University of Groningen
Validation of a pseudo-3D phantom for radiobiological treatment plan verifications
Kartini, Dea Aulia; Sokol, Olga; Wiedemann, Julia; Tinganelli, Walter; Witt, Matthias;
Camazzola, Gianmarco; Kraemer, Michael; Talabnin, Chutima; Kobdaj, Chinorat; Fuss,
Martina C
Published in:
Physics in Medicine and Biology DOI:
10.1088/1361-6560/abb92d
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2021
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Kartini, D. A., Sokol, O., Wiedemann, J., Tinganelli, W., Witt, M., Camazzola, G., Kraemer, M., Talabnin, C., Kobdaj, C., & Fuss, M. C. (2021). Validation of a pseudo-3D phantom for radiobiological treatment plan verifications. Physics in Medicine and Biology, 65(22), [225039]. https://doi.org/10.1088/1361-6560/abb92d
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
PAPER
Validation of a pseudo-3D phantom for radiobiological treatment plan
verifications
To cite this article: D A Kartini et al 2020 Phys. Med. Biol. 65 225039
View the article online for updates and enhancements.
Phys. Med. Biol. 65 (2020) 225039 https://doi.org/10.1088/1361-6560/abb92d
Physics in Medicine & Biology
RECEIVED 9 July 2020 REVISED 31 August 2020 ACCEPTED FOR PUBLICATION 16 September 2020 PUBLISHED 19 November 2020
PAPER
Validation of a pseudo-3D phantom for radiobiological treatment
plan verifications
D A Kartini1,2,3 , O Sokol3 , J Wiedemann3 , W Tinganelli3 , M Witt4 , G Camazzola3 , M Kr¨amer3 , C Talabnin5 , C Kobdaj1,2 and M C Fuss31 School of Physics, Suranaree University of Technology, 111 University Avenue, Muang, Nakhon Ratchasima 30000, Thailand 2 Thailand Center of Excellence in Physics, Ministry of Higher Education, Science, Research and Innovation, 328 Si Ayutthaya Road,
Bangkok 10400, Thailand
3 Biophysics Department, GSI Helmholtzzentrum für Schwerionenforschung GmbH, Planckstrasse 1, 64291 Darmstadt, Germany 4 Marburger Ionenstrahl-Therapie Betriebs-Gesellschaft mbH, Albrecht-Kossel-Strasse 1, 35043 Marburg, Germany
5 School of Chemistry, Suranaree University of Technology, 111 University Avenue, Muang, Nakhon Ratchasima 30000, Thailand E-mail:kobdaj@g.sut.ac.th
Keywords: 3D cell culture, hydrogels, Matrigel, biological dose verification, treatment planning, ion beam therapy
Abstract
Performing realistic and reliable in vitro biological dose verification with good resolution for a
complex treatment plan remains a challenge in particle beam therapy. Here, a new 3D
bio-phantom consisting of 96-well plates containing cells embedded into Matrigel matrix was
investigated as an alternative tool for biological dose verification. Feasibility tests include cell
growth in the Matrigel as well as film dosimetric experiments that rule out the appearance of field
inhomogeneities due to the presence of the well plate irregular structure. The response of CHO-K1
cells in Matrigel to radiation was studied by obtaining survival curves following x-ray and
monoenergetic
12C ion irradiation, which showed increased radioresistance of 3D cell cultures in
Matrigel as compared to a monolayer. Finally, as a proof of concept, a
12C treatment plan was
optimized using in-house treatment planning system TRiP98 for uniform cell survival in a
rectangular volume and employed to irradiate the 3D phantom. Cell survival distribution in the
Matrigel-based phantom was analyzed and compared to cell survival in a reference setup using cell
monolayers. Results of both methods were in good agreement and followed the TRiP98
calculation. Therefore, we conclude that this 3D bio-phantom can be a suitable, accurate
alternative tool for verifying the biological effect calculated by treatment planning systems, which
could be applied to test novel treatment planning approaches involving multiple fields, multiple
ion modalities, complex geometries, or unconventional optimization strategies.
1. Introduction
Ion beam radiotherapy utilizes the physical characteristics (Bragg peak) and enhanced biological
effectiveness as compared to photons or protons to irradiate deep seated tumors while sparing the healthy tissue. In ion beam radiotherapy, the relative biological effectiveness (RBE) is one of the key factors in treatment planning. RBE is a dose ratio between reference radiation such as x-rays and densely ionizing radiation which yields an equal biological effect. When conducting in vitro biological dosimetry i.e.
treatment plan verification by measuring the final distribution of cell inactivation, spatial resolution and the coverage of different areas of a treatment plan remain a challenge. TRiP98, a research treatment planning system developed at GSI Helmholtz Centre for Heavy Ion Research (GSI) for the heavy ion beam therapy patient trial (Kr¨amer et al2000, Kr¨amer and Scholz2000, Kr¨amer2001, J¨akel et al2001), has undergone continuous improvements since then (Kr¨amer and Durante2010, Kr¨amer et al2014, Sokol et al2019). Currently, it allows to optimize multiple radiation fields for cell killing while using a combination of different ions and including the effects of oxygen enhancement ratio. In parallel to these developments, different phantoms have been used at GSI to verify the RBE-weighted dose in ion treatment plans (Kr¨amer et al2003,
Gemmel et al2008, Kr¨amer and Durante2010, Kr¨amer et al2014, Sokol et al2017). Their complexity ranged from rather simple polystyrene slides covered with cell cultures and arranged behind each other at defined depth positions in a medium-filled container to plastic sticks organized into a 2D-pattern by custom-made grids and covered with medium. With increasing sophistication, the laboratory sample preparations and post-irradiation processing increased both in time and complexity, especially since the cells attached only moderately to the respective materials in some cases. This can in some cases lead to a poor reproducibility, ultimately limiting the practical usefulness of the method.
Multiple research groups have used readily available lab-ware such as multi-well plates for their radiobiology experiments (Gemmel et al2010, Gemmel et al2011, Klein et al2017, Patel et al2017, Mein
et al2019) with different purposes. However, the field homogeneity inside 96-well plates has never been subject to a dedicated investigation even though the wells were sometimes not completely filled. Another kind of lab-ware, T-175 cell culture flasks, has been utilized to investigate the cell survival assay and the dose distribution at the entrance to fragmentation tail in one single biological sample (Buglewicz et al2019). However, this setup observed the dose distribution by evaluating the density of stained colonies in T-175 culture flask. Over the past decades, the interest in growing cells in 3D cell culture is increasing (Lin et al
2009, Storch et al2010, Colom et al2014, Caliari and Burdick2016). Here, we explored a different setup based on standard lab materials—multi-well plates filled with Matrigel matrix—which enables cell survival measurements at multiple positions within a rectangular phantom volume with the cells grown in a 3D environment.
Matrigel is a commercial extracellular matrix (ECM) based hydrogel extracted from the
Engelbreth-Holm-Swarm mouse sarcoma, which contains an abundance of ECM proteins. In this study, the CHO-K1 (Chinese hamster ovary) cell culture in Matrigel was adopted as a pragmatic approach to obtain higher cell numbers in each 96-well plate than those achievable in a classical monolayer culture, in which the well surface area acts as a limiting factor under non-confluent conditions and can thereby limit the
measurement of relative survival due to insufficient cell numbers. By using the Matrigel layer thickness in an embedded culture as a third dimension, a significant improvement of the number of cells which could be cultured per well is expected. Furthermore, Matrigel is a step toward a better approximation of an actual biological environment. It provides a large number of growth factors, and the collagen/laminin, proteoglycan etc networks provide structural support that allow cells to proliferate in all directions, to adopt a cellular shape much closer to those in a biological tissue, migrate in three dimensions, and facilitates signal transduction.
To address concerns about the planned setup, extensive initial tests were carried out and are described in Methods. Cell growth was assessed and optimized after seeding in Matrigel and compared to standard monolayer culture. To recover cells from the gel, different protocols were tested, and their suitability regarding the cell numbers and reproducibility was compared. Furthermore, radiochromic film was used to ensure that no difference in absolute dose, variability, or homogeneity were present after irradiation between stacked 96-well plates. In section3(Results), we report whether the radiation response of CHO cells to x-rays and12C ions in Matrigel was compatible with those in monolayers in 96-well plate. Finally, using the
Matrigel setup, we performed a verification of a simple12C treatment plan for a rectangular target.
2. Methods
2.1. Cell culture setups
CHO-K1 cells were chosen for this study because they are easy in handling and GSI has broad experience and reference data in ion irradiation with them. Three different setups for cell culture were used in this study: setup A consisted of cells cultured in Matrigel inside a 96-well plate (flat bottom polystyrene well plate from TPP), setup B consisted of cells cultured in a standard monolayer inside a 96-well plate, and setup C consisted of cells cultured in a monolayer on polystyrene slides arranged in an acrylic glass container (‘stack phantom’). All protocols for preparing the cell culture in different setups are presented below.
2.1.1. Cell culture in Matrigel
Cell culture in setup A was prepared according to the protocol suggested by Corning (Matrigel Matrix 3D In Vitro Protocol2017) with additional modifications. CHO cells were cultured in Matrigel matrix as an embedded culture. For all procedures, Matrigel (from Corning) was diluted with cell culture medium into a concentration of 5 mg ml−1.
The cells were maintained in culture medium (Ham’s F-12 supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin, all from Biochrom) and incubated at standard conditions (37◦C with 5% CO2in humidified atmosphere). Subsequently, the bottom of each well was pre-coated with a layer of
Phys. Med. Biol. 65 (2020) 225039 D Kartini et al
Figure 1. Sketch of the final structure of the Matrigel layer inside a 96-well plate. The light red color represents the Matrigel layers, the orange color represents the mixture layer of cells and Matrigel, and the red color represents the layer of culture medium.
the layer became solid. Next, a cell suspension was prepared by trypsinizing and centrifugation of the cells at 300 g for 5 min at 21◦C. The final cell concentration in the suspension was adjusted to 4× 106cell ml−1. Then, the mixture of cell suspension (5 µl) and Matrigel matrix (45 µl) resulting in final cell concentration of 2× 104cells/well was added on top of the first Matrigel layer and incubated for 1 h. Another layer of Matrigel (16.6 µl) was added to provide more space for cells to proliferate without tearing the Matrigel matrix and incubated for 1 h. For the final layer, 83.3 µl of medium was added to maintain the moisture of the Matrigel matrix. The resulting layered structure inside the 96-well plate is shown in figure1. Matrigel samples were incubated afterwards for 48 h, until reaching the desired cell number. Before irradiation, all wells containing cells were completely filled with medium and closed with micro-plate sealing film.
The same Matrigel protocol was also scaled to fit the size of 384-well plates with the aim to have higher spatial resolution in the biological verification experiments. However, practical difficulties while processing small volumes of the viscous liquid Matrigel caused the gel to adhere frequently to the well walls. As a consequence, finalized samples had an uneven layer structure and reproducible cell growth and recovery were hampered.
To extract the cells from the gel after irradiation, we used Corning dispase. 64 µl of dispase was added to the gel and incubated for 1 h or until the matrix was completely dissolved. Afterwards, the cell suspension was collected into microtubes and centrifuged at 2500 rpm for 5 min. After discarding the supernatant, 300 µl of trypsin (0.05% trypsin/0.02% EDTA in PBS, from PAN Biotech) was added and incubated for 3 min to prevent cell clumping in order to obtain a single cell. Finally, 900 µl of medium was added to stop the trypsin digestion.
In addition to the established cell recovery protocol, an alternative method for cell recovery based on the combination of centrifugation and cooling was tested, following suggestions from the supplier (Corning Matrigel Matrix Frequently Asked Questions2017). In this way, we attempted to avoid the incubation step which might allow cells to partially initiate repair after irradiation. First, the well plate was cooled in the 4◦C refrigerator for 1 h to liquefy the Matrigel matrix. Subsequently, the Matrigel matrix was collected into microtubes and centrifuged at 2500 rpm for 5 min to separate the Matrigel matrix from the cell suspension. Supernatant was discarded. Next, trypsin was added to detach cell clumping and cells were incubated for 3 min. Finally, we added medium to stop the trypsin digestion and counted the cell number using a Beckman Coulter counter. We seeded the cells in Matrigel and recovered them after 48 h of incubation time using the combination method. This alternative protocol yielded inconsistent cell numbers (uncertainty of±50%) and the average number of recovered cells was 62 256 per well, much less than for the standard protocol (c.f. table2). In consequence, we decided to use the dispase method, as described above, to recover cells from Matrigel matrix.
2.1.2. Seeding in monolayers
Cell cultures in setup B and setup C were prepared following the standard monolayer culture procedures, with modifications to adjust to the size of a 96-well plate and polystyrene slides. For culturing the cells in setup B, approximately 280 µl of cell suspension containing 1.1× 104cells per well was seeded into each well of a 96-well plate. Subsequently, all samples were incubated for 24 h prior to irradiation. To recover the cells, trypsin (50 µl) was added to the wells, and the cells were incubated for 5 min or until cell detachment was
Figure 2. Experimental setup for Gafchromic EBT3 film measurements. Four pieces of EBT3 film were placed before, between and after the 96-well plates.
observed under a microscope. Afterwards, 1 ml of medium was added to stop the trypsin reaction, and cells were collected into microtubes for counting.
Cell culture in setup C was performed according to the following protocol as described before (Els¨asser
et al2010). Cell culture treated polystyrene slides imitating the bottoms of the 25 cm2tissue culture flask
(TCF) (produced as special order from Greiner) were placed inside squared bio-assay dishes to ensure easy handling and transportation. Next, cell suspensions (0.5 ml) containing 5× 104cells were center plated on
each slide and incubated for 2–4 h to let the cells attach. Afterwards, the dishes were filled with medium until all the slides were covered, and all samples were incubated for 24 h prior to irradiation. For cell recovery, the slide surface was rinsed with 1 ml of PBS Dulbecco (Biochrom) and covered with trypsin (1 ml) and the cells were incubated for 4–5 min. Finally, the cells were collected into a tube filled with medium (4 ml).
2.2. Clonogenic assay
After irradiation, cells were recovered as described in section2.1and counted using a Beckman Coulter counter to reveal the cell number in each sample. After that, cells were re-seeded in 25 cm2Falcon TCF with
vented caps in 5 ml of cell culture medium. The cell number were adjusted to the expected plating efficiency in the flask (70%) and to the expected survival (depending on the dose) in order to obtain approximately 100 colonies after a week of incubation. Cells were reseeded into triplicates for each irradiated sample. All samples were incubated for one week to allow cells to form colonies. After one week, colonies were stained with methylene blue staining solution, and colonies containing more than 50 cells were considered as viable and were counted.
2.3. Radiochromic film measurements
Film dosimetric tests were carried out to assess the radiation field homogeneity in presence of the well plates and the possible influence of the plate geometry. A phantom was assembled by stacking together three well plates previously filled up with water completely (wells and spaces between them, see figure2) and closed with micro-plate sealing film. Gafchromic EBT3 film (Ashland, RI) was cut to fit the size of a 96-well plate and placed at four different depth positions around and between these plates without further build-up. The films were then irradiated in monoenergetic, uniform12C ion fields at incident beam energies of 145 MeV/u
or 196 MeV/u and additionally in a mixed field, representing simple and well-defined conditions. The doses ranged between 1.4 and 5 Gy depending on the initial energy and exact depth.
The films were scanned 24 h post-irradiation with an Epson flatbed film scanner model ‘Perfection V800 Photo’ in vertical (‘portrait’) orientation, 48-bit rgb color mode, resolution of 200 dpi, and with all image enhancement features disabled to ensure a fully reproducible scanning process. The profiles of signal values recorded by the scanner CCD (‘pixel values’) across the radiation fields downstream of one, two or three well plates and the associated standard deviations were then inspected at different locations for uniformity. For all radiochromic films, only data from the red color channel were analyzed using ImageJ developed by National Institutes of Health (NIH).
Phys. Med. Biol. 65 (2020) 225039 D Kartini et al
Table 1. Irradiation systems employed in this study.
Experiment Beam Energy (MeV/u) Dose range (Gy) Facility
Survival curve X-rays 250 kV 0–11 GSI
Survival curve 12C 180 0– 8 MIT
Treatment plan verification 12C 156.34–211.54 0–6.5 MIT
EBT3 film measurement 12C 145, 195 1.4–5 MIT
2.4. Treatment planning system
The calculation of the dose distribution and prediction of the resulting cell survival for all treatment plans were conducted using the GSI in-house treatment planning system TRiP98. Depending on the phantom used (‘stack’ phantom or 3D-well plate phantom), the cubic target volume had lateral dimensions of 40× 80 mm2 or 60× 40 mm2, respectively, for optimal coverage of the samples, ranging from 55 to 85 mm depth in both cases. Both plans were optimized for a uniform RBE-weighted dose of 6.5 Gy (corresponding to a constant survival rate in the target of ~10%). Different RBE tables based on the corresponding reference (x-ray) survival curve for monolayer and Matrigel samples, respectively (see section3.3), were used to compute the expected cell survival distribution for the radiobiology verification experiment. The CHO Matrigel RBE table was used for the well plate treatment plan, while the standard CHO RBE table (Sokol et al2017) was used for the stack phantom treatment plan.
2.5. Irradiation facilities
In this study, x-rays and monoenergetic12C ion beams were employed in the survival curve experiments, while a 3D actively scanned12C ion beam was employed in the extended target irradiation. For EBT3 film
measurement, monoenergetic12C ion beam with different energies were used to observe the film response to
different energy ranges. The x-ray irradiations were performed at GSI, using an Isovolt DS1 x-rays machine at a dose rate of 2.2 Gy min−1and a peak voltage of 250 kV. The cell samples received doses in the range 0–11 Gy, which were measured with a Farmer-type ionization chamber placed directly below the cell samples.
Carbon ion irradiations were all performed at Marburger Ionenstrahl-Therapiezentrum (MIT, Marburg, Germany). The beam energies used in the different experiments are summarized in table1. Dose
measurements for monoenergetic12C ion beam irradiations were carried out with a Farmer-type ionization chamber placed in the entrance channel for selected fields and the ratio (measured dose / dose expected according to TRiP plan) was applied as a calibration factor to the absorbed dose for all fields at that energy. For one of the extended target irradiations using active scanning, dosimetry was performed inside a water phantom containing a pinpoint chamber (PTW type 31 015) array to verify the central depth dose distribution. The array block consisted of 24 ionization chambers with an outer radius of 1.45 mm and a sensitive volume of 0.03 cm2each. Chambers were arranged in 6 staggered rows of 4 chambers each, covering
5 cm in z-direction and 3 heights (isocenter plane,±7 mm), to avoid disturbing the radiation field. All calibrations were performed in absorbed dose to water, according to the TRS398 protocol (Andreo et al
2000).
2.6. Survival experiment setups
The survival experiment with x-rays was performed using setup A while survival experiment with
monoenergetic12C ion beam was performed using both setup A and setup B. In the survival experiment with x-rays, we prepared eight sample plates including controls. In each 96-well plate, cells were seeded in three wells and irradiated with the same dose. Closed sample plates were placed under the x-ray source and irradiated without further build-up from the top. For the survival experiment with12C ion beam, we
prepared six samples including a control sample in one 96-well plate and irradiated each sample with a different dose. Two independent sample plates were prepared according to section2.1.1, sealed and irradiated in vertical orientation from the plate’s bottom without further build-up.
2.7. Treatment plan verification setups
The irradiation setup for radiobiology verification with setup A was prepared as below. Cells were seeded into five different wells in each 96-well plate as shown in figure3(b). To achieve a homogeneous target geometry, all wells containing cells were completely filled with medium and closed with micro-plate sealing film. After that, 96-well plates were placed vertically and irradiated at four depth positions to obtain the cell survival at the respective position as shown in figure3(a). For each position, two independent sample plates were irradiated and analyzed.
At position 1 which corresponds to the beam entrance channel, we placed the 96-well plates filled with cell samples at a depth of 26.8 mm after additional 23.8 mm water-equivalent thickness (WET) of
Figure 3. Experimental setup for radiobiology verification with Matrigel and 96-well plate (setup (A)). (a) Four 96-well plates were placed in different locations corresponding to entrance channel, two extended target positions, and fragmentation tail region. (b) Cells were seeded in five different wells inside 96-well plate. Four wells (C)-(F) were located inside the target region (red box) while one well (A) was kept outside the target region, mimicking healthy tissue.
Figure 4. Shape of CHO-K1 cells cultured in (a) monolayer after 24 h of incubation, (b) Matrigel after 24 h of incubation, and (c) Matrigel after 48 h of incubation. The shape of cells were observed under the inverted compound microscope.
polymethyl methacrylate (PMMA). The 3 mm difference in thickness correspond to the WET of the polystyrene bottom + 1 mm + 0.75 mm of Matrigel, therefore the depth cited corresponds to the center of the Matrigel layer with cells embedded. At position 2, we placed the well plates at depth of 62.5 mm following a 59.5 mm water-equivalent of PMMA to observe the cell survival in the proximal part of the target. In order to measure cell survival in the distal parts of the target region (position 3), PMMA slabs were combined to give a total water-equivalent depth of 86.3 mm. For position 4, we placed the well plate with cell samples at the depth of 113.9 mm after PMMA slabs and two water-filled well plates for build-up to observe cell survival in the fragmentation tail. For irradiation with setup C, cells were seeded on specific slides at the same depths corresponding to the entrance channel (26.5 mm), two extended target positions (61.5 and 86.5 mm), and fragmentation tail (111.5 mm), respectively. Again, two independent phantoms were prepared and irradiated.
3. Results
3.1. Cell numbers in Matrigel and monolayer cultures
In order to investigate the cell behavior in 3D environment, cell growth and shape were observed over time and compared to conventional cell culture in a monolayer. After 24 h in culture, cells in the monolayer exhibited an elongated shape (figure4(a)), while cells in the Matrigel remained spherical (figure4(b)). After additional 24 h of incubation (48 h in total), the cells in Matrigel became more elongated, stretching in different directions, which demonstrates their interaction with ECM (see figure4(c)). While the cells in monolayer culture became confluent after 48 h—making them not suitable for irradiation due to a possible synchronization in cell cycle—the cells in Matrigel did not show any apparent restrictions to proliferate and/or migrate.
Phys. Med. Biol. 65 (2020) 225039 D Kartini et al
Table 2. Comparison of cell number obtained from different cell culture methods and different incubation times. The uncertainties stated correspond to one standard deviation of the sample.
Incubation time
Cell culture 24 h 48 h
Monolayer 32 596± 17% —
Matrigel — 158 517± 8.7%
Table 3. Standard deviations of Gafchromic film profile results in presence of 96-well plates.
Film sample 145 MeV/u 196 MeV/u Mixed field
Uniform 269 (0.93%) 260 (0.88%) 306 (1.16%)
Film 0 261 (0.90%) 302 (1.05%) 280 (1.08%)
Film 1 260 (0.94%) 285 (1.00%) 317 (1.26%)
Film 2 242 (0.92%) 312 (1.12%) 255 (1.05%)
Film 3 334 (1.40%) - 281 (1.18%)
To test the reproducibility of cell growth and recovery, we seeded 24 samples of cells in Matrigel and 22 samples as a monolayer inside 96-well plates and counted the cell number after samples were incubated for 24 h (monolayer) and 48 h (Matrigel). Cells were recovered according to the protocol in section2.1and all cells were counted using a Beckman Coulter cell counter. Table2shows that obtained cell number in Matrigel after 48 h in comparison to cell number in monolayer after 24 h, both cell cultures were done inside 96-well plate. The obtained cell numbers in monolayer presented a larger uncertainty than cells cultured in Matrigel.
3.2. Gafchromic EBT3 film response
In order to assess the influence of stacked, liquid-filled well plates on the radiation field quality,
representative lateral profiles of the scanner signal readout were analyzed. The results for films in front of the plates as well as downstream after one, two and three plates are shown in figure5(a)–(c) and are compared to the profiles obtained after exposure to a similar dose in a uniform field, with a homogeneous absorber (a 2 mm slab of PMMA) for build-up. The profiles are shown in their original resolution and after applying a simple smoothing filter which reduces the statistical noise present in the data. For up to two well plates before the film, no alterations were detected in the lateral profiles of two monoenergetic12C fields and one
mixed field. On the contrary, the film placed downstream of three well plates in a field of 145 MeV/u presented visible artefacts (marked in figure5by arrows) with a periodicity of ~9 mm, corresponding to the spacing between the single wells. These are also visible as ring patterns on the irradiated film and match the well walls. The extra dose estimated based on the increased WET (+1.3 mm) of these polystyrene walls in the corresponding depth is 5 %. This pattern is reflected as well in the standard deviation across the profile of 1.4% compared to 0.9%–0.94% for the uniform case and up to two well plates. The mixed field yielded slightly higher standard deviations generally (1.05%–1.26%), but no particular trend with the number of plates, indicating a good and persisting field homogeneity for up to three 96-well plates. Table3summarizes the standard deviations of the pixel values across the unfiltered profiles shown, for the different12C ion fields used and different number of well plates present.
3.3. Survival curves
To observe the shape of the survival curve and the radio-sensitivity of cells cultured in Matrigel, cells were irradiated with x-rays and recovered with dispase. The measured survival curves in comparison with the recent reference survival curve for the monolayer are presented in figure6. Matrigel curves demonstrated a slight increase in radio-resistance compared to monolayer curve in the low-dose region. By using the linear-quadratic fit (Hall and Giaccia2012), we obtained new αxMGand βxMGparameters, which will be used to calculate the new RBE table (Scholz et al1997) for CHO cells cultured in Matrigel.
Figure7shows the measured survival curves for Matrigel and monolayer samples irradiated with monoenergetic12C beams with energy of 180 MeV/u in comparison with the respective x-ray survival curves. All measured data points were plotted in semi-logarithmic scale and fitted with the linear-quadratic equation (Hall and Giaccia2012). The curve for the Matrigel samples irradiated with 180 MeV/u12C beams (dark blue
points) is comparable with the monolayer curve (light blue points), especially at lower doses.
At a dose of 8 Gy, an unexpectedly large number of cell colonies was counted implying there was a high possibility of overlap between cell colonies, resulting in unreliable results. As a consequence, the respective data points from Matrigel and monolayer at 8 Gy, represent a lower estimate rather than the best guess and
Figure 5. Homogeneity analysis with Gafchromic film. (a), (b) Film profiles across the center of a 96-well plate irradiated in monoenergetic C ion fields. (c) Profiles across the center of a 96-well plate irradiated in a mixed field.‘Film n’ designates the film sheet placed after n well plates, and ‘uniform’ refers to a piece of film irradiated in a uniform field without any plate. The profiles are shown in native scanner resolution (light-colored lines) and after applying an averaging filter (solid lines; 7-point equivalent to 0.9 mm) for a better appreciation of the macroscopic effects. The arrows in (a) point toward the artefact pattern obtained with film 3.
Figure 6. Survival curves of CHO cells in Matrigel after x-ray irradiation. Triangle symbols and error bars represent the average and the standard deviation of three Matrigel samples irradiated at once. The red line represents recent survival curve of cells in monolayer (Sokol et al2017). All data were plotted in semi-logarithmic scale and fitted using a linear-quadratic equation: S = exp(−αD − βD2).
were not included in the fitting lines. All α and β parameters obtained from different irradiation setups are presented in table4. In contrast to the monolayer case where the α/β ratio is around 11, α/β in Matrigel culture for both x-rays and12C curves showed a ratio ranging from 2 to 4 Gy which might indicate that the
Phys. Med. Biol. 65 (2020) 225039 D Kartini et al
Figure 7. Comparison of CHO cells survival curves in different irradiation setups. All symbols and error bars represent the average and the standard deviation of two independent repetitions. Symbols (△) represent cells in Matrigel (MG) and monolayer (MO) irradiated with monoenergetic12C beam of 180 MeV/u, while the green line and orange line represent x-rays survival
curves in Matrigel and monolayer (2017), respectively. Dashed lines represent the fit of data. All data were fitted with a linear-quadratic equation: S = exp(−αD − βD2).
Table 4. α and β parameters obtained from the survival curves for all irradiation combinations.
Survival curve α(Gy−1) β(Gy−2) α/β (Gy)
X-rays monolayer (Sokol et al2017) −0.216 ± 0.031 −0.019 ± 0.004 11.11
X-rays Matrigel −0.083 ± 0.010 −0.034 ± 0.001 2.435
12C 180 MeV/u monolayer −0.272 ± 0.041 −0.025 ± 0.005 10.94
12C 180 MeV/u Matrigel −0.159 ± 0.021 −0.041 ± 0.003 3.924
3.4. Verification of a TRiP98 treatment plan
We performed the irradiation of the Matrigel-based system with the scanned12C ion beam at different depths to assess its applicability for the verification of more complex three-dimensional treatment plans. The cell samples were seeded in five different wells of a 96-well plate. Four wells on the plate were located in the beam path (inside the field), while one well was kept outside the field for control measurements (see figure3). To evaluate the accuracy of the measurements, the measured cell survival of the 3D bio-phantom was compared to the one of the stack phantom (setup C), which is a standard tool for radiobiology verification at GSI.
The measured cell survival distributions along the beam path for both phantoms in comparison to the TRiP98 calculations are shown in figure8. The TRiP98 calculation of the well plate plan (red line) yielded a slightly elevated survival curve with respect to the stack plan (blue line), especially in the entrance channel and target regions. This difference is caused by the slightly increased radio-resistance of cells in 3D culture consistent with the x-ray results which are used to calculate the Matrigel RBE table.
In the target regions, we have consistent cell survival for both well plate and stack setup. The average survival obtained in target is 0.0919± 14% for Matrigel setup and 0.0716 ± 8.8% for stack setup. In the entrance channel and fragmentation tail regions, we have a similar spread in data points for the cells in the Matrigel setup, and the average survival values obtained are 0.364± 12% and 0.922 ± 11%, respectively. Meanwhile, the average survival measured in stack setup are 0.350± 5.9% and 0.942 ± 1.8%, respectively.
The measured survival from the 96-well plate along the x- and y-axes were also plotted as shown in figure9with the aim to verify the consistency of measurements in different wells within the target region (see section2.7). All the measurements in the lateral distribution give consistent results (figure9) which are in good agreement with the TRiP98 calculation.
Biologically optimized treatment plans were applied in all experiments; this implies that physical dose is not constant along the extended target. Figure10shows the measured physical dose for the biologically optimized plan (using the ‘Matrigel’ RBE table) for the extended target irradiation in the well plate phantom with12C ions. The experimental values given are the average over four pinpoint chamber readings at a given
depth z. The dosimetry results are in very good agreement with the TRiP98 calculation for the present well plate plan.
Figure 8. Cell survival distribution in depth, measured for Matrigel and monolayer setups. Symbols (△) represent cells cultured in Matrigel inside a 96-well plate, while symbols (⋄) represent cells cultured in a monolayer in the stack phantom. All data were plotted in semi-logarithmic scale. All error bars represent the standard deviation of the triplicate of each sample. The red and blue lines represent TRiP98 survival predictions in the well plate plan and in the stack plan, respectively.
4. Discussion
Basic cell tests including the measurements of cell numbers, observation of the cell shape and cell recovery in Matrigel matrix have been conducted in this study with the aim to optimize a new laboratory protocol for biological treatment plan verification experiments. CHO-K1 cells were cultured in Matrigel matrix inside 96-well plates and shape and growth reproducibility was studied. Cells in Matrigel displayed a spread and elongated shape due to their interaction with ECM and we obtained a four times higher cell number compared to ‘flat-growing’ cells in monolayers on the same surface area. A large number of cells obtained is important for quantifying low cell survival level (<1% of survival level). Based on the cell number
reproducibility and the convenient processing steps, dispase was preferred for cell recovery from Matrigel for any quantitative analysis or procedures.
The survival experiment with x-rays presents the comparison of cell response in Matrigel and monolayer which can be observed from the shape of the survival curves (see figure6). The survival curve of cells in Matrigel displayed a slightly elevated curve with a flat response at low dose and a steeper slope at high doses, indicating that cells in a 3D environment are initially less sensitive to irradiation than cells in a monolayer but become more sensitive at higher doses. A different impact of cell culture in 3D environment from higher radioresistance to slightly higher radiosensitivity has been observed in previous studies (Storch et al2010, Eke et al2016, Gomez-Roman et al2017). Because of the differences between the cell response in monolayer and Matrigel, we extracted the linear and quaratic parameters αxMGand βxMGfrom the Matrigel dose response curve to calculate a dedicated RBE table for CHO cells in Matrigel which was subsequently used for treatment planning. We also performed a survival experiment with monoenergetic12C ion beams and the obtained dose response of cells in Matrigel was comparable to the response of the cells irradiated in
monolayers. However, a slight tendency to a broader shoulder for MG samples and a crossing of both curves at higher doses goes in the same direction as obtained with x-ray irradiation.
The radiation field homogeneity analysis carried out with Gafchromic EBT3 supports the use of the relatively small 96-well plates for radiobiological verification purposes. The films exposed in between and downstream of up to two well plates, in different incident12C ion fields, showed no specific alteration of the
field homogeneity, no increased standard deviation of readings, and no spatial patterns linked to the presence of the plates. A third well plate upstream of the film sample caused artefacts in the lateral film profiles for a monoenergetic field which clearly correspond to the spacing of the wells (9 mm). This can be explained by the cumulative effect of the polystyrene well walls (additional WET of 0.44 mm per plate), which have a density (1.04 g cm−3) close to but not equal to water and can locally modify range, LET, and ultimately dose deposition. A smaller influence is expected in mixed fields in macroscopic target volumes (not for single pristine Bragg peaks), which is consistent with the present results (no artefact structure found in the mixed field). Furthermore, the ring pattern observed (for three plates upstream) on the films would come to lie
Phys. Med. Biol. 65 (2020) 225039 D Kartini et al
Figure 9. Distribution of measured cell survival in x- and y-axes at different z depths: (a) entrance channel at 26.8 mm, (b) frontal part of target at 62.5 mm, (c) distal part of target at 86.3 mm, and (d) fragmentation tail at 113.9 mm. Symbols (*) represent the measured cell survival of two independent repetitions except two data points in frontal and distal part of target. The color scale displays the TRiP98 survival calculation.
essentially on top of the next layer’s well walls. For the cell samples, any effect is therefore expected to be smaller than that measured on the films.
The radiobiological verification performed with 3D actively scanned12C ion beams supports the
applicability of the 3D bio-phantom as a verification tool. Using the Matrigel-based setup, we achieved the level of precision of cell survival measurements comparable to those involving the GSI stack phantom. Furthermore, the measured cell survival in lateral cuts at all depths showed a good agreement with TRiP98 calculation where all samples in target region have uniform survival level. In the entrance channel and in the fragmentation tail, Matrigel setup exhibits more disperse data points in comparison to the stack setup where the measurements were more consistent. This might arise due to some cells clumping together during the counting step before re-seeding. The cells clumping would presumably have a larger effect on higher survival levels, where the majority of cells survive to form colonies, than for lower survival level (as in the target area), where a large fraction of cells are unviable. There is also a possibility that the cells in Matrigel exhibit a hypoxic condition (Colom et al2014), which would lead to an increased cell radio-resistance, manifesting in the survival curves (figures6and7). We would suggest to use a cell line with high radiation sensitivity in order to investigate the cell survival in regions with low dose and survival close to one, such as the fragmentation tail, and cell survival in lateral directions.
Figure 10. Depth distribution of the physical (absorbed) dose measured in a water phantom for the TRiP plan optimized for a uniform RBE-weighted dose of 6.5 Gy for Matrigel-embedded cells in a target volume at a depth of 50–90 mm. Symbols (⃝) represent pinpoint ionization chamber measurements, and the solid line depicts the TRiP98 calculation along the central line through the isocenter. The uncertainty (standard deviation) of the measured values is smaller than the symbol size and can therefore not be distinguished.
An attempt to use 384-well plates for verification has been conducted as well by irradiating two duplicate samples with several wells positioned in the distal part of the target area. Compared to 96-well plates, we obtained very different survival values and larger uncertainties when using the same treatment plan, which can be potentially explained by the pronounced practical difficulties and the irregular samples obtained. Therefore, we currently find it complicated to move to a setup with a higher spatial resolution using 384-well plates.
5. Conclusions
We investigated the practical feasibility and performance (accuracy and reproducibility) of a simple setup based on stacked 96-well plates to verify the RBE-weighted dose delivered by a treatment plan in a spatially resolved manner. First, through film dosimetry tests, we quantified the physical homogeneity of the radiation field to assess the effect of the small, periodic structures of the well plates. According to the results and from the dosimetric point of view, it is acceptable to stack up to two well plates together for simultaneous irradiation in the entrance channel. Cell culture and recovery protocols were adapted to yield high and constant cell numbers while using standard labware. The response of CHO cells in Matrigel to x-rays and12C ions irradiation was shown to be slightly less sensitive than when irradiated in a monolayer. Finally, we performed a proof-of-principle experiment where all samples in the target received a flat RBE-weighted dose that allowed a straightforward comparison of the radiobiological survival in this new setup to the predicted cell survival. In view of the agreement with our reference method and the reproducibility achieved among equivalent samples, this method is considered a promising tool for validating exemplary cases of novel treatment planning techniques that include complex fluence and LET variations produced by different radiation fields, or for discovering potential systematic deviations from the expected RBE. Furthermore, we would like to emphasize that the cells are actually in a 3D structure which is more realistic than the
monolayer. This makes the present Matrigel-based phantom a suitable tool for radiobiological cell survival verification experiments.
Acknowledgments
D A Kartini and C Kobdaj are funded by Thailand Center of Excellence in Physics (ThEP) under Grant No. ThEP-61-PHM-SUT4 and SUT-Research Center for Theoretical Physics. D A Kartini is also funded by Suranaree University of Technology (SUT) under SUT-PhD Scholarship for ASEAN. O Sokol is funded by the European Union’s H2020 Research and Innovation Programme, under Grant Agreement No. 730983. The authors acknowledge funding and provision of beam time through the Förderprogramm MIT-Forschung of Philipps-Universit¨at Marburg under grant number MIT-2018-03 funded by Hessisches Ministerium für
Phys. Med. Biol. 65 (2020) 225039 D Kartini et al
Wissenschaft und Kunst. The authors would like to thank Dr Ulrike Schötz for kindly providing her lab infrastructure during the experiments at MIT, Dr Thomas Friedrich who kindly provided the RBE table in Matrigel, and Dr Andreas Maier for designing and 3D-printing sample holders for the well plate setups.
ORCID iDs
D A Kartinihttps://orcid.org/0000-0002-3578-0655 G Camazzolahttps://orcid.org/0000-0002-8204-7763 C Kobdajhttps://orcid.org/0000-0001-7296-5248 M C Fusshttps://orcid.org/0000-0001-5332-6491References
Andreo P, Burns D T, Hohlfeld K, Huq M S, Kanai T, Laitano F, Smythe V G and Vynckier S 2000 Absorbed dose determination in external beam radiotherapy Technical Report 398 Int. Atomic Energy Agency Vienna, Austria
Buglewicz D J, Banks A B, Hirakawa H, Fujimori A and Kato T A 2019 Monoenergetic 290 MeV/n carbon-ion beam biological lethal dose distribution surrounding the Bragg peak Sci. Rep.9 1–9
Caliari S R and Burdick J A 2016 A practical guide to hydrogels for cell culture Nat. Methods13 405
Colom A, Galgoczy R, Almendros I, Xaubet A, Farr´e R and Alcaraz J 2014 Oxygen diffusion and consumption in extracellular matrix gels: implications for designing three-dimensional cultures J. Biomed. Mater. Res. A102 2776–84
Corning Matrigel Matrix Frequently Asked Questions 2017 (https://www.corning.com/catalog/cls/documents/faqs/ CLS-DL-CC-026.pdf)
Eke I, Hehlgans S, Sandfort V and Cordes N 2016 3D matrix-based cell cultures: automated analysis of tumor cell survival and proliferation Int. J. Oncol.48 313–21
Els¨asser T et al 2010 Quantification of the relative biological effectiveness for ion beam radiotherapy: direct experimental comparison of proton and carbon ion beams and a novel approach for treatment planning Int. J. Radiat.78 1177–83
Gemmel A, Bert C, Saito N, Von Neubeck C, Iancu G, Weyrather W K, Durante M and Rietzel E 2010 Development and performance evaluation of a dynamic phantom for biological dosimetry of moving targets Phys. Med. Biol.55 2997
Gemmel A, Hasch B, Ellerbrock M, Weyrather W K and Kr¨amer M 2008 Biological dose optimization with multiple ion fields Phys. Med. Biol.53 6991–7012
Gemmel A, Rietzel E, Kraft G, Durante M and Bert C 2011 Calculation and experimental verification of the RBE-weighted dose for scanned ion beams in the presence of target motion Phys. Med. Biol.56 7337
Gomez-Roman N, Stevenson K, Gilmour L, Hamilton G and Chalmers A J 2017 A novel 3D human glioblastoma cell culture system for modeling drug and radiation responses Neuro-oncol.19 229–41
Hall E and Giaccia A 2012 Radiobiology for the Radiologist (Philadephia, PA: Wolters Kluwer Health/Lippincott Williams & Wilkins) J¨akel O, Kr¨amer M, Karger C and Debus J 2001 Treatment planning for heavy ion radiotherapy: clinical implementation and application
Phys. Med. Biol.46 1101
Klein C et al 2017 Overcoming hypoxia-induced tumor radioresistance in non-small cell lung cancer by targeting DNA-dependent protein kinase in combination with carbon ion irradiation Radiat. Oncol.12 208
Kr¨amer M 2001 Treatment planning for heavy-ion radiotherapy: biological optimization of multiple beam ports J. Radiat. Res.42 39–46 Kr¨amer M and Durante M 2010 Ion beam transport calculations and treatment plans in particle therapy Eur. Phys. J. D60 195 Kr¨amer M, J¨akel O, Haberer T, Kraft G, Schardt D and Weber U 2000 Treatment planning for heavy-ion radiotherapy: physical beam
model and dose optimization Phys. Med. Biol.45 3299
Kr¨amer M and Scholz M 2000 Treatment planning for heavy-ion radiotherapy: calculation and optimization of biologically effective dose Phys. Med. Biol.45 3319
Kr¨amer M, Scifoni E, Schmitz F, Sokol O and Durante M 2014 Overview of recent advances in treatment planning for ion beam radiotherapy ion beam transport calculations and treatment plans in particle therapy Eur. Phys. J. D68 306
Kr¨amer M, Wang J F and Weyrather W 2003 Biological dosimetry of complex ion radiation fields Phys. Med. Biol.48 2063–70 Lin Y F, Nagasawa H, Peng Y, Chuang E Y and Bedford J S 2009 Mammary epithelial cells cultured as 2D monolayers or 3D acinar
stuctures in Matrigel Radiat. Res.171 708–15
Matrigel Matrix 3D In Vitro Protocol 2017 (https://www.corning.com/catalog/cls/documents/application-notes/Application_ Note_CLS-DL-AN-414_Matrigel_Matrix_3D_In_Vitro_Protocol.pdf)
Mein S et al 2019 Biophysical modeling and experimental validation of relative biological effectiveness (RBE) for 4 He ion beam therapy Radiat. Oncol.14 123
Patel D et al 2017 Optimization of Monte Carlo particle transport parameters and validation of a novel high throughput experimental setup to measure the biological effects of particle beams Med. Phys.44 6061–73
Scholz M, Kellerer A, Kraft-Weyrather W and Kraft G 1997 Computation of cell survival in heavy ion beams for therapy Radiat. Environ. Biophys.36 59–66
Sokol O et al 2017 Oxygen beams for therapy: advanced biological treatment planning and experimental verification Phys. Med. Biol.62 7798–7813
Sokol O, Kr¨amer M, Hild S, Durante M and Scifoni E 2019 Kill painting of hypoxic tumors with multiple ion beams Phys. Med. Biol.64 045008
Storch K, Eke I, Borgmann K, Krause M, Richter C, Becker K, Schröck E and Cordes N 2010 Tree-dimensional cell growth confers radioresistance by chromatin density modification Cancer Res.70 3925–34
