Dosimetric evaluation of a novel modular cell irradiation platform for multi-modality in vitro studies including high dose rate brachytherapy

Abstract

PURPOSE:

High dose-rate (HDR) brachytherapy is integral for the treatment of numerous cancers. Preclinical studies involving HDR brachytherapy are limited. We aimed to describe a novel platform allowing multi-modality studies with clinical HDR brachytherapy and external beam irradiators, establish baseline dosimetry standard of a preclinical orthovoltage irradiator, to determine accurate dosimetric methods.

METHODS:

A dosimetric assessment of a commercial preclinical irradiator was performed establishing the baseline dosimetry goals for clinical irradiators. A 3D printed platform was then constructed with 14 brachytherapy channels at 1cm spacing to accommodate a standard tissue culture plate at a source-to-cell distance (SCD) of 1 cm or 0.4 cm. 4-Gy CT-based treatment plans were created in clinical treatment planning software and delivered to 96-well tissue culture plates using an Ir192 source or a clinical linear accelerator. Standard calculation models for HDR brachytherapy and external beam were compared to corresponding deterministic model-based dose calculation algorithms (MBDCAs). Agreement between predicted and measured dose was assessed with 2D-gamma passing rates to determine the best planning methodology.

RESULTS:

Mean ( ±standard deviation) and median dose measured across the plate for the preclinical irradiator was 423.7 ± 8.5 cGy and 430.0 cGy. Mean percentage differences between standard and MBDCA dose calculations were 9.4% (HDR, 1 cm SCD), 0.43% (HDR, 0.4 cm SCD), and 2.4% (EBRT). Predicted and measured dose agreement was highest for MBDCAs for all modalities.

CONCLUSION:

A 3D-printed tissue culture platform can be used for multi-modality irradiation studies with great accuracy. This tool will facilitate preclinical studies to reveal biologic differences between clinically relevant radiation modalities.

Introduction

Despite the long history of high dose-rate (HDR) brachytherapy as a therapeutic modality, there are few preclinical HDR brachytherapy studies incorporating clinical irradiators and techniques. While there are few preclinical studies using clinical HDR brachytherapy remote afterloaders, there are numerous studies investigating the biological effects of dose rate with results varying by dose rate and cell lineage among other things (1). In fact, one such study shows evidence for a biological difference between HDR brachytherapy and external beam radiation therapy (EBRT) linear accelerators (2). Historically, preclinical in vitro and in vivo studies have predominantly relied on ⁶⁰Co or ¹³⁷Cs sources (3-5). More recent preclinical studies mostly utilized orthovoltage X-ray irradiators, which do not mimic clinically relevant dose distributions, source characteristics, and dose rate of HDR treatment delivery systems and EBRT linear accelerators (6-10). Even when a proper calibration protocol such as that from the American Association of Physicists in Medicine (AAPM) Task Group Report Number 61 (11) is implemented, preclinical irradiators are subject to large dosimetric error (7-9,12-16). In particular, insufficient scatter conditions can result in a 50% dose difference depending on experimental setup with a 10% dose difference attributable to the choice in single-well cell dish size (15). These large dose differences are seen even when following standard use protocols. Despite these uncertainties, preclinical orthovoltage irradiators are still commonly relied upon for in vitro studies.

In light of this, other groups have developed systems for performing preclinical brachytherapy studies that utilize clinical irradiation sources including therapeutic HDR brachytherapy sources or linear accelerators to elicit more clinically relevant dosimetry (17-21). Most groups utilized machined acrylic (Poly(methyl methacrylate)) platforms with one or more source channels in their studies (17,19-21). Acrylic has certain disadvantages including its limited adaptability to other irradiation geometries, modalities, or experimental setups. Given this, a 3D printed platform would offer certain advantages to using a machined acrylic platform, one of which would be the ease of designing and printing components to adapt a base platform to various experimental designs. With the exception of one study (21), heterogeneity corrections were not reported in any of the aforementioned in vitro studies, but as high-lighted by Chen et al., such corrections are essential for meaningful in vitro studies with clinical irradiators (15). Moreover, to fully incorporate clinical relevance into preclinical in vitro studies, the proposed platform much allow for simple multi-modality irradiation to mimic various clinical irradiation schemes. An ideal platform will also be compatible with a clinical treatment planning system; by which commercially available dose calculation algorithms may be used eliminating the need for correction factors and dosimetric error and uncertainty found in preclinical orthovoltage irradiators.

This study first establishes a baseline dosimetry needed to be achieved by the clinical irradiators by acquiring dose measurements from a commercially available orthovoltage preclinical irradiator. Following this, the fabrication and initial evaluation of a modular, cost-effective 3D printed platform capable of multi-modality studies including HDR brachytherapy with subsequent EBRT irradiation is discussed. The dosimetric characteristics of the platform accoutered with a typical 96-well tissue culture plate were established via a comparison of calculation algorithms for the HDR afterloader and EBRT linear accelerator. After this, dose measurements were done using radiochromic film and optically stimulated luminescent dosimeters with the HDR and EBRT irradiators to assess the impact of material heterogeneities. This is due to the less than full scatter conditions inherent to the 96-well tissue culture plate. The ability of dose calculation algorithms with advanced heterogeneity corrections to properly account for this was then investigated.

Methods and materials

Cellular preclinical irradiation platform

The platform was designed in Solid Edge (Siemens Digital Industries Software, Plano, TX) and printed using a fused deposition modeling 3D printer (Ultimaker S5, Utrecht, Netherlands) with a polylactic acid (PLA) filament. The platform was inspected for inclusions via the simulation CT scan, from which the printing of the platform was found to be uniform. Aside from the HDR catheters, material for the platform cost less than 50 USD and took 1 to 2 days to print, owing largely to slow print speed. The platform base houses 14 standard 6 FR button ended catheters spaced 1 cm apart. Additional inserts allow variable distances between an HDR source and a standard tissue culture plate (Fig. 1), thus varying the dose rate. The choice of a 1 cm spacing between catheters was selected to mimic our standard clinical interstitial approach. 12 channels were found to sufficiently cover the entire tissue culture plate with a 1 cm spacing. From this, we added an extra channel to either side to aid in uniform coverage of the wells. However, the number and spacing of channels can be customized at the discretion of the end user. Given the platform was designed to be used with up to 14 source channels, various dose distributions can be created depending on experimental design, by varying the number of channels used, source positions and dwell times within each channels, and the distance between catheters and plates via the aforementioned inserts. Tissue culture plates come with a standard footprint size, so any standard size tissue culture plate can be placed inside the platform as there are nylon set screws that register the plate to one corner of the base. This study utilized a 96-well tissue culture plate (Techno Plastic Products, Trasadingen, Switzerland). The distance between the source channels and the adherent layer of cells (source-to-cell distance, SCD) within the wells of the 96-well tissue culture plate is approximately 0.4 cm and is comprised of approximately 0.1 cm of plastic with 0.3 cm of air. Additional components have been designed and printed to place the cells variable distances from the source (0.4 cm and 1 cm). As needed, nylon screws are used to attach additional components, which can allow for alternative setup geometries and/or multi-modality use. The implementation of the platform for this study uses all 14 source channels (Fig. 1). The use of a 96-well tissue culture plate represented challenging conditions for creating a uniform dose to the cells across the entire plate due to the amount of air pockets within the plate. While the platform does largely act as a holder of the tissue culture plate for EBRT irradiations, using the platform for these irradiations allows us to only need the one set of planning CT scans resulting in more consistency in dosimetry and setup both within and between modalities. Additionally, if desired, an adapter can be printed to fit a couch registration bar for the platform or a buildup cap that would screw onto the corner posts.

Fig. 1.

Well plate holder platform in the two source-to-cell distance (SCD) configurations and 14 6-French catheters spaced 1 cm apart. The 1 cm spacer is shown in the middle image giving the platform an SCD of 1 cm. The right image is of a 96-well tissue culture plate with the footprint size shown. Standard tissue culture plates like this will into the bounds of the platform. Catheters are manufactured in different colors for clinical distinguishability purposes and their order in this application is random.

Treatment planning and delivery

Preclinical orthovoltage irradiator

Orthovoltage irradiations were performed using a standard commercial preclinical irradiator (X-Rad 225iR or XRad, Precision X-ray Irradiation, Madison, CT); with adjustable jaws and adjustable source-to-surface distance (SSD). Service engineers from the manufacturer inspect and calibrate the machine annually. To capture the dosimetry for the normal operating conditions and standard experimental setups for this machine, tissue culture plates were placed directly on the irradiation table centered and aligned using the light field of the XRad. Irradiations on the XRad were delivered at 40 cm SSD with a 13 × 9 cm² field size to fully incorporate the tissue culture plate into the irradiation field. The XRad self-reports its delivered dose as well as dose rate and uses these factors to stop the irradiation. Using this system, the XRad was set to irradiate all plates to 4 Gy at a dose rate of 2.2 Gy/min.

Treatment planning

CT-based treatment plans were created using clinical treatment planning software (BrachyVision and Eclipse, Varian Medical Systems, Palo Alto, CA). CT scans were taken on two clinical simulation CT scanners (Siemens SOMATOM Confidence, SOMATOM go.Open Pro, Siemens Healthineers, Erlangen, Germany) each with a 0.1 cm slice thickness. While no dosimetric differences were expected between CT scanners, we wanted to remark on the feasibility of using any clinical CT scanner. In total, 4 separate CT scans were taken. All scans included the platform with wire markers in each of the 14 catheters and a 96-well tissue culture plate with 100 μL of water in each well. Two scans were acquired with the platform in the 0.4 cm SCD configuration: one with and one without OSLDs. Similarly, two scans were acquired with the same conditions but in the 1 cm SCD configuration. To account for variation in the filled volume of each well, a water contour was created and assigned to 0 HU in each of the 96 wells to be consistent with normal irradiation conditions. Furthermore, nitinol wire markers were inserted into the catheters for digitization purposes and were included in the CT scan. These markers were contoured and assigned to be air for EBRT treatments.

HDR brachytherapy irradiation

HDR brachytherapy irradiations were delivered using a clinical HDR afterloader (Varisource iX, Varian Medical Systems, Palo Alto, CA) with dosimetry evaluated using homogeneous water-based dose calculations (TG-43) and a commercial model-based dose calculation algorithm (MB-DCA) that includes heterogeneity corrections (Acuros BV or ABV, Varian Medical Systems, Palo Alto, CA). Clinical treatment plans frequently use 0.5 cm to 1 cm for their step size. 0.6 cm was found to be an optimal compromise between plan delivery efficiency and dose uniformity. Dwell positioning was constant throughout the HDR plans with a 0.6 cm step size spanning up to 10 cm in length per channel. The dose grid resolution was constant across BrachyVision plans at 0.25 cm for both the X and Y planes and 0.05 cm in the Z plane. To assess the dose the cells would receive, reference lines consisting of 5 points were place at the bottom of each well which is the location of the adherent layer of cells (Fig. 2). HDR plans were optimized by manual isodose reshaping and evaluated using the aforementioned reference lines until a uniformity of approximately 5% across the tissue culture plate was achieved.

Fig. 2.

Transversal (left) and coronal (right) cuts of the platform with a 96-well tissue culture plate. Labeled in the transversal slice is the tissue culture plate sitting atop the 1 cm spacer, which is all contained in the 3D printed platform. In the coronal slice, the tissue culture plate is sitting within the 3D printed platform and visible are the water contour and the reference lines placed at the bottom of the sampled wells. The outer wells do not contain reference lines because the outer wells are often excluded in biological studies due to excessive evaporation. The coordinate system given is consistent with the coordinate system displayed in Eclipse but differs from that given for the dose grid in Eclipse.

EBRT irradiation

EBRT irradiations used a clinical linear accelerator (Edge, Varian Medical Systems, Palo Alto, CA) and akin to the brachytherapy dosimetry, dosimetry for EBRT was calculated using a commercial convolution/superposition algorithm (Analytical Anisotropic Algorithm or AAA, Varian Medical Systems, Palo Alto, CA) and an MBDCA (Acuros XB or AXB, Varian Medical Systems, Palo Alto, CA). Strictly speaking, AAA is an MBDCA with heterogeneity corrections, albeit somewhat limited, through radiological scaling; however, from here on out, AAA will be referred to as a standard calculation model and compared against the Acuros XB (AXB) dose calculation algorithm. The dose grid resolution was adjusted to ensure complete sampling coverage of relevant contours. For these plans, 4 Gy was delivered at 600 MU/min (5.9 Gy/min) to the cells by placing the isocenter in the plane of the cells and treating with a single anterior/posterior 15 × 15 cm² field at a 100 cm source-to-axis distance (SAD). To assess the dose to the cells, 5 reference points were placed at the bottom of each well (Fig. 2).

Dose calculation

For both HDR and EBRT, plans were made to deliver 4 Gy uniformly to the adherent layer of cells across the tissue culture plate using the TG-43 or AAA calculations (Fig. 2). The plans were copied and recalculated using the ABV or AXB calculations to assess the necessity for improved heterogeneity corrections. Both MBDCA calculations were reported as dose to medium.

Dosimetry

Film

Film was chosen as the primary dosimeter as it was recommended (16) and used in this capacity in previous studies (4,8,11,15). The film was taped to the bottom of a tissue culture plate to verify the dose near the adherent cells for XRad, HDR, and EBRT irradiations, which is a similar approach to other published studies (8). For all film irradiations, each well of the 96-well tissue culture plate was filled with 100 μL of water, equivalent to the level as would be used in an in vitro cell study. The radiochromic film (EBT3, Ashland Global, Wilmington, DE) used for the dosimetry in this study has a dynamic dose range of 0.1 cGy to 20 Gy with minimal energy dependence from 100 keV to the MV range and is all from the same batch. Following irradiation, the film was allowed to photopolymerize for 48 hours and placed under a glass compression plate in the center of the scanner and scanned subsequent to 5 warm up scans (Epson Expression 10000 XL, Seiko Epson Corporation, Suwa, Nagano, Japan). Scanning parameters were 48 bits, 300 dpi, and color corrections disabled, and images were saved in TIFF format.

Film: XRad

Prior to creating our calibration curve for the XRad, we ensured the manufacturer had completed their annual inspection and calibration. Squares of film (3 × 3 cm²) were irradiated on the XRad to 15 known doses ranging from 0 to 5 Gy. Film squares were scanned and analyzed and a calibration curve was created specific to the XRad. The dose uniformity for the XRad was assessed by irradiating a 96-well tissue culture plate with film taped to the underside of the tissue culture plate. This measurement was repeated 4 times. The SSD was set at 40 cm and the field size was set to the boundary of the tissue culture plate. The film was analyzed by selecting a suitable region within each well of the 96-well tissue culture plate and gathering the mean dose with standard deviation from this area. The mean and median of these results was reported along with the standard deviation in dose across each well and the mean standard deviation within each individual well.

Film: HDR brachytherapy and EBRT

For the HDR brachytherapy and EBRT film irradiations, a calibration curve was created on an Edge linear accelerator (Varian Medical Systems, Palo Alto, CA) at 6 MV. Following our standard clinical calibration protocol, 16 jaw defined 3 × 3 cm² squares of film were irradiated to known doses ranging from 0 to 7 Gy. A separate calibration for the HDR afterloader was not needed as the mean discrepancy for EBT3 film between a 6 MV Linac beam and a ¹⁹²Ir calibration curve was found to be close to zero for doses above 250 cGy (22). HDR dosimetry was evaluated at an SCD of 1 cm (2.5–4.5 Gy/min for a 10 Ci [370GBq] Ir-192 source) using radiochromic film. Consistent with the XRad dosimetry, setup for the HDR and EBRT film measurements consisted of taping film to the underside of a 96-well tissue culture plate with each well filled with 100 μL of water. For the HDR film irradiations, 14 transfer guide tubes are connected to each of the 14 source channels on the platform. The tissue culture plate is placed within the platform with the 1 cm spacer. Film measurements for the clinical irradiators were done in triplicate and used a 2D gamma analysis, in which the film image and corresponding dose plane were normalized to the median value. For each film, an average passing rate was calculated from five 2D gamma evaluations each using a 5%/2 mm criteria. Each evaluation was done using a different normalization value chosen from high dose, low gradient regions of interest to account for the passing rate variability due to positional uncertainty in the selection of the normalization point.

OSLD: HDR brachytherapy

For the dosimetric evaluation at 0.4 cm SCD (30–40 Gy/min for a 10 Ci [370GBq] Ir-192 source), optically stimulated luminescent dosimeters (OSLDs) (nanoDot^™ , Landauer, Glenwood, IL) were selected. This was because the film appeared to over-respond at a 0.4 cm SCD. The measurement photon energy range for the OSLDs used is 15 kV to 25 MV with a minimum and maximum dose measurement value of 0.05 mGy and 13 Gy, respectively. Additionally, a linear response was found for OSLDs up to 600 cGy with HDR brachytherapy sources (23). In total, 8 OSLDs were placed on the bottom of the 96-well tissue culture plate at wells and scanned using the clinical CT simulation scanners. In the same manner as the HDR film irradiations, 14 transfer guide tubes were connected to each of the 14 source channels on the platform. The water filled tissue culture plate with the 8 OSLDs was placed into the platform and set screws were tightened to keep the platform from shifting relative to the source channels. Following a 24-hour recombination period, the OSLDs were read (microSTAR ii, Landauer, Glenwood, IL) using an HDR brachytherapy calibration curve with a dose range from 2.5 to 22 Gy to give an absolute dosimetry reading of the OSLDs.

Results

Film: XRad

In all four measurements, the XRad self-reported a delivered dose of 400 ± 2 cGy. Each film was irradiated using the same protocol and conditions, and yet, the range in measured dose to the film was 394.5 ± 10.1 cGy to 440.2 ± 3.1 cGy. Even though the delivered dose was reported by the machine to be nearly the same across irradiations, the standard deviation across all measured film doses was 17.4 cGy. The mean dose across the sampled wells was 423.7 ± 8.5 cGy. To better show the data without an outlier, the median dose ( ±standard deviation) across the sampled wells for the XRad was 430.1 ± 5.6 cGy, which is a percent dose difference of 7.2% from the prescribed 400 cGy. The mean and median standard deviation in dose within individual wells was 6.1 cGy and 5.6 cGy, respectively.

Dose calculation comparison

The relative dose within each well for both HDR and EBRT calculations are shown as heat maps to depict dose uniformity across the sampled plate (Fig. 3). In the HDR case, calculations are compared at both SCDs. There is a notable difference in dose uniformity between the 0.4 cm SCD HDR plan and the 1 cm SCD HDR or EBRT plans (Fig. 3). Even though the range in dose for the ABV dose calculation at a 4 mm SCD is 9% compared to the plate average, the standard deviation in the ABV dose calculation across the entire plate at a 4 mm SCD is 2.3% (Fig. 3). At a 1 cm SCD using the ABV dose calculation, the same range and standard deviation gives 2.5% and 0.6%, respectively. In the case of EBRT dose calculations, AXB gives a range and standard deviation of 7% and 0.9%, respectively.

Fig. 3.

Heat map of relative dose in each well calculated with standard calculation models (TG-43 and AAA) (left) and MBDCAs (ABV and AXB) (right) calculations at an SCD of 1 cm (top) and 0.4 cm (middle) for HDR. The outer rows (A and H) and columns (1 and 12) have been excluded from this analysis as these wells are often not seeded in an experiment due to excessive evaporation.

HDR dose calculation: 1 cm SCD

For a single 4 Gy fraction, within each well at an SCD of 1 cm, the mean standard deviation was 0.58 cGy (0.14%) and 0.64 cGy (0.17%) for TG-43 and ABV calculations, respectively. The mean dose (±standard deviation) in this plane is 409.0 ± 2.3 cGy and 372.3 ± 2.0 cGy with coefficient of variations of 0.56% and 0.54% for TG-43 and ABV calculations, respectively (Table 1). The percent difference for the mean dose between TG-43 and ABV calculations in each sampled well across the tissue culture plate is 9.4% (Table 1).

Table 1.

Parameters describing dose calculation uniformity between and within the sampled wells of the 96-well tissue culture plate for both brachytherapy and EBRT irradiations.

	Brachytherapy				EBRT
	SCD = 1 cm		SCD = 0.4 cm
Calculation models	TG-43	ABV	TG-43	ABV	AAA	AXB
Mean σ within a well	0.58 cGy (0.14%)	0.64 cGy (0.17%)	10.2 cGy (2.6%)	10.3 cGy (2.6%)	8.5 cGy (2.1%)	2.1 cGy (0.52%)
Mean dose ± σ across plate	409.0 ± 2.3 cGy	372.3 ± 2.0 cGy	398.3 ± 9.2 cGy	400.0 ± 9.3 cGy	406.4 ± 10.0 cGy	390.0 ± 3.4 cGy
Coefficient of variation across plate	0.56%	0.54%	2.3%	2.3%	2.5%	0.87%
%DStandard,MBDCA	9.4%		0.43%		2.4%

Abbreviations: %D_{Standard,MBDCA} = percent dose difference between standard calculations (TG-43, AAA) and MBDCA calculations (ABV, AXB), σ = standard deviation.

HDR dose calculation: 0.4 cm SCD

At an SCD of 0.4 cm, the mean standard deviation within each well is 10.2 cGy (2.6%) and 10.3 cGy (2.6%) for TG-43 and ABV calculations, respectively. The mean dose (±standard deviation) across the 96-well tissue culture plate is 398.3 ± 9.2 cGy and 400.0 ± 9.3 cGy for TG-43 and ABV calculations giving coefficient of variations of 2.3% and 2.3%, respectively (Table 1). A percent difference of 0.43% was found between the TG-43 and ABV calculations for the mean dose in each well across the tissue culture plate.

EBRT dose calculation

Gathering data from reference points in the EBRT plans, a standard deviation within individual wells of 8.5 cGy (2.1%) and 2.1 cGy (0.52%) was found for AAA and AXB calculations, respectively. The mean dose (±standard deviation) across the 96-well tissue culture plate is 406.4 ± 9.3 cGy and 390.0 ± 3.4 cGy with coefficient of variations of 2.5% and 0.87% for AAA and AXB calculations, respectively (Table 1).

HDR brachytherapy and EBRT dose measurements

Our film measurements were analyzed using a 2D gamma evaluation with a 5%/2 mm criteria. HDR brachytherapy film resulted in passing rates (±standard deviation) of 94.8 ± 1.9% and 98.0 ± 0.9% for TG-43 and ABV calculations, respectively. EBRT film resulted in passing rates of 84.9 ± 0.5% and 97.4 ± 1.8% for AAA and AXB calculations, respectively.

Placement of the OSLDs was chosen to sample the dose in various locations across the plate (Fig. 4). In two cases, OSLDs were placed under neighboring wells. With OSLDs placed at the bottom of the water filled 96-well tissue culture plate, three independent measurements were carried out, each with a new set of OSLDs, at an SCD of 0.4 cm resulting in a mean percent difference (±standard deviation) relative to TG-43 and ABV calculations of 10.7 ± 3% and 9.4 ± 4%, respectively (Fig. 4).

Fig. 4.

OSLD placement is represented as white wells. OSLD dose averaged between three sets of measurements plotted with the TG-43 and ABV calculated doses. Error bars indicate standard deviation in measurements. OSLD location corresponds to the row (letter) and column (number) position of that well.

Distances to agreement (DTA ± standard deviation) were gathered by finding the distance between expected (dose within scintillating material) and measured doses along a 1D dose profile through the location of the OSLD along the Z-axis (refer to Fig. 2 for the orientation of the Z-axis). This was done to highlight that the distance between the OSLD and the calculated dose point equal to the measured dose falls within the uncertainty in locating the scintillating material within the OSLD housing (25). The DTA ± standard deviation for the TG-43 and ABV dose calculations were 0.12 ± 0.04 cm and 0.13 ± 0.04 cm, respectively (Fig. 5).

Fig. 5.

Distances to agreement between OSLD measurements and TG-43 and ABV calculations. Error bars indicate standard deviation in the distance to agreement for the three OSLD measurements. OSLD location corresponds to the row (letter) and column (number) position of that well.

Discussion

We have designed, printed, and dosimetrically evaluated a 3D printed solution for preclinical in vitro HDR brachytherapy studies. 3D printing the platform gives the design a high degree of modularity allowing it to be easily adaptable to other clinically relevant SCDs for HDR irradiation as well as multi-modality studies including clinical EBRT irradiators. This study highlights further advantages of this system as it has been implemented with a clinically relevant workflow utilizing CT-based treatment planning with corrections for heterogeneities. Another benefit of using a clinical workflow lies in the ease with which in vitro studies can be performed using multiple clinical irradiators. Since a 96-well tissue culture plate has many air gaps, and thus produces insufficient scatter conditions, it was chosen to assess the extent that advanced heterogeneity corrections can account for insufficient scatter conditions.

This platform offers various advantages over what was used in previous studies. In particular, the switch from preclinical orthovoltage irradiators offers the ability to create simple or complex treatment plans for in vitro studies and assurance in the accuracy of delivered dose. Large differences in programed and measured dose from preclinical orthovoltage irradiators can, in part, be attributed to insufficient scatter conditions in experimental design (15). Two studies created platforms, one acrylic and one 3D printed, using a single catheter (17,18). There are many benefits to using a single catheter including several interesting studies; however, including multiple catheters in a design allows for more possibilities while retaining the capability to use a single channel. A notable aspect of our platform is the use of 14 catheters giving us uniform dose distributions or, if desired, complex dose distributions by using any combination or selection of the 14 catheters. Studies including multiple catheters have previously been done (19-21), and, in fact, in a prior study (21) the need for full scatter conditions in experimental design with a HDR brachytherapy remote afterloader was determined through Monte Carlo simulation and radiochromic film analysis. From this, a dose correction factor was created to account for the difference in measured dose from the expected dose. A distinct advantage of using a clinical workflow, and therefore commercial model-based dose calculation algorithms, lies in the ability to account for material heterogeneity and insufficient scatter conditions all while using tools already present in a clinic. Thus, their use can eliminate the need for dose correction factors.

Given that orthovoltage irradiators are still the current standard for preclinical in vitro studies, it is important to establish the baseline dosimetric properties of a common commercially available preclinical irradiator. The XRad 225iR uses a parallel plate ion chamber open to atmosphere and does not correct for temperature or pressure. This could potentially lead to uncertainty in the measured dose as it uses this chamber to stop the irradiation. The dosimetry of the XRad proved to be inconsistent throughout the film analysis; therefore, it was performed in quadruplicate. The inconsistency is apparent in a 17.4 cGy standard deviation across the four film readings with one reading showing 10% higher dose than prescribed despite the XRad displaying a delivered dose of 400 ± 2 cGy for each of the irradiations. The large variation in measured dose from the XRad could be attributed to the calibration of the machine. Specifically, the often drastically different irradiation conditions seen with experimental use of the machine versus the conditions by which it was calibrated. It might be inferred that our calibration curve may exhibit a similar deviation from the prescribed and delivered dose resulting in a possible median percent dose difference of 16.7% from the prescribed 400 cGy. Given these large deviations from the prescribed dose, it is apparent that by utilizing clinical machines, preclinical studies can be performed with a higher degree of certainty in the delivered dose.

Prior to using the platform in preclinical in vitro studies, it is important to first demonstrate that uniform dose distributions may be delivered to the cell plane at both small and large SCDs for HDR irradiation as well as for the clinical EBRT irradiator that meet or exceed the XRad.

The need for heterogeneity corrections is suggested in prior studies (14,15,21) and confirmed by the 9.4% difference between brachytherapy dose calculations at 1 cm SCD. In agreement with this, the 2D gamma analysis showed better agreement with MBDCA calculations than the standard calculation models suggesting that the abundant air pockets in the 96-well tissue culture plate and/or lack of material surrounding the tissue culture plate engender a deviation from full scatter conditions. Claridge Mackonis et al. 2018 found that the monitor units required to deliver 2 Gy to cells in various tissue culture plate designs would need to be scaled by approximately 6% and 4% for AAA and AXB calculations, respectively, for a 6 MV beam (14). As shown by Chen et al. 2019, attaining full scatter conditions depends largely on experimental setup including irradiation platform design (15). In addition, Geraldo et al. 2021 reports a difference of 12% between their MC calculations with full scatter conditions and their MC code with the 96-well plate, which has been attributed to the lack of full scatter inherent to the 96-well plate (15). Our platform does not produce full scatter conditions as evident in the percent dose difference between standard calculation models and MBDCAs. However, with that being said, the film passing rates of the MBDCA calculations show that the workflow used here can reasonably account for the lack of full scatter conditions, thus establishing the dosimetric characteristics of the platform exceed the current standard for preclinical in vitro studies.

Conceivable uncertainties in the printed platform include possible air pockets within the platform and the fitment of a tissue culture plate, in particular its position relative to the source channels. To address this first point, we inspected for inclusions by reviewing the simulation CT scan of the platform and found none. The latter point was addressed in the platform design. The four set screws are used to ensure that each tissue culture plate will be located in the same spot relative to the source channels. There could be a possible uncertainty in locating the scintillating material within the OSLD and correlating this position to dose calculation points within the treatment planning system, which could explain at least a portion of the 10.7% and 9.4% dose difference in the OSLD measured dose to the TG-43 and ABV calculated doses, respectively. Even still, a calculated dose matching the OSLD measured dose is located very near to the approximate OSLD location in the Z-axis as evident by a mean DTA of 0.12 ± 0.04 cm and 0.13 ± 0.04 cm, respectively. OSLD agreement to both dose calculations, TG-43 and ABV, is sufficiently good to suggest that heterogeneity corrections due to excess air in the 96-well plate are not as important when very close to the source. Furthermore, the approximately 10% dose difference between measurement and the calculations is within published OSLD experimental uncertainty (23-25). Given this published uncertainty, the minimal DTA, and nearly identical pattern between measured and calculated OSLD dose (Fig. 5), we believe that even when nearly adjacent to the HDR source, the workflow and platform used can accurately describe the dosimetry to the location of adherent cells.

To improve upon the design of the platform a buildup cap could be printed helping with EBRT and HDR brachytherapy dosimetry. Additionally, an enclosure could be printed or bolus added to encase the tissue culture plate to help establish full scatter conditions. If a higher average dose rate is wanted, plans could be optimized to use fewer source channels as long as the experimental design allowed for a smaller irradiation area. To reduce the material heterogeneities, present in multi-well plates, a component could be printed to fill these air gaps. For our study, we decided to present a device with the most universal application hence the absence of a buildup cap, enclosure or bolus, and gap filling components.

The ability of a preclinical in vitro irradiation system to uniformly irradiate any tissue culture plate or flask, particularly a 96-well tissue culture plate (suboptimal scatter conditions), is tantamount to its ability to deliver clinically relevant and non-uniform irradiation schemes. While our study has thus far focused on uniform irradiation of a 96-well tissue culture plate, our platform is capable of complex and clinically relevant irradiation schemes for any standard size tissue culture plate and numerous irradiation modalities. Perhaps the most remarkable aspect of the platform is the ability to perform in vitro studies with the robustness of a clinical workflow scheme and the precision and accuracy of clinical irradiators all the while being readily implemented in situations with limited resources. Reducing the SCD to 0.4 cm allows for preclinical studies in regions of high dose gradients and high dose rates without the need for dose correction factors. While prior studies (17,19) have investigated biological endpoints near an HDR brachytherapy source, our design readily allows for successive EBRT irradiations without the need for a different experimental setup. This opens the door to investigating biological endpoints arising from mixed modality irradiations seen clinically such as that of the standard cervical cancer treatment regime.

Conclusion

In this study, the dosimetry for a commercially available, preclinical orthovoltage irradiator was assessed establishing a baseline for our novel, 3D printed platform to be compared to. This platform is adaptable to other experimental designs including multi-modality experiments, and has addressed the current limitations of preclinical in vitro brachytherapy studies through the delivery of uniform fields to a 96-well tissue culture plate in regions of high dose rate and, therefore, also further from the source. Additionally, this study has been consistent with clinical workflow in part by using commercially available dose calculation algorithms, some including advanced heterogeneity corrections, giving the platform the ability to seamlessly transition from one treatment modality to another with the same setup. Through dosimetric evaluation, the study has shown the need for heterogeneity corrections when full scatter conditions are unable to be met. The dosimetric studies of this platform indicate feasibility for novel preclinical in vitro studies incorporating both HDR brachytherapy and EBRT irradiations which allow for more clinically relevant studies particularly in the context of cervical cancer.

Abbreviations:

HDR

high dose rate

EBRT

external beam radiotherapy

AAPM

American Association of Physicists in Medicine

MC

Monte Carlo

SCD

source-to-cell distance

SSD

source-to-surface distance

XRad

X-Rad 225iR

TG-43

homogeneous water-based dose calculations

MBDCA

model-based dose calculation algorithm

ABV

Acuros BV

AAA

Analytical Anisotropic Algorithm

AXB

Acuros XB

OSLD

optically stimulated luminescent dosimeter

%D_TG-43,MBDCA

percent dose difference between; TG-43 and MBDCA, calculations

σ

standard deviation

DTA

distance-to-agreement