Development and characterization of a prototype selenium‐75 high dose rate brachytherapy source

Abstract

Background

⁷⁵Se ($t_{1/2}$ $\approx$120 days, $E_{\gamma ,\text{avg}}$ $\approx$215 keV) offers advantages over ¹⁹²Ir ($t_{1/2}$ $\approx$74 days, $E_{\gamma ,\text{avg}}$ $\approx$360 keV) as a high dose rate brachytherapy source due to its lower gamma energy and longer half‐life. Despite its widespread use in industrial gamma radiography, a ⁷⁵Se brachytherapy source has yet to be manufactured.

Purpose

A novel ⁷⁵Se‐based source design with a vanadium diselenide core, titled the SeCure source, was proposed. This study aimed to evaluate the feasibility of this source design for dosimetry and manufacturability purposes and to develop an activated prototype source.

Methods

The source was modeled and integrated into the Monte Carlo‐based treatment planning system RapidBrachyMCTPS, where its TG‐43U1 parameters, photon spectrum, and broad beam first half‐value layers (HVL₁) and tenth‐value layers (TVL₁) in lead, tungsten, and concrete were calculated. A prototype source was manufactured, and the vanadium diselenide content of the capsule was verified with neutron radiography. The source was then activated to a nominal activity of $8.5\pm 0.9$ mCi at the McMaster Nuclear Reactor. The activity was measured with two separate dose calibrators. Gamma spectroscopy was used to characterize any activated radioactive contaminants in the source, and wipe testing was performed to check for any leakage of ⁷⁵Se from the encapsulation.

Results

The SeCure source's TG‐43U1 parameters were computed, showing that $2.056\pm 0.003$ times the activity of ⁷⁵Se is required relative to ¹⁹²Ir to achieve the same dose rate in water at (1 cm, 90). The mean spectral energy of the source is $214.695\pm 0.005$ keV, resulting in reduced first half‐value and tenth‐value layers relative to ¹⁹²Ir in attenuating materials. For example, the ${\mathrm{HVL}}_1$ was reduced from $2.795\pm 0.002$ mm to $1.020\pm 0.001$ mm in lead, from $2.049\pm 0.002$ mm to $0.752\pm 0.001$ mm in tungsten, and from $70.63\pm 0.04$ mm to $61.37\pm 0.03$ mm in concrete. The activated source achieved the desired activity, indicated as $9.2\pm 0.2$ mCi and $8.5\pm 0.9$ mCi at the end of irradiation on the two dose calibrators. All identified radionuclide contaminants decaying below $0.1\%$ of the ⁷⁵Se activity after 5 days post‐irradiation. Wipe testing only identified radioactive contaminants present in activated titanium, with only $1.24\pm 0.01\times {10}^{-7}$ mCi of detected 72 h post‐irradiation, indicating that the integrity of the encapsulation was maintained.

Conclusions

The SeCure design possesses the dosimetric, spectral, and physical properties necessary for a feasible high dose rate brachytherapy source. Next, manufacturing of a high‐activity SeCure source will be pursued.

1. INTRODUCTION

1.1. Sources and shielding in high dose rate brachytherapy

In high dose rate (HDR) brachytherapy, a sealed, highly radioactive, photon‐emitting source is temporarily placed inside or near the tumor. Due to the steep dose gradients from brachytherapy sources, a high dose can be delivered to the tumor while minimizing the dose to the surrounding healthy tissues. This distinctive characteristic makes HDR brachytherapy one of the most effective and precise radiation delivery modalities for specific tumor types, especially when used with image guidance. ¹ , ² , ³ , ⁴ , ⁵ , ⁶ HDR brachytherapy is currently delivered with ¹⁹²Ir ($E_{\gamma ,\text{avg}}$= 380 keV, $t_{1/2}$ = 74 days) and ($E_{\gamma ,\text{avg}}$= 1250 keV, $t_{1/2}$ = 5.3 years), chosen for their high mean energies ($>$200 keV), high dose rates, cost‐effectiveness, and practicality in manufacturing. To shield an ¹⁹²Ir HDR brachytherapy treatment room, typical wall thickness requirements are 54 cm of concrete, or 5 cm of lead, while for sources, typical shielding requirements are 85 cm of concrete or 14 cm of lead, assuming a desired transmission level of $2.7\times {10}^{-4}$ for uncontrolled areas. ⁷ Metallic shielding, usually lead, tungsten, or platinum, is also increasingly employed inside catheters and applicators to deliver intensity modulated brachytherapy (IMBT), shielding healthy tissue from radiation. ⁸

A gamma‐emitting isotope with lower mean energy than ¹⁹²Ir is desirable to reduce the required shielding thickness for brachytherapy treatment rooms and increase the attenuation of IMBT shields. For a novel isotope to be viable, it also must have a long enough half‐life to allow for transportation of the source globally and must produce dose rates comparable to current HDR brachytherapy sources to ensure similar treatment times. Additionally, the source must be feasible and cost‐effective to manufacture. To date, several intermediate energy isotopes (50 keV $<$ $E_{\gamma ,\text{avg}}$ $<$ 200 keV) have been explored for brachytherapy applications, including , ⁹ , ¹⁰ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹² , ¹⁶ and . ¹⁷ , ¹⁸ However, the clinical implementation of these sources has been hindered by challenges such as high production costs, low specific activities, and short half‐lives. Another potential source that meets the necessary criteria is ($E_{\gamma ,\text{av}}=215$ keV, $t_{1/2}$ = 119 days).

1.2. Selenium‐75

⁷⁵Se is a radioisotope commonly used in industrial gamma radiography. Its relatively low gamma energy enhances image quality and reduces required shielding, and its long half‐life reduces the need for frequent source exchanges; it has thus been identified as a cost‐effective option for various radiographic inspection applications. ¹⁹ The same characteristics make it a viable source for HDR brachytherapy. Several studies have already identified the large potential benefits of its lower gamma energy for reduced applicator and room shielding requirements. ¹⁵ , ²⁰ , ²¹ , ²² , ²³ , ²⁴

To date, developments in the manufacturing of ⁷⁵Se sources have primarily been toward industrial gamma radiography applications. ¹⁹ , ²⁵ , ²⁶ , ²⁷ The toxicity, volatility, reactivity, and corrosiveness of elemental selenium, combined with its significant expansion coefficient near its low melting point of 217, pose challenges in designing an encapsulated source. ²⁷ Improper handling can result in explosions, fires, and the production of harmful gases. ²⁸ In light of these challenges, industrial radiography sources are less compact than required for brachytherapy sources, ¹⁹ , ²⁵ , ²⁶ and typically encapsulate elemental selenium twice, once pre‐irradiation and again post‐irradiation. An alternative approach was introduced by Shilton, using ${\mathrm{VSe}}_2$ with greater chemical and thermal stability than elemental selenium, ¹⁹ , ²⁷ largely mitigating safety concerns during irradiation.

Since ⁷⁵Se  and ${\mathrm{VSe}}_2$  powder require encapsulation before irradiation, ²⁵ , ²⁶ it is important to select an encapsulation material that either remains nonactivated during neutron activation or produces only short‐lived radioisotopes compared to ⁷⁵Se. Long‐lived radioactive isotopes present in the source's encapsulation risk compromising the source's classification as a “sealed source” according to specifications of the Canadian Nuclear Safety Commission. ²⁹ Stainless steel, used to encapsulate ¹⁹²Ir  HDR brachytherapy sources due to its malleability and cost‐effectiveness, contains chromium, iron and nickel; neutron irradiation of stainless steel produces radioactive contaminants such as ($t_{1/2}$  = 27.7 days), ($t_{1/2}$  = 2.7 years), and ($t_{1/2}$  = 44.6 days). Consequently, stainless steel is not a suitable encapsulation material for a ⁷⁵Se  source, and a material with a low thermal neutron cross‐section must be chosen in place.

1.3. Objectives

This study aimed to design, fabricate, and characterize a novel ⁷⁵Se HDR brachytherapy source. Monte Carlo simulations were used to calculate the TG‐43U1 parameters of the source, ³⁰ , ³¹ investigate the photon spectrum, and compute half‐value layer (HVL) tenth‐value layer (TVL) in concrete, lead, and tungsten for shielding considerations. A prototype source was then manufactured, and quality assurance was performed using neutron radiography. Finally, the source was irradiated at the McMaster Nuclear Reactor, and leakage tests and gamma spectroscopy were performed following the irradiation, the latter used to identify potential radioactive contaminants.

2. MATERIALS AND METHODS

The first stage of this study designed and characterized an HDR brachytherapy source containing ⁷⁵Se using RapidBrachyMCTPS, ³² , ³³ a validated research treatment planning system available to the brachytherapy community upon request. RapidBrachyMCTPS offers a user‐friendly graphical user interface, a Monte Carlo‐based dose calculation engine, two dose optimization algorithms, contouring, dose analysis tools, and TG‐43 parameter and dose calculation tools. ³⁴ , ³⁵ , ³⁶ Following the design phase, the source was manufactured and subsequently irradiated at the McMaster Nuclear Reactor.

2.1. Design and characterization of the source

The final source design, dubbed the “SeCure” source, had a vanadium diselenide (${\mathrm{VSe}}_2$) active core with a length of 6.7 mm, an outer diameter of 0.55 mm, and a mass density of 5.078 ${\mathrm{g}\mathrm{cm}}^{-3}$. Maintaining a similar diameter to HDR ¹⁹²Ir brachytherapy sources, the extended length accounted for the nearly double activity that ⁷⁵Se sources require to achieve a comparable dose rate to ¹⁹²Ir. ²⁰ The active core was placed inside a 7.85 mm long titanium capsule with an inner diameter of 0.55 mm and outer diameter of 0.9 mm with a mass density of 4.54 ${\mathrm{g}\mathrm{cm}}^{-3}$. The capsule was attached to a titanium wire 16.5 mm long and 0.7 mm in diameter, but only 5 mm of the wire was simulated. The source design and encapsulation dimensions are illustrated in Figure 1.

FIGURE 1.

Schematic of the SeCure ⁷⁵Se source developed for this study (not to scale).

2.1.1. TG‐43U1 parameters and spectra

Simulations were performed with the Geant4‐based ³⁷ , ³⁸ , ³⁹ RapidBrachyTG43 module of RapidBrachyMCTPS ³⁶ to characterize the SeCure source design. TG43U1 parameters, including air‐kerma strength per unit activity, dose‐rate constant, radial dose function, and 2D anisotropy function, were computed using the default workflow and parameters of RapidBrachyTG43 outlined in Kalinowski and Enger. ³⁶ An active length of $L=0.67$ cm was used, corresponding directly to the length of the ${\mathrm{VSe}}_2$ core. For comparison purposes, TG‐43U1 parameters of the vendor‐independent generic ¹⁹²Ir  HDR source developed by Ballester et al., ⁴⁰ hereafter referred to as the “GenericHDR” source, were computed. The dose rate in water per unit activity at the reference position of $(r_0=1\text{cm},{\theta }_0={90}^{\circ }$) was determined as the product $\frac{S_K}{A}\times \mathrm{\Lambda }$ for both sources. Additionally, the photon spectra of both sources were determined by recording the energies of all photons escaping the source encapsulation after ${10}^8$ radioactive decays. Spectra were also computed for the GenericHDR ¹⁹²Ir source, SeCure source design with a pure ⁷⁵Se core, both ${\mathrm{VSe}}_2$ and pure ⁷⁵Se cores without the titanium encapsulation, and the raw ⁷⁵Se decay spectrum for comparison. A detailed description for all Monte Carlo simulations performed for this study compliant with TG‐268 recommendations ⁴¹ may be found in Table S1.

2.1.2. Attenuation data

For shielding considerations, attenuation data in the form of first HVL and TVL in lead, tungsten, and concrete were determined for the SeCure ⁷⁵Se and GenericHDR ¹⁹²Ir sources. A broad‐beam geometry for HVL and TVL calculations was selected due to ease of simulation, comparability with existing literature values, and applicability to shielding scenarios.

RapidBrachyTG43 was used to simulate a vacuum world, with spherical shells of a given thickness and attenuating material placed surrounding the source, and the air kerma in a 1 mm thick spherical shell beyond the attenuating material was scored. Air kerma at various attenuator thicknesses was computed for the SeCure source with lead, tungsten, and concrete attenuators. The composition of concrete was taken as the default Geant4 concrete material G4_CONCRETE. ${10}^7$ radioactive decays were used for each simulation, resulting in a maximum of 0.4% uncertainty in air kerma values. Air kerma at each attenuator thickness was normalized to the value at 0 mm thickness, and this normalized data was interpolated with cubic splines; the thickness with which the resulting interpolated function intersected with a normalized air kerma value of 0.5 and 0.1 was interpreted to be the first broad beam HVL and TVL, respectively (${\mathrm{HVL}}_1$ and ${\mathrm{TVL}}_1$).

2.2. Manufacturing, activation, and quality assurance of the source

2.2.1. Manufacturing process

Spectrum Safety Inc. (Lowell, Massachusetts, USA) was tasked to manufacture an inert SeCure source model with the same dimensions as the simulated source, shown in Figure 1. It contained a ${\mathrm{VSe}}_2$ active core consisting of ⁷⁴Se enriched to greater than 99.5% with naturally occurring vanadium, according to the patent by Shilton. ²⁷ The ${\mathrm{VSe}}_2$  compound was ground into a fine powder and pressed into three pellets using a 0.6 mm die at approximately 965 MPa. Based on manufacturer specifications, the total mass of ⁷⁴Se  contained in the source is approximately $6.08\pm 0.01$ mg, and the total length of the ${\mathrm{VSe}}_2$  pellets was $6.7\pm 0.1$ mm. The three pellets were then inserted into a titanium capsule with a total length of $7.85\pm 0.05$ mm. A $16.5\pm 0.1$ mm long piece of titanium wire, $0.40\pm 0.01$ mm in diameter, was laser welded to the encapsulation to seal the assembly, allowing for source handling. The total mass of the source was measured as $42.5\pm 0.2$ mg.

2.2.2. Neutron radiography

Neutron radiography was used to verify the physical dimensions of the ${\mathrm{VSe}}_2$ active core within the titanium encapsulation, completed at the McMaster Nuclear Reactor (Hamilton, Ontario, Canada) neutron radiography facility, operated by Applus+ Nray Services (Hamilton, Ontario, Canada). A rectangular, divergent collimator shaped the neutron flux before passing through the inert source. The source was irradiated for 5 min, with a thermal neutron flux at the image plane of $3.3\times {10}^6$ n ${\mathrm{cm}}^{-2}$ ${\mathrm{s}}^{-1}$. This short exposure time and low flux ensured the source was not activated during the process. ²⁰ The interaction between the thermal neutrons and the source was recorded on radiographic film positioned at the end of the beam path, enabling a detailed visualization of the internal structure of the source without inducing radioactivity.

2.2.3. Source activation and activity measurement

A vial containing the inert SeCure source prototype was capped in a quartz tube separately wrapped in aluminum foil, and placed in site 8C of the McMaster Nuclear Reactor core, with a nominal thermal neutron flux of $2.2\times {10}^{13}$ n ${\mathrm{cm}}^{-2}$ ${\mathrm{s}}^{-1}$. A 10% uncertainty in this neutron flux was assumed for activity calculations based on typical observed flux variations at this reactor site. This site was selected because of its highly thermalized neutron spectrum, which advantages thermal neutron activation of ⁷⁴Se to form the desired radionuclide ⁷⁵Se, rather than the production of unwanted radioactive impurities due to fast neutron reactions on vanadium and titanium. The vial was irradiated for 24 h to achieve an expected activity for the source of $8.5\pm 0.9$ mCi, low enough to allow manual source handling. A 23‐hour cooling period allowed short‐lived isotopes to decay before handling. The activity of the source was subsequently measured at 335 h post‐irradiation using the Capintec CRC‐55tR (Florham Park, New Jersey, USA) and AtomLab500 (Biodex Medical Systems, Shirley, New York, USA) nuclear medicine‐style ion chamber dose calibrators. Both detectors were previously calibrated with activity standards from the National Institute of Standards and Technology (NIST). Manufacturer‐specified coefficients were used to provide direct readout of ⁷⁵Se  activity, and no geometry correction was applied.

2.2.4. Contaminants analysis

In a separate experiment, a vanadium piece (215.4 mg) and titanium wire (16.8 mg) were irradiated for 24 h in site 5C of the McMaster Nuclear Reactor core, with a nominal neutron flux of $4.67\times {10}^{13}$ n ${\mathrm{cm}}^{-2}$ ${\mathrm{s}}^{-1}$. These samples were activated to investigate potential contamination radionuclides that could arise from the non‐⁷⁴Se components of the source.

Gamma spectroscopy was utilized to characterize radionuclide contaminants produced in the activated vanadium and titanium samples. Two samples of each material were analyzed using a high‐efficiency HPGe detector (ORTEC/AMETEK, Oak Ridge, Tennessee, USA) in sample position #5 with a lead shield to reduce background noise. Spectra were recorded at two time points, 48 and 96 h post‐irradiation, to monitor changes in gamma line intensities associated with the decay of radioactive species and to validate peak assignments based on decay characteristics. Gamma lines were first corrected for detector efficiency based on an efficiency curve based on a NIST‐certified multi‐gamma disc source. Radionuclides such as , , and were then identified by matching observed gamma energies with known reference values from the Live Chart of Nuclides, ⁴² confirming the presence of correlated gamma lines with expected relative intensities, and verifying half‐lives against established values. Once the contributing radioisotopes were identified, gamma lines were adjusted by their relative intensities obtained from the Live Chart of Nuclides ⁴² to obtain per‐radionuclide count rate. Finally, each radionuclide was decay‐corrected to reflect the activity levels at the end of irradiation.

2.2.5. Leakage tests

To assess the integrity of the titanium encapsulation and verify no radioactive leakage, wipe tests were performed at 24, 48, and 72 h post‐irradiation. The irradiated SeCure source was carefully extracted from its container using sterilized tweezers and wiped on all sides. Wipes were then analyzed via the same gamma spectroscopy workflow described above to quantify the activity of any potential radioactive isotopes that may have escaped the encapsulation.

3. RESULTS

3.1. Design and characterization of the source

3.1.1. TG‐43U1 parameters and spectra

The default workflow of the RapidBrachyTG43 module of RapidBrachyMCTPS ³² , ³³ , ³⁶ was used to calculate TG‐43U1 parameters for the SeCure ⁷⁵Se and the GenericHDR ¹⁹²Ir sources. The resulting parameters $S_K/A$ and $\mathrm{\Lambda }$ are shown in Table 1 and $g(r)$ and $F(1\text{cm},\theta )$ in Figure 2. The ratio of dose rates in water at $(r_0=\text{1 cm},\theta ={90}^{\circ })$ between the GenericHDR ¹⁹²Ir source and the SeCure ⁷⁵Se source at a given activity is $2.056\pm 0.003$. An uncertainty analysis of computed values of $S_K/A$ and $\dot{D}(r_0,{\theta }_0)$ in line with TG‐138 recommendations ⁴³ may be found in Table S2, yielding an estimated expanded uncertainty ($k=2$) of 1.5% and 1.6%, respectively.

TABLE 1.

Comparison of the mean source spectral energies computed by RapidBrachyTG43 for various core/encapsulation materials.

Isotope	Core material	Encapsulation material	$E_{\gamma ,\text{avg}}$ (keV)	Notes
75Se	None	None	$165.488\pm 0.004$	Raw 75Se decay spectrum
75Se	Se	None	$207.922\pm 0.005$	75Se core only
75Se	${\mathrm{VSe}}_2$	None	$210.489\pm 0.005$	${\mathrm{V}}^{75}\mathrm{Se}$ core only
75Se	${\mathrm{VSe}}_2$	Ti	$214.695\pm 0.005$	Manufactured SeCure source
75Se	Se	Ti	$214.883\pm 0.004$	Hypothetical source w/ pure 75Se core
192Ir	Ir	AISI 316L	$359.246\pm 0.008$	GenericHDR 192Ir source

FIGURE 2.

Comparison of TG‐43U1 parameters $g(r)$ (left) and $F(1\text{cm},\theta )$ (right) for the SeCure ⁷⁵Se source and GenericHDR ¹⁹²Ir source.

RapidBrachyTG43 was also used to compute the spectrum of photons escaping the SeCure ⁷⁵Se source, shown in Figure 3. For comparison, decay spectra with other ⁷⁵Se‐based sources with various core compositions and encapsulation scenarios, as well as the spectrum of the GenericHDR ¹⁹²Ir source, are given in Table 2. The mean gamma energy of the SeCure ⁷⁵Se source is $214.695\pm 0.005$ keV.

FIGURE 3.

Normalized energy spectrum of photons escaping the SeCure ⁷⁵Se source computed by RapidBrachyTG43.

TABLE 2.

TG‐43U1 parameters air‐kerma strength per unit activity, dose‐rate constant, and their product, $\dot{D}(r_0,{\theta }_0)/A$, the dose rate in water per unit activity at the reference position $(r_0=1\text{cm},{\theta }_0={90}^{\circ })$ for the SeCure source and GenericHDR source for comparison.

Source	$\frac{S_K}{A}$ (10 U/Bq)	$\mathrm{\Lambda }{(\mathbf{cm}}^{-2}$)	$\frac{\dot{D}(r_0,{\theta }_0)}{A}\left({10}^{-8}\frac{\text{cGy}}{\mathrm{h}·\text{Bq}}\right)$
SeCure	$4.760\pm 0.003$	$1.1186\pm 0.0008$	$5.324\pm 0.005$	$2.056\pm 0.003$
GenericHDR	$9.859\pm 0.006$	$1.1104\pm 0.0003$	$10.948\pm 0.009$	—

The dose rate ratio for the two sources is computed in the last column.

3.1.2. Attenuation data

RapidBrachyTG43 was used to simulate the attenuation of the SeCure ⁷⁵Se  source and the GenericHDR ¹⁹²Ir source through lead, tungsten, and concrete, with the resulting air kerma as a function of attenuator thickness shown in Figures 4 and 5. The cubic spline‐interpolated curves were used to estimate the ${\mathrm{HVL}}_1$ and ${\mathrm{TVL}}_1$ of both sources in each material, given in Table 3. Uncertainties in ${\text{HVL}}_1$ and ${\text{TVL}}_1$ were estimated by calculating their resulting variations when varying computed air kerma values by one standard deviation. Equivalent data for lead and concrete computed by Papagiannis et al. ⁷ for an ¹⁹²Ir point source are also compared.

FIGURE 4.

Broad beam air‐kerma as a function of lead (Pb) and tungsten (W) thickness normalized non‐attenuated air‐kerma, calculated by RapidBrachyTG43 for the SeCure ⁷⁵Se and GenericHDR ¹⁹²Ir sources.

FIGURE 5.

Broad beam air‐kerma as a function of concrete attenuator thickness normalized to a thickness of 0 mm, calculated by RapidBrachyTG43 for the SeCure ⁷⁵Se and GenericHDR ¹⁹²Ir sources.

TABLE 3.

Broad beam ${\mathrm{HVL}}_1$ and ${\mathrm{TVL}}_1$ and their associated uncertainties for the source and GenericHDR source through lead (Pb), tungsten (W), and concrete.

	Pb		W		Concrete
Source	${\mathrm{HVL}}_1$ (mm)	${\mathrm{TVL}}_1$ (mm)	${\mathrm{HVL}}_1$ (mm)	${\mathrm{TVL}}_1$ (mm)	${\mathrm{HVL}}_1$ (mm)	${\mathrm{TVL}}_1$ (mm)
SeCure	$1.020\pm 0.001$	$4.747\pm 0.003$	$0.752\pm 0.001$	$3.527\pm 0.002$	$61.37\pm 0.03$	$162.15\pm 0.05$
GenericHDR	$2.795\pm 0.002$	$11.226\pm 0.006$	$2.049\pm 0.002$	$7.982\pm 0.004$	$70.63\pm 0.04$	$190.73\pm 0.05$
‐ Papagiannis et al. 7	2.8	11	—	—	70	180

3.2. Manufacturing and quality assurance of the source

3.2.1. Neutron radiography

Figure 6 shows the neutron radiograph of the manufactured SeCure source. Three pellets were identified, with the pellet boundaries at $0.0\pm 0.1$ mm, $2.7\pm 0.1$ mm, $5.5\pm 0.1$ mm, and $6.7\pm 0.1$ mm. The uncertainty in the location of these pellet boundaries was estimated as 0.1 mm, one full‐length graduation, due to the difficulty in localizing the exact pellet boundary due to the presence of noise in the image. This yields individual lengths of $2.7\pm 0.1$ mm, $2.8\pm 0.1$ mm, and $1.2\pm 0.1$ mm, yielding a total length of $6.7\pm 0.2$ mm, thus matching the active length of the source design.

FIGURE 6.

Neutron radiography image showing the three pellets, with the boundaries of the pellets identified in red.

3.2.2. Source activation and activity measurements

Post‐irradiation, activity measurements at 335 h post‐irradiation using the Capintec CRC‐55tR and AtomLab500 detectors indicated a source activity of $9.2\pm 0.2$ mCi and $8.5\pm 0.9$ mCi, respectively, when corrected to end‐of‐irradiation. Uncertainties in measured activities are based on ⁷⁵Se  uncertainty thresholds reported by each device's manufacturer. The anticipated activity based on the nominal neutron flux during irradiation was $8.5\pm 0.9$ mCi.

3.2.3. Contaminants analysis

Table 4 shows the activities of contaminating radioisotopes found in the irradiated titanium (16.8 mg) and vanadium (215.4 mg) samples as measured with gamma spectroscopy, $A_{\text{Ti}}$ and $A_{\mathrm{V}}$, corrected to end of irradiation. Short‐lived isotopes ($t_{1/2}$ $<12$ h) were not quantified due to the delay between irradiation and spectrum acquisition. Uncertainty in measured per‐isotope activities was estimated at 1% based on data for comparable neutron activation experiments. ⁴⁴ The total radioisotope contaminant activities present in the activated SeCure source was desired, but could not be measured directly with gamma spectroscopy due to its high ⁷⁵Se content. Therefore, the sum of both materials' contaminant activities was taken, each scaled by the mass content of each material in the source relative to the sample mass, and corrected for the differing flux of the source and sample activation experiments, given as $A_{\text{source}}$ in Table 4. The largest estimated contaminant activity was ($t_{1/2}$ = 14.956 h), with an activity of $1.52\pm 0.01$ mCi. Only two other radionuclides had estimated activities about 10 $\mu \mathrm{Ci}$, ($t_{1/2}$ = 26.824 h) at ($3.04\pm 0.03)\times {10}^{-2}$ mCi and ($t_{1/2}$ = 80.3808 h) at ($2.14\pm 0.02)\times {10}^{-2}$ mCi. After a cooldown time of 5 days or 120 h, their activities would reduce to $5.84\pm 0.04$ $\mu \mathrm{Ci}$, $0.0250\pm 0.0002$ $\mu \mathrm{Ci}$, and $1.52\pm 0.02$ $\mu \mathrm{Ci}$ for , , and , respectively. In contrast, the nominal ($t_{1/2}\approx$120 days) activity of the SeCure source after 5 days would decay from 8.49 mCi to 8.25 mCi, meaning that each contaminant would contribute less than 0.1% the nominal ⁷⁵Se  activity after 5 days of cooldown, and proportionately less thereafter.

TABLE 4.

Radionuclide impurities detected via gamma spectroscopy in titanium and vanadium samples, with their respective activities corrected to end of irradiation, $A_{\text{Ti}}$ and $A_{\mathrm{V}}$, and the estimated contaminant activities in the SeCure source prototype, $A_{\text{source}}$.

Radionuclide	$t_{\text{1/2}}$ (h)	$A_{\text{Ti}}$ (mCi)	$A_{\mathrm{V}}$ (mCi)	$A_{\text{source}}$ (mCi)
	14.956	$1.69\pm 0.02$	$(5.89\pm 0.06)\times {10}^{-4}$	$1.52\pm 0.01$
	12.355	—	$(2.63\pm 0.03)\times {10}^{-2}$	$(1.43\pm 0.01)\times {10}^{-4}$
	2010.96	$(5.97\pm 0.06)\times {10}^{-4}$	—	$(5.37\pm 0.05)\times {10}^{-4}$
	80.3808	$(2.38\pm 0.02)\times {10}^{-2}$	$(1.03\pm 0.01)\times {10}^{-4}$	$(2.14\pm 0.02)\times {10}^{-2}$
	43.71	$(5.57\pm 0.06)\times {10}^{-3}$	$(7.41\pm 0.07)\times {10}^{-3}$	$(5.04\pm 0.05)\times {10}^{-3}$
	664.896	—	$(8.92\pm 0.09)\times {10}^{-4}$	$(4.86\pm 0.05)\times {10}^{-6}$
	1067.76	—	$(1.08\pm 0.01)\times {10}^{-5}$	$(5.88\pm 0.06)\times {10}^{-8}$
	12.7006	$(3.41\pm 0.03)\times {10}^{-3}$	$(8.03\pm 0.08)\times {10}^{-2}$	$(3.50\pm 0.03)\times {10}^{-3}$
	14.10	$(1.70\pm 0.02)\times {10}^{-4}$	$(7.60\pm 0.08)\times {10}^{-3}$	$(1.94\pm 0.02)\times {10}^{-4}$
	2874.72	—	$(1.03\pm 0.01)\times {10}^{-3}$	$(5.63\pm 0.06)\times {10}^{-6}$
	26.254	$(9.32\pm 0.09)\times {10}^{-4}$	—	$(8.38\pm 0.08)\times {10}^{-4}$
	35.282	—	$(1.85\pm 0.02)\times {10}^{-4}$	$(1.01\pm 0.01)\times {10}^{-6}$
	65.3712	$(6.24\pm 0.06)\times {10}^{-4}$	—	$(5.61\pm 0.06)\times {10}^{-4}$
	1444.8	$(1.44\pm 0.01)\times {10}^{-5}$	$(3.03\pm 0.03)\times {10}^{-5}$	$(1.31\pm 0.01)\times {10}^{-5}$
	40.28592	$(4.70\pm 0.04)\times {10}^{-3}$	—	$(4.23\pm 0.04)\times {10}^{-3}$
	26.824	$(3.38\pm 0.04)\times {10}^{-2}$	$(1.17\pm 0.01)\times {10}^{-3}$	$(3.04\pm 0.03)\times {10}^{-2}$
	100.44	$(4.62\pm 0.05)\times {10}^{-4}$	—	$(4.15\pm 0.04)\times {10}^{-4}$
	23.80	$(2.04\pm 0.02)\times {10}^{-4}$	$(1.10\pm 0.01)\times {10}^{-4}$	$(1.84\pm 0.02)\times {10}^{-4}$

Note: Half‐lives were obtained from the Live Chart of Nuclides. ⁴²

3.2.4. Leakage tests

Isotope activities measured via gamma spectroscopy from the wipe testing conducted at 24, 48, and 72 h post‐irradiation and are shown in Table 5. These tests detected , , and as surface contaminants, consistent with impurities observed in irradiated titanium samples. No activity from ⁷⁵Se was detected.

TABLE 5.

Activities of radionuclidic impurities detected by gamma spectroscopy for wipe testing at three time points post‐irradiation.

Time Elapsed (h)	Radionuclide	Activity (mCi)
24		$3.62\pm 0.04\times {10}^{-6}$
		$9.12\pm 0.09\times {10}^{-9}$
		$8.14\pm 0.08\times {10}^{-7}$
48		$5.24\pm 0.05\times {10}^{-7}$
72		$1.24\pm 0.01\times {10}^{-7}$

4. DISCUSSION

Previous studies investigating novel isotopes for HDR brachytherapy have identified potential hypothetical alternatives using Monte Carlo simulations, including , ⁹ , ¹⁰ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹² , ¹⁶ and . ¹⁷ , ¹⁸ Each of these radionuclides exhibits a combination of low specific activity, short half‐life, high production costs, or an unfavorable decay spectrum with a high beta emission yield. Out of these alternatives, only a source has been manufactured and activated, the pursuit of which was since abandoned due to its 32‐day half‐life, rendering the manufacture and distribution of a high‐activity source challenging. ⁷⁵Se is another promising isotope due to its lower gamma energy and longer half‐life relative to ¹⁹²Ir. A lower mean energy will ensure local dose deposition, minimize dose spillage to healthy tissue, reduce the shielding requirements of brachytherapy treatment rooms, and increase the dose attenuation of shielded applicators. ⁷⁵Se has also been shown to have a higher relative biological effectiveness than ¹⁹²Ir in radiobiological modeling. ⁴⁵

This work represents the first step toward producing a clinically viable ⁷⁵Se brachytherapy source. The proposed SeCure source design was characterized using Monte Carlo simulations, and a low‐activity source was manufactured and activated to realistically assess the feasibility of the source design.

4.1. Design and characterization of the source

4.1.1. TG‐43U1 parameters and spectra

TG‐43U1 parameters were calculated for the ⁷⁵Se‐based SeCure source using the validated RapidBrachyTG43 module of RapidBrachyMCTPS, ³² , ³³ , ³⁶ and were compared against the GenericHDR ¹⁹²Ir source ⁴⁰ parameters calculated with the same workflow. The TG‐43U1 dose rate ratio at the reference position $(r_0,{\theta }_0)$ in water at a given activity between the SeCure and GenericHDR sources was determined to be $2.056\pm 0.003$, implying that roughly twice the activity of ⁷⁵Se is necessary to achieve a comparable dose rate to ¹⁹²Ir. Weeks and Schulz previously calculated an exposure rate ratio between the two isotopes of 2.3. ²⁰ The lower value calculated in this work, likely directly attributable to the differences between exposure rate and dose rate in water, reduces the required activity of hypothetical clinical ⁷⁵Se sources and represents a more meaningful quantity in terms of patient dosimetry.

The TG‐43U1 dosimetry parameters $g(r)$ and $F(\text{1 cm},\theta )$, shown in Figure 2, revealed differences in the spatial dosimetric properties of both sources. The radial dose function $g(r)$ of the SeCure ⁷⁵Se source exhibits the same general trend as the GenericHDR ¹⁹²Ir source, with a larger maximum value characteristic of sources with a lower energy than ¹⁹²Ir due to an increased scatter dose contribution. ²¹ , ⁴⁶ $F(1\text{cm},\theta$), the 2D anisotropy function at $r=1$ cm, indicates a less anisotropic azimuthal dose profile for the SeCure source versus the GenericHDR due to its greatly extended active length. Looking towards patient dosimetry, ⁷⁵Se has been previously investigated only for static‐shielded endorectal ²² and ocular brachytherapy, ²⁴ and cervical IMBT. ¹⁵ Further patient dosimetry research across various sites, especially in non‐shielded scenarios, is necessary to evaluate the viability of ⁷⁵Se sources to replace ¹⁹²Ir in clinical practice.

Photon spectra were computed for the SeCure source model and several related scenarios. The SeCure spectrum, shown in Figure 3, has a mean gamma energy $E_{\gamma ,\text{avg}}$ = $214.695\pm 0.005$ keV, consistent with the mean energy of 215 keV reported by Hadadi and Ghanavati for a ⁷⁵Se‐based HDR source calculated by FLUKA. ²³ Photon spectra resulting from several other permutations of source and encapsulation materials were simulated, with their mean energies given in Table 1. Notably, the decay spectrum of has a mean energy of $165.488\pm 0.004$ keV, in the range of an intermediate energy brachytherapy source. However, considering the self‐attenuating effect in the core alone, the mean spectral energy rose to $207.922\pm 0.005$ keV and $210.489\pm 0.005$ keV for a pure and ${\mathrm{V}}^{75}\mathrm{Se}$ core, respectively. When considering the titanium encapsulation, slightly more beam hardening occurred, increasing the mean energies of both core compositions to $214.695\pm 0.005$ keV and $214.883\pm 0.004$ keV. For the encapsulated SeCure source design, the practical necessity of using a ${\mathrm{VSe}}_2$ core had minimal effect on mean energy versus pure ⁷⁵Se. The mean energy of the SeCure source is expectedly substantially lower than that of the GenericHDR ¹⁹²Ir source, $359.246\pm 0.008$ keV. The result of this energy difference in terms of attenuation was further explored with HVL/TVL calculations.

4.1.2. Attenuation data

First broad beam half‐value layers and tenth‐value layers (${\mathrm{HVL}}_1$ and ${\mathrm{TVL}}_1$) of the SeCure ⁷⁵Se and GenericHDR ¹⁹²Ir sources were computed for lead, tungsten, and concrete, given in Table 3, computed from transmission curves in each material (Figures 4 and 5). These three materials were selected due to their wide use in brachytherapy room and/or in‐patient shielding. The equivalent values computed for ¹⁹²Ir by Papagiannis et al. ⁷ were comparable for lead, validating the calculation methodology, but differed for concrete, likely due to differences in simulated elemental composition. The transmission curves in lead and tungsten for the SeCure source also appear visually to be in good agreement with those computed by Currier et al., ⁴⁷ though no values of ${\mathrm{HVL}}_1$ and ${\mathrm{TVL}}_1$ are reported. In general, the SeCure source's ${\mathrm{HVL}}_1$ and ${\mathrm{TVL}}_1$ were approximately half or less of those of the GenericHDR source.

Lesser thickness required for equivalent dose attenuation underpins the strong potential for a ⁷⁵Se‐based source to greatly reduce shielding requirements for new brachytherapy treatment rooms, and to provide less dose transmission through shielded applicators and catheters, providing more dose sparing of organs at risk. For example, a novel rectal IMBT applicator uses 7.5 mm of tungsten shielding, ⁴⁸ , ⁴⁹ providing less than one ${\mathrm{TVL}}_1$ for ¹⁹²Ir (${\mathrm{TVL}}_1$ = 7.99 mm) but much more than one ${\mathrm{TVL}}_1\mathrm{s}$ for ⁷⁵Se (${\mathrm{TVL}}_1$ = 3.53 mm). It is important to note, however, that subsequent HVL/TVLs will increase in thickness due to beam hardening. A full attenuation characterization of ⁷⁵Se in relevant treatment room and applicator shielding materials beyond the first HVL/TVL is therefore warranted according to the methodology of Papagiannis et al. ⁷

4.2. Manufacturing and quality assurance of the source

Though recognized for many years as a promising isotope for use in brachytherapy, the development and clinical introduction of ⁷⁵Se‐based brachytherapy has been substantially impeded by manufacturing difficulties, namely the high volatility of elemental selenium near its melting point of 217 . Encapsulating enriched ⁷⁴Se before irradiation in a reactor also risks activating the capsule material, a new challenge versus ¹⁹²Ir and post‐irradiation. The SeCure source employs a stable ${\mathrm{VSe}}_2$ compound, based on Shilton's patent. ²⁷ Vanadium enhances stability during manufacturing and irradiation without affecting the mean photon energy emitted. Titanium was selected for the capsule due to its stable isotope composition and low contamination risk under thermal neutron flux, which limits unwanted activation products.

${\mathrm{VSe}}_2$ powder was pressed into pellets and placed into the source encapsulation. Uncertainties in the pellet thicknesses and densities were therefore identified, which could, in turn, substantially affect the activity of the activated source. Neutron radiography performed on the manufactured source, shown in Figure 6, enabled an independent verification of the source dimensions. Future fabrication of the SeCure source prototype must, therefore, control for ${\mathrm{VSe}}_2$ pellet size and density and include neutron radiography or an equivalent measure for source quality assurance after assembly.

4.2.1. Source activation and activity measurements

Activation of the prototype SeCure source was achieved by irradiation in a thermal neutron flux provided by the McMaster Nuclear Reactor. This initial activation to an expected activity of $8.5\pm 0.9$ mCi tested the integrity of the encapsulation when subject to the conditions in a nuclear reactor core during and post‐irradiation. To confirm the approximate accuracy of activation calculations, activity measurements were taken at 335 h post‐irradiation using two different dose calibrators, allowing for substantial decay of the high‐activity contaminant isotopes identified in Table 4. The SeCure prototype source activity at the end of irradiation was measured as $9.2\pm 0.2$ mCi from the Capintec CRC‐55tR and $8.5\pm 0.9$ mCi from the AtomLab500 with no geometry corrections. While these activities agreed with the expected activated source activity within uncertainty, these measurements only serve as an approximate confirmation of activity on an order‐of‐magnitude level. Both detectors are calibrated for use with aqueous solutions of radiopharmaceuticals with a standard positioning, mass and sample geometry. Geometry corrections to account for an encapsulated solid source were thus neglected, further limiting the accuracy of measured activity values by an unknown degree. Future development of the SeCure source must establish a protocol for precise source strength and/or activity measurements, which should shift toward well‐chamber measurements traceable to primary dose standards in line with current societal recommendations. ⁵⁰ , ⁵¹ Expected source activity in future experiments should also be more accurately computed with a neutron flux value measured throughout activation, rather than the nominal flux of the reactor site.

Due to regulatory considerations related to post‐irradiation handling, the activity achieved in this study was well below the 20 Ci needed to match the dose rate of a typical 10 Ci ¹⁹²Ir  source. Achieving this level of activity would require irradiation in a high‐flux research reactor beyond the capabilities of the McMaster Nuclear Reactor. Suitable facilities include the Belgian Reactor 2 and the High Flux Isotope Reactor at Oak Ridge National Laboratory, both of which provide maximum thermal neutron fluxes on the order of ${10}^{15}$ n ${\mathrm{cm}}^{-2}$ ${\mathrm{s}}^{-1}$. ⁵² , ⁵³ Based on standard activation calculations, Weeks and Schulz ²⁰ demonstrated that a specific activity of approximately 1.5 Ci‐mg⁻¹ can be achieved during a 30‐day irradiation at these flux levels. Given the $6.08\pm 0.01$ mg of ⁷⁴Se  contained in a SeCure source, a single activation could produce around 9 Ci of ⁷⁵Se, which could be further increased through extended or repeated irradiation cycles.

4.2.2. Contaminants analysis and leakage tests

Gamma spectroscopy was used to characterize the contaminant isotopes present in titanium and vanadium samples during the activation and estimate the total contaminant activities in the activated source, as presented in Table 4. The dominant contaminant in the source is , with $A_{\text{source}}=1.52\pm 0.01$ mCi and $t_{1/2}$ = 14.956 h. Secondary contaminants were identified as , $A_{\text{source}}=(2.14\pm 0.02)\times {10}^{-2}$ mCi and $t_{1/2}$ = 12.355 h, and , $(3.38\pm 0.04)\times {10}^{-2}$ mCi and $t_{1/2}$ = 26.824 h, both more than 10 $\mu \mathrm{Ci}$ at end of irradiation. Each of these contaminant activities would reduce below to < 0.1% of the ⁷⁵Se  activity after 5 days of cooldown. The joint American Association of Physicists in Medicine (AAPM) and European Society for Therapeutic Radiology and Oncology (ESTRO) Task Group 167 (TG‐167) recommend that, for novel brachytherapy sources, radiation contaminants should contribute less than 5% of the dosimetric contributions of the primary radionuclide. ⁵⁴ Though we did not explicitly evaluate their dosimetric contributions, radioactive contaminants from the SeCure source's encapsulation are extremely unlikely to exceed this 5% threshold with a very low (< 0.1%) activity contribution relative to ⁷⁵Se.

Wipe testing on the activated SeCure source prototype was performed to evaluate the integrity of the encapsulation throughout the irradiation process, with contaminant activities detected on the wipe at 24, 48, and 72 h given in Table 5. Wiped radionuclides were low and consistent with contaminants consistent with the titanium encapsulation, verifying that the ⁷⁵Se activity was successfully contained inside the source.

5. CONCLUSIONS

We have successfully designed, developed and activated a prototype ⁷⁵Se‐based HDR brachytherapy source model known as the SeCure source. The source model's TG‐43U1 parameters, photon spectrum, and ${\mathrm{HVL}}_1$ and ${\mathrm{TVL}}_1$ in attenuating materials were characterized. Activated in a low thermal neutron flux to a nominal activity of 8.5 ± 0.9 mCi, the prototype source contained a low activity of contaminating isotopes after activation, and the titanium encapsulation was sufficient to contain the ⁷⁵Se activity. With the feasibility of a ⁷⁵Se‐based HDR brachytherapy source proven, the creation a source at the desired 20 Ci will be pursued further.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

Supporting Information

Supporting Information

Kalinowski J, Tal O, Reid J, et al. Development and characterization of a prototype selenium‐75 high dose rate brachytherapy source. Med Phys. 2025;52:e18088. 10.1002/mp.18088

DATA AVAILABILITY STATEMENT

Authors will share data upon request to the corresponding author.