study

(15)

MPVs for each line profile were imported to Matlab, divided by the mean value and the absolute relative difference from the mean was calculated. A threshold for the relative difference was chosen to 1.5 % and the number of pixels exceeding this threshold were calculated for the four edges of the film. The number of pixels were converted into centimetres by knowing that 1 pixel is 0.035 cm. This procedure was done for all twelve films and the mean distance from each edge of the films was calculated including the result from each film. The results shows that not more then 0.2-0.3 cm as an edge margin is needed to get stabilized MPVs within a ROI. Thus making it possible to conclude that the use of small pieces of film during the treatment will not add uncertainties.

3.5 Anthropomorphic phantom study

3.5.1 Treatment planning

(16)

Figure 5: The treatment plan for the left breast, showing the dose distribution in colour wash and the dose volume histogram in Eclipse

3.5.2 Phantom irradiation

The phantom was placed in the isocentre by using the lasers in the room and at an SSD of 100 cm. Before placing the films on the phantom, On-Board Imager (OBI) system was used to take kV-images. These images were matched to the CT-images taken previously and the system calculated if further adjustments of the couch were needed to match the positioning of the phantom. Thereafter, 21 pieces of film were placed on the left breast of the female phantom, as shown in Figure 6. The phantom was irradiated according to the treatment plan.

(17)

3.5.3 Entrance and exit dose

To estimate how much the entrance dose and exit dose contribute to the absorbed dose re-spectively, a set of 21 new films were placed at exactly the same positions as presented in Figure 6. The phantom was irradiated with only one of the fields in the treatment plan and its corresponding compensation field. The same procedure was done for the second field with its corresponding compensation fields using 21 new films. It was decided visually by using the field-light on the phantom for a certain field (124◦ and 305◦) which films that should represent the entrance dose and which films that should represent the exit dose. For the 124◦ field, the following films represent the entrance dose: 7 and 15-17, while the films at placement 9 and 10 represent the exit dose. For the 305◦ field, the entrance dose will be represented by the films placed at 9 and 10. The exit dose will be represented by the films placed at 6, 7, 15 and 16. All films were scanned 24 h after exposure.

3.6 Patient study

3.6.1 Treatment of patient

The patient study protocol was approved by the Regional Ethics Board in Link¨oping, Sweden (DNr 2014/299-31) and a female patient with breast cancer agreed to be a part of this study. The patient has undergone surgery for the removal of the tumour prior to the start of radiotherapy. The tumour was located in the right breast and the prescribed dose for this patient is the same as chosen for the phantom, namely 2.66 Gy in 16 fractions using 6 MV beam. The treatment plan included two opposed fields: one field at 57◦ with two compensations fields and the second field at 234◦ with one compensation field. The patient was told to not use any kind on lotion on that area the same day as the treatment to not affect the films in any manner.

During the treatment, the patient laid in supine position using an overhead arms positioner that supports the arms and a Prostep that supports the legs. These are fixation techniques that are usually used during breast cancer radiotherapy treatments. The placement of the films was done before the images were taken with OBI in order to not affect the positioning of the patient. Twenty-one pieces of film (2×1 cm2) were taped on the breast as shown in Figure 18. The placement of the films is similar to that described in section 3.5.2 but instead it is mirrored on the right breast.

3.6.2 Laser Speckle Contrast Imaging

To assess the changes in skin blood flow in the breasts during treatment, a laser speckle con-trast imaging (LSCI) camera was used (PeriCam PSI, Perimed AB, J¨arf¨alla, Sweden). LSCI is a high resolution and fast technique that uses coherent light for visualization of the microcir-culation.

(18)

and is detected by the CCD camera which is a part of a PeriCam PSI system. If there is no motion, the speckle pattern is stationary, while the speckle pattern will change over time if the illuminated volume contains moving objects. It is affected in such way that the local speckle pattern becomes blurred out and a lower contrast observed at higher blood flows since the moving RBCs are those that scatter the light. Contrary, high contrast regions represent areas where the blood flow is low. Thus, an adequate correlation between the speckle contrast and blood perfusion is possible. The speckle contrast C is usually defined as [52]:

C ≡ σ¯

I (1)

where σ is the standard deviation of the intensity I and ¯I is the mean intensity of the speckle pattern. When the blood flow increases in the illuminated tissue volume, the standard deviation will decrease due to blurring and result in a low speckle contrast. The mean intensity does not change.

It was chosen that the camera should take 21 frames per second. For each of the the frames, the speckle contrast is determined for every 3×3 pixel matrix over the whole frame. By overlapping all 21 frames and taking the mean contrast over every 3×3 pixel matrix from all the 21 frames, a resultant image is reconstructed. Thus, the noise in the image can be reduced. In total, five images (15×15 cm2 each) are obtained with one second time difference in between. The spatial resolution of the system is 20 pixels per millimetre. The PeriCam PSI version that is used in this study uses normal resolution (100 µm/pixel at 10 cm camera-surface-distance) with variable measurement area [53].

3.6.3 Polarised Light Spectroscopy

Polarised Light Spectroscopy is a noninvasive technique that is based on a digital camera and has the ability to measure the concentration of the RBCs in the upper dermis using polarised light. This technique has been shown to be insensitive to motion artefacts, has the ability to measure on relatively large skin areas and produce images with high spatial resolution. In vitro study showed that this technique is independent of the oxygen concentration in the blood[7].

A polarised light spectroscopy imaging (PLSI) system (TiVi600, Wheelsbridge AB, Link¨oping, Sweden) was used in this study for the purpose of measuring the changes in RBC concentration (RBCC). This camera system is equipped with 96 white light LED-sources and two polarisation filters, one in front of the flash and one in front of the detector lens. The light source produces randomly polarised (RP) light that becomes linearly polarised as it passes through the polari-sation filter. As the light hits the skin, a portion becomes randomly polarised while a portion of the light remain linearly polarised, see Figure 7. The linearly polarised light is directly re-flected from the epidermis, while the randomized light changes characteristics as it is randomly scattered in the tissue. The polarisation filter that is located in front of the lens filters away the directly reflected light so that only the randomly polarised light can reach the detection array. The higher the RBCC is in the blood, the higher is the total detected intensity/signal.

(19)

and blue (RGB) images. The RGB values are then used to calculate the RBCC by an algorithm that utilises the differences in absorption in the different wavelengths ranges. The algorithm subtracts each pixel value in the green colour image from the corresponding pixel value in the red colour image. The difference is then divided by the detected signal and the TiVi-indices (Tissue Viability indices) are obtained. The TiVi-indices are linearly correlated to the concentration of the RBCs in the volume of tissue. (A more detailed description of the technology is done by O’doherty et. al [7]). Due to the fact that the PLSI technology measures at relatively large areas, it has the ability to include variability in the response between different sites of the skin due to inhomogeneities in the microcirculation. These heterogeneities are due to difference in innervation, physiological response and vessel density.

Figure 7: A schematic overview of how the polarised light spectroscopy works. The randomly polarised (RP) white light from 96 LED-sources gets linearly polarised by a first filter before hitting the skin surface. A part of the light will be backscattered from the surface while another part will be randomly polarised by the tissue. Only the randomly polarised light will be able to pass through the second filter, becomes linearly polarised before it reaches the detection array and a 8-bit RGB-images can be produced by digitised image processing. (Adapted from O’Doherty et. al [7])

3.6.4 Skin microcirculation measurements

(20)

4

Results

4.1 Gafchromic EBT Film

4.1.1 Energy dependence

The results of the comparison between 6 MV and 15 MV are presented in Table 3. The MPVs for both beam qualities (as the films are scanned 24 h and 48 h after exposure) are normalised to approximately 1 Gy. The MPVs are normalised to the absorbed dose to get a relative difference that is independent of the ionisation chamber signal and thus considering one parameter less. The relative differences are obtained by comparing the normalised MPVs (24 h) for both energies. The same calculation was done for the normalised MPVs (48 h). According to Table 3, the largest relative difference for the MPVs is 0.59 %.

Table 3: The MPVs obtained 24 h and 48 h after exposure normalised to 1 Gy, given for the two energies: 6 MV and 15 MV. The relative differences between the two energies for both time-after-exposure are represented as well.

Energy [MV] MPV (24 h)/absorbed dose MPV (48 h)/absorbed dose

[Gy−1] [Gy−1]

6 28652.2 28487.9

15 28483.7 28344.2

Difference [%] 0.59 0.51

4.1.2 Variability with film sheet

The results of the variability with film sheet are shown in Table 4. Film 1 is chosen to be the reference film since the irradiated pieces of film in previous sections originate from the same sheet. Thus, the MPVs and absorbed doses for Film 2 and Film 3 are compared to Film 1 and the absolute relative differences are calculated. The results indicate not more than 2.2 % deviation for both time-after-exposure.

Table 4: The variation in MPV and absorbed dose for 3 different film sheets from the same lot. The films were irradiated to ∼1 Gy. Film 2 and 3 are compared to Film 1 and the relative differences are represented in this table.

MPV (24 h) Absorbed dose MPV (48 h) Absorbed dose

(24 h) [Gy] (48 h) [Gy]

Film 1 28853.9 0.92 28758.0 0.91

Film 2 28799.6 0.92 28673.7 0.92

Film 3 28952.6 0.90 28812.9 0.91

Diff. (Film 2/Film 1) [%] 0.19 % 0.00 % 0.29 % 1.10 %

Diff. (Film 3/Film 1) [%] 0.34 % 2.17 % 0.19 % 0.00 %

4.2 Scanner

4.2.1 Temporal variability

(21)

of the MPVs (± 1 standard deviation (SD)) is calculated to be 43121 ± 18.

Figure 8: Mean pixel value variation over a time of 49 minutes for 15 scans using the same film sheet. The error bars are calculated by ImageJ and represent the variation of the pixel values in the chosen ROI within 1 SD.

4.2.2 Variability with ROI size

The variations in MPVs for the small, medium and large sized ROIs are shown in Figure 9. A linear fitting was done for all three ROIs. The mean of the MPVs over a ROI (± 1 SD) is calculated to be 43166 ± 13, 43032 ± 13 and 43028 ± 16 for small ROI, medium ROI and large ROI, respectively.

(22)
(23)

Figure 10: Mean pixel value for the three different ROIs: small (256 pixels), medium (9600 pixels) and large (81 472 pixels).

(24)

Figure 11: The image of scan number 15 with changed contrast (brightness). The contrast was changed by changing the window settings in ImageJ. The MPVs from the ROIs in image 1 and image 2 are compared.

Table 5: The obtained MPVs for the ROIs in image 1 and image 2 using scan number 15. MPV Image 1 MPV Image 2 Difference [%]

32118.6 50669.3 36.61

(25)

Figure 12: Image 1: One paper sheet scanned. Image 2: Two paper sheets scanned. Image 3: Ten paper sheets scanned. The contrast (brightness) was alternated by changing the window settings in ImageJ for all three images.

The contrast of the image of the exposed film was altered and is shown in Figure 13. Two ROIs with an area of 6.3 cm2 each where drawn on the darkest and brightest part of the image, see image 1 and 2 respectively in Figure 13. The MPVs were obtained and the absorbed doses were calculated from the exponential fitted curve (24 h) in Figure 15. The MPVs and absorbed doses for both ROIs were compared and are represented in Table 6.

(26)

Table 6: The obtained MPVs for the ROIs in image 1 and image 2 using a film sheet irradiated to 200 MUs with 6 MV beam. The two ROIs are compared in terms of MPV and absorbed dose (obtained from the calibration curve (24 h)) and the difference is represented as relative difference.

MPV Absorbed dose [Gy]

Image 1 21879.8 2.09

Image 2 22264.7 2.00

Difference [%] 1.76 4.50

4.2.3 Spatial variability

The MPVs for different placements of the films on the scanner surface are shown in Figure 14. The positioning of the films was varied according to Table 2. A ROI was chosen in the centre of the each film and included 5621 pixels.

The recommend placement of the film is in the centre, as in scan number 2 (x=11 cm, y=15 cm). Hence all the obtained MPVs are compared to scan number 2. The two placements that result in the largest discrepancy are scan number 1 and 5 with an absolute percentage deviation of 0.56 % and 0.48 %, respectively. Nevertheless, the absolute mean percentage deviation compared to scan number 2 is calculated to be only 0.23 %.

Figure 14: The MPVs for different placements of the film on the scanner area.

4.3 Calibration

(27)

from ImageJ and shown in the plots. For comparison reason, the two plots in Figure 15 are plotted together and shown in Figure 16. The absolute mean relative difference between the MPVs for scanning after 24 h and 48 h was calculated to be 0.33 %. The absolute mean relative difference between the doses for 24 h and 48 h was calculated to be 0.34 %.

Figure 15: The calibration curves are obtained using the MPVs and by calculation of the absorbed dose using the ionisation chamber readout. Exponential fitting of second degree is used to obtain the calibration curve. The scanning of the films is done 24 hours and 48 hours after exposure. Two films were irradiated to unknown doses, scanned after 24 h and 48 h and plotted in the figures as well.

(28)

Using the fitted curves from Figure 15 for respective time-after-exposure, the absorbed doses for the films irradiated to ”unknown” doses were obtained and are presented in Table 7. The absorbed doses that are calculated using the ionisation chamber signal, are shown in the same table. The film dose values are compared to the ionisation chamber dose values and the absolute relative difference is shown in the table as well. According to the results, the relative difference can reach up to approximately 2 %.

Table 7: The film doses are obtained from the fitted calibration curves. The ionisation cham-ber absorbed doses are calculated from chamcham-ber signal. The film doses are compared to the ionisation chamber absorbed doses and represented as absolute relative differences.

Film dose Film dose Ionisation chamber (24 h) [Gy] (48 h) [Gy] absorbed dose [Gy]

Film 1 0.1415 0.1387 0.1389

Film 2 0.8216 0.8220 0.8125

Film 1 Diff. [%] 1.87 0.14

Film 2 Diff. [%] 1.12 1.17

4.4 Anthropomorphic phantom study

(29)

Table 8: The absorbed dose at different film placement using the female phantom. The relative dose gives percentage of the absorbed dose relative to the prescribed dose of 2.66 Gy.

Film placement Absorbed dose [Gy] Relative dose [%]

1 1.60 60.2 2 1.57 59.0 3 1.62 61.0 4 1.36 51.1 5 1.68 63.2 6 1.47 55.3 7 1.39 52.3 8 1.55 58.3 9 1.30 48.9 10 1.24 46.6 11 1.51 56.8 12 1.68 63.2 13 0.10 3.8 14 0.51 19.2 15 1.37 51.5 16 1.18 44.4 17 1.30 48.9 18 0.11 4.1 19 0.86 32.3 20 1.07 40.2 21 0.14 5.3

(30)

4.4.1 Entrance and exit dose

The absorbed doses using the field at 124◦ and two compensation fields with the same incident angle for irradiation of the phantom are shown in Table 9. The doses are compared to the prescribed dose and presented in the table as relative doses given in percent. In Table 10, the doses are represented for the field at 305◦ with one compensation field.

Table 9: The absorbed dose at different film placement with treatment at 124◦ gantry angle using the female phantom. The relative dose gives percentage of the absorbed dose relative to the prescribed dose of 2.66 Gy.

Film placement Dose field 124◦ [Gy] Relative dose [%]

(31)

Table 10: The absorbed dose at different film placement with treatment at 305◦ gantry angle using the female phantom. The relative dose gives percentage of the absorbed dose relative to the prescribed dose of 2.66 Gy.

Film placement Dose field 305◦ [Gy] Relative dose [%]

1 0.63 23.7 2 0.86 32.3 3 0.59 22.2 4 0.38 14.3 5 0.70 26.3 6 0.70 26.3 7 0.71 26.7 8 0.58 21.8 9 0.37 13.9 10 0.36 13.5 11 0.64 24.1 12 0.64 24.1 13 0.07 2.6 14 0.38 14.3 15 0.60 22.6 16 0.64 24.1 17 0.70 26.3 18 0.04 1.5 19 0.09 3.4 20 0.28 10.5 21 0.05 1.9 4.5 Patient study

(32)

Figure 18: The film placement on the right breast of a female patient before irradiation. The right image shows the breast is taken in anterior direction. The left image shows the breast from right in the lateral direction.

Table 11: The absorbed dose at different film placement during a patient treatment. The relative dose gives percentage of the absorbed dose relative to the prescribed dose.

Film placement Dose treatment [Gy] Relative dose [%]

(33)

Figure 19: A 3D Matlab illustration of the female patient with 21 films placed on the skin of the right breast. The colour bar gives the absorbed doses in Gy for each film. The doses are obtained from one fraction with a prescribed dose of 2.66 Gy.

Figure 20 shows how the change in perfusion (blood flow) and TiViindex (RBCC) due to the 16 irradiations of the patients breast vary with the skin doses obtained at the same regions. The absorbed doses are calculated from the doses obtained from Table 11 and multiplied by a factor 16 to get the total accumulated dose at the end of the treatment.

(34)

Table 12: For all film placements: the total absorbed skin dose after 16 fractions (2.66 Gy/fraction), the change in mean perfusion and in TiViindex (RBCC) before the start of the treatment and after the 16th fraction.

Film Absorbed Change in mean Change in TiViindex placement dose [Gy] perfusion [PU] [A.U.]

1 25.00 190.37 184.35 2 26.08 154.93 244.91 3 24.48 107.31 211.14 4 22.40 66.69 129.37 5 26.08 30.75 67.27 6 23.84 7.86 130.33 7 25.00 74.71 105.86 8 24.80 17.06 141.33 9 20.64 21.29 74.32 10 20.48 28.42 49.09 11 25.92 22.25 77.33 12 27.04 16.94 136.00 13 3.04 27.79 23.53 14 24.48 16.60 65.00 15 21.44 2.11 -35.03 16 21.28 6.05 -3.95 17 21.76 7.64 -33.21 18 5.92 1.21 13.97 19 19.68 17.81 65.46 20 18.72 24.06 65.76 21 18.08 33.54 67.58

4.5.1 Pearson’s correlation test

Pearson’s correlation test measures the correlation between two variables to see if a linear correlation exist. The correlation is referred to as Pearson’s r and indicates a perfect correlation if r = +1/-1. If there is no correlation between two variables, the correlation coefficient r equals 0. However, some assumptions must be accounted for when performing this type of correlation test. These assumptions include that the two variables must be from the same sample group, measured independently and be sampled from a Gaussian distributed sample. The covariance is assumed to be linear [54].

(35)

Table 13: Pearson’s correlation coefficient r calculated to test the correlation between the change in mean perfusion and absorbed dose as well as between the change in TiViindex and the absorbed dose. A two-tail test with a significance level of 0.05 was done.

X/Y Change in mean PU/ Change in TiViindex/

Absorbed dose Absorbed dose

Pearson’s r 0.30 0.48

P-value (two-tailed) 0.18 0.03

Significance (α = 0.05) No Yes

5

Discussion

The discussion will include analysis of the measurement devices used, the methods and the results. In the analysis of the data obtained in this study, it is worth considering all effects that may influence the results. Some of these effects affects the measurements and results more than others which are considered to be of less importance.

5.1 Gafchromic EBT Film

The LRA will be more pronounced in the red channel due to its high dose dependency for doses above 1 Gy. However, no literature has discussed the reason for why the dose affects the LRA but only the fact that a lower relative response is observed towards the edges of the scan area and this effect increases with increased dose. Since the doses in the this study do not exceed 2 Gy and all the films were placed in the centre of the scan area, it is concluded that the variation in response due to LRA will not affect the results.

The low variation in MPVs (maximum of 0.59 %) with beam quality (6 MV and 15 MV) confirm the statement by the manufacturer as the films being minimally energy independent and agree with the conclusions from other studies [40], [41]. However, since both the phantom and patient had relatively small breast, they were irradiated with only 6 MV. Thus, the variation in MPVs due to the beam quality will not have an impact of the results in this project but may be useful for future studies. In future studies when including a larger population of patients, it would further be of interest to investigate the energy dependence of the films in an energy range down to a few kV. This is to see how the film response will vary due to the OBI which will have an impact of the film response during patient irradiation. Other studies have confirmed the variations in the film response due to energy dependence in kV-range [56], [57].

The relative difference for the variation of MPVs and absorbed doses with film sheet are con-sidered small with a maximum variation of 0.34 % for the MPVs and 2.17 % for the absorbed doses. The conclusion drawn from these results is that all sheets of film in this lot should approximately give the same correlation between the darkening of the film and the absorbed dose.

(36)

When considering the active depth of the films, it is important to have in mind that the pixel values are a result of the integrated ionisation over the active volume/layer. The active layer of the film is ∼ 28 µm (∼3.36×10−3 g/cm2) and the obtained absorbed doses are thus considered to be at a depth between 100-128 µm (1.35×10−2 - 1.69×10−2 g/cm2).

5.2 Scanner

The results show a small increase in MPV over time, see Figure 8. The small increase was unpredicted since scanning the same film a few times and assuming that the LED-lamp in the scanner may affects the film, the result would be a reduction in MPV with time instead. This is since the LED-lamp may darken the film and give lower MPVs. A study done by ˚Astrand shows that the MPV decreases with time [48], although she used another scanner (Epson Perfection V700 Photo). A conclusion that can be made from these results is that the LED-lamp affecting the films is not the reason for the observed increase but it is expected however to be another unknown property of the scanner that gives these results.

Comparing the sizes of the error bars calculated by ImageJ in Figure 8 with the standard deviations of the mean MPVs, it is observed that the error bars covers a larger range of the MPVs (a variation in pixel value of approximately ±230) than the calculated SDs (a variation of ±18). These results indicates that the variation of the pixel values within a chosen ROI is larger than the variation of the MPVs over time. Considering this, it is concluded that no warm up time is needed for the scanner.

The same trend between the error bars in Figure 9 and the calculated SDs in section 4.2.2 is observed, where the SDs are small in comparison to the error bars. Thus, the slightly increase of the MPVs within a chosen ROI size after scanning the same film sheet 30 times, will not have an impact on the absorbed dose determinations in the study. Figure 10 shows that the MPVs decrease markedly between the large ROI and medium ROI, while the difference between the medium and small ROI is less pronounced. However, the results clearly shows that mean of the MPVs for the small and medium ROI are comparable (0.01 % mean difference), while the difference is more pronounced when comparing the mean of the MPVs from medium and large ROI (0.3 % mean difference). The optical density increases towards the lateral edges of the scan area which can be a probable cause for the higher MPV for the large ROI. When analysing the images in Figure 11 with alternated contrast, the results show that certain areas of the film appear as more green-coloured and have reduced MPVs. It is expected that the LED-lamp in the scanner contributes to an uneven distribution of the optical density over the scan area and lead to variations in MPVs that can vary up to 37 %.

(37)

The heterogeneities appearing due to the scanning of the films are less pronounced when the film has been exposed to radiation and the MPVs vary with only 1.76 % (see Figure 13 and Table 6). The variation in the absorbed dose can vary up to 4.5 %. The large difference in the relative differences between the MPVs and the absorbed dose is not surprising, since according to the calibration curve the relationship between the MPVs and absorbed dose is not linear. If a linear relationship would be obtained instead, the relative differences would be the same. However, it must be noted that the variations in optical density found in this study may be specific for this scanner and the same trend will not necessarily be found in another scanner of the same type. Note as well that the most extreme regions in term of colour darkening are compared here and the selection of the regions was done only subjectively. Despite the fact that the films used in the phantom and patient studies were placed centrally on the scan area, the variation in MPVs of 4.5 % is considered to be large and may have affected the absorbed dose determinations in this study to some extent.

The spatial variability of the MPVs depending on the film placement on the scanner area was shown to be minimal (< 1 %). Nevertheless, this effect will not affect the absorbed dose determination in any manner since all the films were placed centrally on the scanner surface as recommended by the manufacturer of the Grafchromic EBT3 film.

5.3 Calibration

The same trend is observed for both plots in Figure 15, where the MPVs decrease with in-creased dose. A similar exponential trend of the calibration curve is observed by ˚Astrand [48]. However, the absolute MPVs are not identical to those represented in this study due to different factors including the calibration method and the scanner used. The results give a mean relative difference of 0.33 % and 0.34 % between the scanning after 24 h and 48 h for the MPVs and dose, respectively. The small relative differences are considered to be negligible in the absorbed dose determinations in the study. It was therefore concluded to scan the films after 24 h after exposure and not necessary 48 h after exposure, due to the time limitation in this study. Further uncertainties appear when comparing the calculated doses for the films that were irradiated to ”unknown” doses with the absorbed doses that were calculated using the ionisation chamber signal. The maximum relative difference of approximately 2 % is somehow an estimation of uncertainties in the calibration curve.

An important consideration regarding the calibration is the irradiation setup used in this study. The setup do not completely match the reference geometry setup recommended by the IAEA TRS-398, since the correction factors are calculated for a water phantom and 10 cm depth in water. Here, a PMMA phantom is used and a depth of 3 cm. Seuntjens et al. concluded that a small error in the perturbation factors is introduced due to the fact that a non-reference geometry is used and the dose ratio water to PMMA is accurate within 0.4 %[58].

5.4 Anthropomorphic phantom study

(38)

The fact that the highest surface doses are observed at the midline of the breast is not surprising since those films are irradiated with both of the tangential radiation fields. The films that were completely out of the irradiation field are film placement 13, 18 and 21. The results from Table 8 show that these films represent the lowest doses. However, the low doses observed for these films are due to scattered radiation. The fact that mainly within 45-64 % of the prescribed dose is deposited in the skin is in good agreement with the founding by Almberg [59] and Rudat [60]. Almberg et al. concluded that around 45-65 % of the target dose results in surface dose while Rudat et al. concluded that 40-60 % of the target dose is deposited at the surface. Both studies used TB-IMRT and Gafchromic film dosimetry.

A surprising result is obtained if the doses from Table 9 and Table 10 are added up together and compared to the doses in Table 8. A mean relative difference of 8.3 % is observed. The expectation was to see that the added doses from the fields at 124◦ and 305◦ would result in the same doses as the full treatment. It is less likely to be caused by the film placement, even though the films were not moved during the full treatment but replaced during the separate field irradiation. This is since the tape pieces from the full treatment were left on the phantom in order to place the new films at the exact same position when irradiating with separate fields. The mean relative difference may in this case not be representative of the differences since the overall relative differences seems to be around 4-5 % while the largest relative differences are observed for film placement 18-21 where the relative difference vary between 17-42 %. These films receive relatively low doses and have probably been affected by the amount of scattered radiation. A contribution to this large difference may be the fact that the full treatment irradiation and the separated fields irradiation were done on different days. Thus, the set-up of the phantom is not identical and also the accelerator output is expected to vary slightly.

When the films were taped on the phantom surface, it was complicated to keep the films totally attached to the surface. Instead, the edges of the film tended to lift up from the phantom surface. Although, the ROIs were chosen with a small margin to the edges to exclude those areas where the films were lifted up, this film behaviour may has affected the absorbed doses. Considering the lack of backscatter contribution from the phantom at these regions on the film pieces, the absorbed doses may be underestimated.

5.4.1 Entrance and exit dose

For the gantry angle 124◦, the maximum dose is observed at film placement 12 with a dose of 0.96 Gy which corresponds to 36.1 % of the prescribed dose of 2.66 Gy. For the gantry angle 305◦ the maximum dose is observed at film placement 2 with a dose of 0.86 Gy which corresponds to 32.3 % of the prescribed dose. The reason for the higher doses at 124◦ than at 305◦ is due to the extra compensation field. However, the same trend as for the full treatment is observed here, where the highest doses appear centrally and in the midline direction of the breast.

(39)

contributes more to the surface dose than the entrance dose for both fields.

For some of the films, the beam was incident with a large obliquity. Due to the fact that the effective depth of the electrons resulting from photon interaction is larger for obliquity incident beams than for perpendicular incident beams, the result will be a higher dose for those films. For the 124◦ field, the beam was incident on the films placed on the midline of the breast (1, 3, 5, 8, 11 and 12) with an angle of approximately 90◦. This explains the higher dose at those regions. The beam incident angle at those six films with the 305◦ field is estimated to be 70◦. Comparing the results in Table 9 and Table 10 at those regions, it is observed that a higher absorbed skin dose is obtained for the 124◦ field. This is a result due to the larger anglulation of incident beam which gives a larger effective depth of the electrons. Note that it is of importance to take into account the extra compensation field that was added at the 124◦ compared to the 305◦ field and the beam angle are only subjective estimations. It is furthermore important to consider the rough estimations of the angulation due to the difficulty to determine the angle between the beam and the skin at a particular point.

5.5 Patient study

The results show that the absorbed doses are in a range between 0.19-1.69 Gy. The three highest doses appear to be at film placement 2, 5 and 12 , which is in agreement with the findings from the phantom study. These doses correspond to 61.3 %, 61.3 % and 63.5 % of the prescribed dose, respectively. The lowest absorbed doses are observed at film placement 13 with a dose of 0.19 Gy and film placement 18 with a dose of 0.37 Gy. These low doses are probably due to the fact the films were positioned outside the radiation field. Furthermore, the same trend is observed for the patient as for the phantom, where the higher doses seems to be situated on the midline of the breast.

In patient study, it is worth considering that the films were attached directly to the skin and may therefore have been affected by skin oils and sweat in a larger extent than in the phantom study. Furthermore, the films in the patient study were irradiated during the OBI in comparison to the films in the phantom study which were placed on the phantom after the OBI. However, these differences between the phantom and patient study does not seem to affect the absorbed doses if comparing the dose ranges from both studies. The dose range (0.10-1.68 Gy, see Table 8) from the phantom study is in good agreement with the dose range obtained from the patient study (0.19-1.69 Gy, see Table 11). Similarity between both studies were also found in the overall absorbed skin dose distribution over the breast. The fact that the size of the patient’s breast happened to be of comparable size to the phantom breast, may explain the results. However, it is important to consider the impact of the diaphragmatic motion in the patient study that adds further uncertainties but is absent in the phantom study.

(40)

Note that there are uncertainties related to the absorbed doses in Table 12, since these doses are obtained by multiplying the doses in Table 11 by a factor 16 to get the total accumulated skin dose after the end of the treatment. Thus, it is assumed in this study that the patient will receive the same skin doses in every fraction in the treatment, which is an estimation. The skin dose may vary between the fractions, not only due to the small variation in the accelerator output but also due to the diaphragmatic motion of the patient.

Since the patient had undergone surgery for tumour-removal before the breast cancer radiother-apy was scheduled, it was expected that marked difference in the change in microcirculation would be observed in the scar region. The surgery scar was located below the film placements 3, 8 and 12. The scar does not seems to affect the mean perfusion since the change in blood flow does not seems to be markedly affect by the scar in comparison to other regions on the breast. However, the change in RBCC is clear at film placements 3, 8 and 12 (see Table 12). The redness around the scar was also observed visually. The increase RBCC in the scar can be explained by an increase in capillary density and diameter and by vasodilation due to tissue repair mechanisms.

The results of the Pearson’s r in Table 13 show that for both tests, the two variables tend to increase or decrease together due to the positive sign. However, the correlation is better between the change in TiViindex with the absorbed dose, than it is for the change in mean perfusion. Furthermore, the p-value is lower for TiViindex which indicates that there is a low probability that the correlation is by chance. The conclusion from these statistical tests is that there is no statistical correlation between the change in mean blood perfusion and the absorbed dose for this patient. Whereas, a significant statistical correlation exist between the change in RBCC (TiViindex) and the absorbed dose (α = 0.05).

When discussing the microcirculation change, it is of importance to account for the biological factors influencing the results, such as the thickness of the epidermis. A conclusion drawn in a study by Sandby-Moller et. al is the positive correlation between the thickness of the cellular epidermis and the blood content, i.e. a thicker layer of the cellular epidermis indicate a greater blood supply. This result was mainly observed in males [61]. Observations of the increase in blood perfusion with radiation skin dose was found by Nystr¨om. Nystr¨om et al. investigated the possibilities to correlate the skin reaction with radiation dose, using inter alia laser Doppler technique which measures the blood perfusion. A general result was observed, that in 79 % of the cases a significant increase in perfusion is observed with increased radiation dose, i.e. the blood flow in the treated area increases with increased skin dose [62]. However, a conclusion could not be drawn regarding how much a specific radiation dose will increase the blood perfusion. This is since the radiosensitivity of the individuals varies and thus also the increase in the perfusion.

6

Conclusions

(41)

conclude whether a general correlation exist or not between the changes in skin microcirculation and the absorbed dose due the small population sample. In order to get more reliable results, a larger sample must be included in future work. It was also found to be difficult to perform a concrete error-analysis which sums up all the error-sources in the study. Furthermore, there is biological properties that are difficult to take into account for, such as the thickness of the epidermis and general health condition. Therefore, the skin doses found in this study can not simply be generalized to other patients. In order to make this possible in the future, a new study should test the correlation over a larger sample with different breast geometries. It would also be of interest to include patients that have undergone mastectomy that often experience more pronounced skin reactions due to the use of bolus during radiotherapy.

The uncertainties related to subjective scoring criteria of the skin reactions have been stated. Future studies should continue with quantification of the skin reactions that appear as a con-sequence of the radiation therapy and include biological related properties (skin thickness, age, sex etcetera) in the prediction of the absorbed skin dose based on the microcirculation proper-ties. In this way, it may be possible to predict the radiosenstivity of the patient and the risk of skin damage in the beginning of the radiotherapy treatment.

Even though the results from this study can not contribute to the development of the treatment technique in radiotherapy or be directly implemented in the clinic, this study shows that the method is reliable and a correlation may exist between the changes in microcirculation and the absorbed skin dose in breast cancer radiotherapy.

7

Acknowledgement

I would like to thank my supervisors E. Tesselaar, E. Adolfsson and B.Nilsson for your help and guidance throughout the project. I would also like to thank all the staff at department of Medical and Health Sciences at Link¨oping University Hospital for making me feel welcome. Another thank goes to my family and soon-to-be husband Dr. Milad Gabro for all the motivation and support you have giving me during the project.

8

Bibliography

[1] Cancerfonden. Aktuell statistik om cancer; 2013. Available from: https://www. cancerfonden.se/om-cancer/aktuell-statistik-om-cancer.

[2] De Conno F, Ventafridda V, Saita L. Skin problems in advanced and terminal cancer patients. J Pain Symptom Manage. 1991;6(4):247–56.

[3] Porock D, Kristjanson L. Skin reactions during radiotherapy for breast cancer: the use and impact of topical agents and dressings. Eur J Cancer Care (Engl). 1999;8(3):143–53. [4] Devic S, Seuntjens J, Abdel-Rahman W, Evans M, Olivares M, Podgorsak EB, et al.

Accurate skin dose measurements using radiochromic film in clinical applications. Med Phys. 2006;33(4):1116–1124.

(42)

[6] Akino Y, Das IJ, Bartlett GK, Zhang H, Thompson E, Zook JE. Evaluation of superficial dosimetry between treatment planning system and measurement for several breast cancer treatment techniques. Med Phys. 2013;40(1):011714.

[7] O’Doherty J, Henricson J, Anderson C, Leahy MJ, Nilsson GE, Sj¨oberg F. Sub-epidermal imaging using polarized light spectroscopy for assessment of skin microcirculation. Ski Res Technol. 2007;13(4):472–484.

[8] Tesselaar E, Bergkvist M, Sj¨oberg F, Farnebo S. Polarized Light Spectroscopy for Mea-surement of the Microvascular Response to Local Heating at Multiple Skin Sites. Micro-circulation. 2012;19(8):705–713.

[9] ICRP Publication 26. International Commission on Radiological Protection; 1977.

[10] ICRU Report 39. International Commission on Radiation Units and Measurements; 1985. [11] Kry SF, Smith SA, Weathers R, Stovall M. Skin dose during radiotherapy: a summary

and general estimation technique. J Appl Clin Med Phys. 2012;13(3):3734.

[12] Nilsson B, Sorcini B. Surface dose measurements in clinical photon beams. Acta Oncol. 1989;28(4):537–42.

[13] Wells M, MacBride SM. Radiation skin reactions. Support Care Radiother. 2003;p. 135– 159.

[14] Archambeau JO, Pezner R, Wasserman T. Pathophysiology of irradiated skin and breast. Int J Radiat Oncol Biol Phys. 1995;31(5):1171–85.

[15] Nystr¨om J, Geladi P, Lindholm-Sethson B, Rattfelt J, Svensk AC, Franzen L. Objective measurements of radiotherapy-induced erythema. Ski Res Technol. 2004;10(4):242–50. [16] Crane J. Extending the Role of a New Hydrogel. J Tissue Viability. 1993 jul;3(3):98–99. [17] Dische S, Warburton MF, Jones D, Lartigau E. The recording of morbidity related to

radiotherapy. Radiother Oncol. 1989;16(2):103–8.

[18] Farley KM. Cornstarch as a treatment for dry desquamation. Oncol Nurs Forum. 1991;18(1):134.

[19] O’Rourke ME. Enhanced cutaneous effects in combined modality therapy. Oncol Nurs Forum. 1987;14(6):31–5.

[20] Sitton E. Early and late radiation-induced skin alterations. Part I: Mechanisms of skin changes. Oncol Nurs Forum. 1992;19(5):801–7.

[21] Cox JD, Stetz J, Pajak TF. Toxicity criteria of the Radiation Therapy Oncology Group (RTOG) and the European Organization for Research and Treatment of Cancer (EORTC). Int J Radiat Oncol Biol Phys. 1995;31(5):1341–6.

[22] Metcalfe P, Kron T, Hoban P. The Physics of Radiotherapy X-rays from Linear Accelera-tors. 1st ed. Medical Physics Pub Corp; 1997.

[23] Podgorsak EB. Radiation Oncology Physics: A Handbook for Teachers and Students. Podgorsak EB, editor. Vienne, Austria: International Atomic Energy Agency; 2005. [24] Nilsson B, Brahme A. Absorbed dose from secondary electrons in high energy photon

(43)

[25] Dupont S, Rosenwald JC, Beauvais H. Convolution calculations of dose in the buildup regions for high energy photon beams obliquely incident. Med Phys. 1994;21(9):1391–1400. [26] Gerbi BJ, Meigooni aS, Khan FM. Dose buildup for obliquely incident photon beams. Med

Phys. 1987;14(3):393–399.

[27] Lin JP, Chu TC, Lin SY, Liu MT. Skin dose measurement by using ultra-thin TLDs. Appl Radiat Isot. 2001;55(3):383–91.

[28] Quach KY, Morales J, Butson MJ, Rosenfeld AB, Metcalfe PE. Measurement of radio-therapy x-ray skin dose on a chest wall phantom. Med Phys. 2000;27(7):1676–80.

[29] Gagnon WF, Horton JL. Physical factors affecting absorbed dose to the skin from cobalt-60 gamma rays and 25-MV x rays. Med Phys. 1979;6(4):285–90.

[30] Purdy JA. Buildup/surface dose and exit dose measurements for a 6-MV linear accelerator. Med Phys. 1986;13(2):259–62.

[31] Klein EE, Purdy JA. Entrance and exit dose regions for a Clinac-2100C. Int J Radiat Oncol Biol Phys. 1993;27(2):429–35.

[32] Kron T, Butson M, Hunt F, Denham J. TLD extrapolation for skin dose determination in vivo. Radiother Oncol. 1996;41:119–123.

[33] Butson MJ, Cheung T, Yu PKN, Currie M. Surface dose extrapolation measurements with radiographic film. Phys Med Biol. 2004;49(13):N197–N201.

[34] Nilsson B, Montelius A. Fluence perturbation in photon beams under nonequilibrium conditions. Med Phys. 1986;13(2):191–195.

[35] Shea EO, Mccavana P. in Practice Review of surface dose detectors in radiotherapy. J Radiother Pract. 2003;3(2):69–76.

[36] Gerbi BJ, Khan FM. Plane-parallel ionization chamber response in the buildup region of obliquely incident photon beams. Med Phys. 1997;24(6):873–8.

[37] Butson MJ, Yu PKN, Cheung T, Metcalfe P. Radiochromic film for medical radiation dosimetry. Mater Sci Eng R Reports. 2003;41(3–5):61–120.

[38] Chu RDH, Lewis DF, O’Hara KPJ, Buckland BW, Dinelle F, Van Dyk G. GafchromicTM dosimetry media: A new high dose, thin film routine dosimeter and dose mapping tool. Int J Radiat Appl Instrumentation Part C Radiat Phys Chem. 1990;35(4-6):767–773.

[39] Dreindl R, Georg D, Stock M. Radiochromic film dosimetry: considerations on precision and accuracy for EBT2 and EBT3 type films. Zeitschrift fur Medizinische Phys Phys. 2014;24(2):153–63.

[40] Chiu-Tsao ST, Ho Y, Shankar R, Wang L, Harrison LB. Energy dependence of response of new high sensitivity radiochromic films for megavoltage and kilovoltage radiation energies. Med Phys. 2005;32(11):3350–4.

[41] Rink A, Vitkin IA, Jaffray DA. Energy dependence (75 kVp to 18 MV) of radiochromic films assessed using a real-time optical dosimeter. Med Phys. 2007;34(2):458–63.

References

Related documents

Parallellmarknader innebär dock inte en drivkraft för en grön omställning Ökad andel direktförsäljning räddar många lokala producenter och kan tyckas utgöra en drivkraft

Närmare 90 procent av de statliga medlen (intäkter och utgifter) för näringslivets klimatomställning går till generella styrmedel, det vill säga styrmedel som påverkar

I dag uppgår denna del av befolkningen till knappt 4 200 personer och år 2030 beräknas det finnas drygt 4 800 personer i Gällivare kommun som är 65 år eller äldre i

Den förbättrade tillgängligheten berör framför allt boende i områden med en mycket hög eller hög tillgänglighet till tätorter, men även antalet personer med längre än

På många små orter i gles- och landsbygder, där varken några nya apotek eller försälj- ningsställen för receptfria läkemedel har tillkommit, är nätet av

Detta projekt utvecklar policymixen för strategin Smart industri (Näringsdepartementet, 2016a). En av anledningarna till en stark avgränsning är att analysen bygger på djupa

DIN representerar Tyskland i ISO och CEN, och har en permanent plats i ISO:s råd. Det ger dem en bra position för att påverka strategiska frågor inom den internationella

Den här utvecklingen, att både Kina och Indien satsar för att öka antalet kliniska pröv- ningar kan potentiellt sett bidra till att minska antalet kliniska prövningar i Sverige.. Men