Impact of shield location on staff and caregiver dose rates for I-131 radiopharmaceutical therapy patients

The goal of this study is to investigate the effect of the location and width of a single lead shield on the dose rate of staff and caregivers in a hospital room with an I-131 patient. The best orientation of the patient and caregiver relative to the shield was determined based on minimizing staff and caregiver radiation dose rates. Shielded and unshielded dose rates were simulated using a Monte Carlo computer simulation and validated using real-world ionisation chamber measurements. Based on a radiation transport analysis using an adult voxel phantom published by the International Commission on Radiological Protection, placing the shield near the caregiver yielded the lowest dose rates. However, this strategy reduced the dose rate in only a tiny area of the room. Furthermore, positioning the shield near the patient in the caudal direction provided a modest dose rate reduction while shielding a large room area. Finally, increased shield width was associated with decreasing dose rates, but only a four-fold dose-rate reduction was observed for standard width shields. The recommendations of this case study may be considered as potential candidate room configurations where radiation dose rates are minimized, however these findings must be weighed against additional clinical, safety, and comfort considerations.


Introduction
Radiopharmaceuticals, specifically designed to deliver radiation to target cells, may be used to treat cancer and other diseases. When administered to patients, these molecules can either kill cancer cells directly or activate other molecules to do so, thereby providing a therapeutic approach for treating various diseases [1,2]. Cancer hospitals may treat patients with significant quantities of radiopharmaceuticals, which can expose hospital staff, visitors, and other patients to radiation unless protection measures are undertaken. Thus, radiation protection is critical when providing care in a cancer hospital [3]. Optimisation and limitation of radiation doses must be considered in the management of radiopharmaceutical therapy (RPT) patients because the radiation sources used to treat cancer have the potential to create significant doses for staff and caregivers who spend time near RPT patients [4]. Appropriate planning and facility design are necessary to manage and maintain doses to caregivers as low as reasonably achievable below regulatory limits. To accomplish this goal, the size and positioning of treatment rooms, patient position, patient orientation, use of mobile lead shields, and caregiver positioning are essential considerations when optimizing and limiting the dose to caregivers and medical workers [3,5,6]. Appropriate radiation shielding is paramount when delivering RPT care in a cancer hospital.
Iodine-131 (I-131) is a radioactive isotope that has multiple medical applications, such as treating hyperthyroidism, thyroid ablation for cancer treatment, and radiolabeling molecular targeting agents like metaiodobenzylguanidine (MIBG) [7] for targeting and killing cancer cells. In some cases, significant activities are administered, sometimes beyond 37 GBq, resulting in considerable dose rates to the patient and, as this paper is concerned, individuals near the patient. To protect family members and members of the public, these patients may be held in the hospital room for a few days until the dose rate drops to acceptable levels [8]. Upon patient release, the family is instructed on methods to reduce their dose through minimizing the time spent with and maximizing the distance from the patient.
Because these high-dose-rate patients are held in the hospital, reducing radiation dose rates is of utmost importance to protect the health and safety of staff, caregivers, and nearby patients. Implementing well-designed safety protocols to reduce radiation exposure is essential to ensure that all individuals in the hospital environment are safeguarded from the potential adverse effects of radiation. Standard shielding calculations are excellent at estimating the dose rate but do not consider radiation which passes around the lead shielding [9]. In a typical hospital room, scattering from the walls, ceilings, and floor constitute a significant fraction of the dose to the caregiver [10]. Unlike the extensive body of work on therapy vault shielding design, little rigorous work has been done to analyse optimal radiation shield size and positioning in a hospital room. Thus, the overarching goal of this paper is to investigate the effectiveness of various shielding configurations in reducing the dose rate from an I-131 patient to a caregiver.

Materials and methods
Monte Carlo radiation transport software was used to simulate the geometry and materials in a typical hospital room to support optimizing dose rates for caregivers using a single, mobile lead shield. These findings were then validated using real-world radiation detector measurements near an I-131 patient in a hospital room. The International Commission on Radiological Protection (ICRP) adult male voxel phantom was used to simulate I-131 source in the patient's thyroid, bladder, and remaining tissue. Monte Carlo N particle code (MCNP) [11,12] was used for radiation transport calculations employing the air kerma track length estimator as a surrogate for caregiver dose. Kerma is a reasonable approximation to radiation dose under conditions of charged particle equilibrium, which is likely to exist for the caregiver in the case of I-131 treatments. This is because the human body is primarily composed of low atomic number elements like hydrogen, carbon, nitrogen, and oxygen. Additionally, the primary and scattered radiation from I-131 has relatively low energies, typically in the range of a few 100 keV.
Material compositions and thicknesses were based on representative hospital treatment rooms used for RPT. A standard mobile lead shield was modelled in various configurations to investigate its effect on air kerma rate. As a validation step, the air kerma rate at 1 m from a point source was estimated using MCNP, and this value was compared to published values. Next, the air kerma rate at 1 meter from an I-131 patient was calculated. The shielding efficiency associated with two patient orientations and two radiation shield positions was investigated. Finally, the effect of using a perfect theoretical shield (i.e. zero shield transmission) of various widths was analysed to determine the role of lead shield size in caregiver dose rates.
In this work, the ICRP 110 [13] adult male reference computational voxel phantom was employed to realistically simulate the I-131 source distribution and patient self-shielding. This computational phantom was constructed with the anatomical and physiological characteristics defined in the ICRP Task Group on Reference Man [14,15] and is an idealised person for use in radiation transport modelling. Based on biokinetic modelling published in the ICRP Occupational Intake of Radionuclides Part 1, I-131 activities are distributed in voxel source regions within the phantom, representing organs and tissues such as the thyroid, urinary bladder, and gastrointestinal tract. The Monte Carlo source was then defined by assigning emission probabilities of I-131 in each patient phantom voxel to match the biokinetic model output.
MCNP version 6.1.1 was used for all radiation transport calculations associated with this work. Air kerma was estimated in this work using the MCNP F6 tally [12] to determine the dose rates in the room. This tally works by sampling the average track length of photons within the receptor location folded with the energy-dependent kerma factors within the MCNP code. The Monte Carlo simulation was performed with explicit transport of only photons. That is, air kerma contributions from secondary electrons are not directly considered by explicit electron transport but are estimated based on the MCNP thick target bremsstrahlung model. The simulations were performed using MODE P and not MODE P E). Thus the dose estimates were conducted under the assumption that kerma is numerically equal to absorbed dose where photoelectric effects and Compton scattering was fully simulated. Sufficient events were simulated such that the statistical error associated with the Monte Carlo-derived air kerma was less than 5% at the caregiver location [11].
The primary quantity used in this work to compare dose rates is air kerma, K a . In this work, the numerical value of air kerma is a reasonable approximation of the effective dose to ICRP reference persons under radiopharmaceutical exposure scenarios. It is well established that there is a correlation between organ dose and air kerma for multiple incident radiation types (photons, electrons, neutrons, muons, etc) and incident radiation geometries (A.P., PA, left-lateral (LLAT), etc) [16]. In addition, air kerma is a reasonable estimate of effective dose (i.e. H * (10)) when dealing with photons above 50 keV and when the receptor is facing the source [17]. Because the I-131 emission energies meet these criteria, air kerma is used as a direct surrogate for dose rates in this work. Exposure rate measurements can also be used to approximate dose rate to the air under the assumption that one cGy h −1 is numerically approximately the exposure rate in Roentgens/hour times a factor of 0.88 [18].
The geometry of the hospital room employed in this work was based on a frequently occupied radiopharmaceutical inpatient room of Memorial Sloan Kettering. The room dimensions were measured to be 243 cm in height, 280 cm in width, and 610 cm in length. In this study, the ceiling, one wall, upper floors, and lower floor of the room were simulated as concrete, while the three remaining walls and false ceiling were simulated as gypsum. This was done to accurately represent the room's physical characteristics and provide a realistic simulation of the environment. All elemental compositions and densities included in the Monte Carlo simulation were consistent with the PNNL Compendium of Material Composition Data for Radiation Transport Modelling [19].
The simulated lead shield employed in this work was based on the design of the existing mobile lead shields currently in use at Memorial Sloan Kettering Hospital [20]. A mobile lead shield, with a height and width of 137.16 and 91.44 cm, respectively, and a thickness of 1.27 cm, was simulated and employed for comparing patient orientation and shield positioning. The shield was placed 46 cm above the concrete floor. The effect of shield width on caregiver dose rates was investigated by modifying the shield width from 0 cm to 180 cm. The impact of shield thickness was also investigated by replacing the lead shield with a perfect absorber (terminating all photons which enter the shield by setting the MCNP importance to zero).
The shielding effectiveness was investigated as a factor of patient orientation, shield positioning, and shield width. The two patient-caregiver orientations investigated in this work were the LLAT and caudal direction (i.e. feet pointing towards caregiver), as described in ICRP Pub. 116 [16]. For each of these two patient-caregiver orientations, air-kerma values were estimated for three configurations: no lead shield, the shield 30 cm from the patient, and the shield 30 cm from the caregiver. The 30 cm distance represents the closest patient-to-shield, and caregiver-to-shield measurements. These shielding configurations are shown in figure 1. Additionally, the shielding efficiency associated with shields of different widths was assessed for the patient in the LLAT orientation with respect to the caregiver. For this assessment, the shield was placed close to the patient, and shield lengths up to 180 cm were simulated.
Air kerma rates for a point source in the air were calculated and compared to previously published values to validate the Monte Carlo model. In addition, the unshielded air kerma rate at 1 m from an I-131 patient was calculated as this quantity has regulatory implications concerning patient release criteria. Finally, air kerma values associated with the previously described shielding configurations were calculated. These kerma values, which estimate caregiver dose rates, employed a 14 cm diameter sphere of air located 110 cm above the hospital room floor. These results were compared to previous studies to validate the accuracy of the Monte Carlo models developed in this study.
As an additional validation, the Fluke 451B Ion Chamber was used to generate dose rate measurements from an I-131 patient in a hospital room at three shield positions. First, the unshielded distance from the patient to the caregiver was 370 cm. Next, the shield was positioned, and measurements were taken at 120, 210, and 320 cm from the patient, corresponding to the near and far simulated measurements with an additional position in between. At each shield position, a reading was taken both above and behind the shield and at the caregiver's seat.

Results
Our findings indicate that air kerma rates were significantly influenced by both the positioning of the patient-caregiver and the placement of the shielding. Six LLAT and Caudal patient-caregiver orientations were considered, as shown in figure 1, with the resultant air kerma rates presented in table 1. Across all shielding configurations, lower air kerma rates were consistently observed when the shield was positioned closer to the caregiver than to the patient.
When comparing the LLAT and Caudal patient-caregiver orientations, the minimum dose rates corresponded to the Caudal orientation. The configuration which yielded the most substantial reduction in kerma rate (82% reduction compared to unshielded conditions) was the Caudal patient-caregiver orientation with the shield placed near the caregiver.
The influence of the mobile lead shield's width and thickness on caregiver dose rates was also investigated. Air kerma values for the LLAT orientation with shield widths ranging from 0% to 100% (0 cm to 180 cm) of the patient height were simulated and illustrated in figure 2.
Dose rates for the validation experiment were measured at three shield positions in a hospital room (presented in table 2). Consistent with the previous findings, the lowest dose rates corresponded to the shield placement closest to the caregiver across all configurations. As a specific example, when the shield was moved away from the patient and towards the caregiver's seat, dose rates decreased from 7.8 µGy h −1 to 6.0 µGy h −1 . These results corroborate the observed trend in dose rates as the shield is moved away from the patient, affirming the validity of both our experimental measurements and the Monte Carlo simulations.
Finally, the air kerma rate at 1 meter from an I-131 point source with 1 GBq activity in the air was measured to be 70.7 µGy h −1 per GBq. In comparison, the air kerma rate measured 1 meter anterior to a standing patient administered 1 GBq of I-131 yielded a rate of 38.5 µGy h −1 per GBq. This 46% reduction in dose rate is attributed to the combined effects of patient self-shielding and the spatial distribution of the source, aligning with previously published data [21].

Discussion
Recent studies emphasise the importance of considering the effects on family caregivers when calculating dose rates from RPT patients. In recognition of the comfort and care they provide, as well as the likely infrequency of their exposure, NCRP Report 155 and ICRP Report 94 recommend higher dose limits for family caregivers than for the general public [8,22]. U.S. regulatory guidance offers several methods to calculate the dose rate from RPT patients, including replacing the patient with an equivalent-activity point source and calculating the dose based on the specific gamma-ray dose constant (gamma factor) [23]. However, this approach has drawbacks such as oversimplifying the source geometry, neglecting attenuation from the patient's body tissues, and not accounting for scattering from walls, floor, and ceiling. Monte Carlo simulations, on the other hand, effectively compute estimated shielded caregiver doses, considering the full physics complexity of the problem. As a result, the point-source assumption falls short of accurately characterizing the variables considered in this work [6,24].
Patients held in the hospital until they are releasable may benefit from the comfort and care of family caregivers, even though these caregivers may be exposed to higher doses than the general public. Consistent with the recommendations from NCRP report 155 and ICRP report 94, the US Nuclear Regulatory Commission (NRC) permits approved visitors of patients who cannot be released to receive a dose up to 5 mSv (total effective dose equivalent) [25], compared to the 1 mSv limit for other members of the public, and may grant exemptions in caregiver situations where the dose limits for members of the public may be insufficient [26]. Despite regulatory allowances for higher caregiver radiation doses, optimizing the use of shielding to keep doses as low as reasonably achievable remains crucial.
Radiation transport simulation results comparing the relative shielding efficiency are shown in table 1 as a function of patient orientation. The caudal orientation was associated with a lower dose rate than the LLAT   orientation. Additionally, the caregiver dose rate was lower when the shield was placed closer to the caregiver than the patient. The dose rate was lowest when the caregiver was in the caudal direction and when the lead shield was placed near the caregiver. However, even though the dose rate was the lowest, this configuration may not be optimal due to other factors. Factors such as the caregiver's occupancy in various places in the room and the dose rate to the caregiver should also be considered when evaluating room shielding, especially when positioned near the patient. If the caregiver spends an extended amount of time in one place in the room, it might be advantageous to shield that area. However, the caregiver typically will move around the room throughout the patient's hospital stay. The caregiver may need to approach the patient and thus be within close proximity. In such a case, if the mobile lead shield is not present between the caregiver and patient, this will result in a high, unshielded dose rate. The strategy of shielding only the caregiver area has been studied to assess its efficacy in reducing occupational doses. Monte Carlo simulations were conducted to evaluate the dose rate in the room when the shield was placed in different configurations. These simulations showed that when the shield was placed only around the caregiver area, a large area in the room remained unshielded and had a relatively high dose rate. This configuration resulted in a larger area where hospital staff may be exposed to a higher dose rate, leading  (2) the shield placed near a patient (bottom). The simulated activity of I-131 in the patient was 10 GBq, and dose rates are presented in mSv/hr. Plaching the shield further away from the patient (top) leads to a lower dose rate next to the shield, but the dose rate the remainder of the room is much higher compared to where the shield is placed near to the patient (bottom).
to higher occupational doses. The results of the Monte Carlo simulations are illustrated in figure 3, which demonstrates the phenomenon associated with shield positioning.
This study's findings suggest that using a mobile lead shield of a standard thickness is an effective way to reduce the dose rate from I-131. This is because the standard lead thickness was found to be effective as a primary shield based on the comparison of the standard shield thickness to the theoretical ideal absorber. Furthermore, increasing the width of the shield was associated with a large reduction in the dose rate as shown in figure 2. This implies that further reductions of the dose rate may be obtained by reducing scatter from the ceiling, walls, and floor. To do this, one could shield the scattering surfaces or place the patient and caregiver in a larger room where the scattering services surfaces are further away from the caregiver. This could be especially beneficial for staff exposed to radiation during their duties, as it could help reduce their exposure to radiation and protect them from any potential health risks.
The effectiveness of the lead shield was found to be similar to that of an ideal theoretical absorber. However, this should not be generalised to other radionuclides [27]. Photons with higher energies, such as those from Cobalt-60, are less attenuated through the lead shield. Therefore, in order to minimize the dose, it is important to investigate the use of a thicker shield. Nevertheless, the results of this study can be applied to photon emitters with emission energies lower than I-131. It is important to note that the lead shield may not be as effective for higher energy photons, and thus a thicker shield may be necessary to ensure adequate protection.
Several published works have described very high I-131 dose rates and associated safety precautions. According to Flori et al [28], the highest dose rate is at the bedside and in front of the lead shield. Estimates indicate that a team member can remain in front of the lead shield for the first five hours after infusion, the most significant radiation exposure time, before reaching 5 mSv (500 mrem). The next highest dose rate is at the head of the bed, although the dose rate was often less than 25% of the dose rate in front of the lead shield. Their reported shielded/unshielded ratio is consistent with the Monte Carlo results of the current work.
According to Chu et al [29], high-dose MIBG Therapy was possible using lead shielding, contamination reduction, and education of patients, parents, and caregivers. In that study, radiation protection methods may result in a lower radiation exposure rate for parents, caregivers, and the surrounding environment. The mean dose administered to patients was 17 ± 11 GBq. The mean exposure to caregivers throughout the patient's stay was 0.98 ± 0.58 mSv, and the mean exposure for parents was 0.54 ± 0.32 mSv/parent when two parents were involved. As a result of shielding, maximum exposure rates to unrestricted areas surrounding the patient's room were less than 0.02 mSv h −1 for patients who were injected with a dose of less than 30 GBq. Through educating caregivers and staff on radiation safety protocols and the use of lead shielding and implementing safety precautions, the authors' establishment was able to administer high dose MIBG therapy while maintaining low exposure rates.
Cougnenc et al [30] reported that many preventative measures could be implemented to reduce radiation exposure associated with 131I-MIBG therapy. These measures included a mobile protective lead screen and toilets in patients' rooms, a lead-shielded injection trolley, and homemade syringe shields. Caregivers were allowed to stay in a room next door to the patients during their stay at the hospital to minimize radiation exposure. The effective dose was estimated using a sodium iodide NaI(TI) scintillation detector installed in the ceiling of the patient's room. For relatives, the effective dose ranged from 0.018 to 2.79 mSv while at the hospital. Caregivers were given basic radiation protection instructions for the time they spent with patients in the hospital and instructions for coming home from treatment which helped minimize exposure. The mean effective dose for night nurses and day nurses was 33.6 µGy and 20.2 µGy, respectively, and external gamma ray dose rate of nuclear physicians and paediatric oncologists was <5 µGy d −1 . Based on Cougnenc et al, the use of lead shielding can be a part of an overall strategy of procedures and precautions set in place for high-dose I-131 MIBG therapy.
Finally, Chuamsaamarkkee et al [10] presented results indicating a variety of ambient dose range equivalents in different locations relative to the patient. These included measurements from the floor beneath the patient's bed, which ranged from 131.64 to 149.61 µGy·h −1 , and from the patient's bed itself, which ranged from 73.39 to 140.00 µGy·h −1 . Further measurements were taken at the location of the caregiver, where the ambient dose range equivalents were found to be significantly lower, ranging from 22.97 to 30.57 µGy·h −1 . Additionally, the highest observed directional dose rate equivalents at the caregiver's bed was 30.8 µGy·h −1 , with an uncertainty of ±5.9 µGy·h −1 . These findings are difficult to compare with the dose rates in the current work because the patient-to-caregiver distance were not published by Chuamsaamarkkee. Their larger dose rates are likely due to smaller treatment room which resulted in the higher reported dose rates.
In the current paper's context, the optimal placement of the shield depends on the primary positions of the patient and caregiver, which may change several times a day. However, moving the lead shield during the patient's stay may not be optimal from an occupational and patient safety standpoint. There are falling/crushing risks associated with moving unwieldy, heavy shields while the room is occupied. The mobile lead shield has a relatively large floor footprint to ensure that the shield's centre of gravity remains comfortably within the base of support. Thus, changing the position of the shield may be difficult when there is minimal floor space for manoeuvrability. The lead shield supporting legs also constitutes a significant tripping hazard if placed inappropriately in the hospital. When optimizing shield positioning, these physical risks should be weighed against the radiological risk. Moving the mobile shield continuously with the caregiver would provide the lowest theoretical dose but may be impractical. The caregiver wearing a lead apron as an alternative to the mobile lead shield provides a minimal benefit because standard 0.5 mm lead-equivalent aprons shield only 15% of the primary I-131 radiation emission based on the lead x-ray mass attenuation coefficient.

Limitations and future work
This work has several limitations that suggest potential avenues for future research. Firstly, the study only considers one fixed-height lead shield and does not explore other geometric strategies for reducing scatter. The use of two shields is standard practice, and investigating the optimal configuration with a third shield could be a fruitful avenue for further investigation. Alternative strategies for reducing scatter could include increasing the height of the top of the shield and lowering the gap between the bottom of the shield and the floor, shielding primary photons before they reach the scattering surfaces, or shielding lower-energy photons after they are scattered from the room surfaces. Shielding configurations, such as those which seek to minimize ceiling scatter, are promising options which should be investigated. Radiation dose may also originate from caregiver contamination and subsequent ingestion. This important radiation pathway has been considered in other publications [31]. Tangential to this point, the source distribution in the patient was assumed to be static over the course of this work. This assumption ignores the time-dependent movement of the radionuclide within the patient's body, which could potentially affect the patient self shielding factor.
Secondly, the study only considers Monte Carlo radiation transport in one anthropomorphic computational phantom, which represents only adult males. It may not adequately capture the shielding phenomenon associated with paediatric patients, whose self-shielding characteristics may differ depending on height and BMI. Further research should investigate shielding effectiveness associated with paediatric patients and explore the optimal shield geometry for different patient sizes.
Thirdly, the study focused only on the attenuation of Iodine-131, and other radionuclides with higher energies may have photons that are less effectively attenuated by the shield. Shield positioning and thickness may need to vary depending on the radionuclide's energy, and further research should investigate the role of scattering and direct irradiation associated with mobile lead shields as a function of photon energy Fourthly, the study only considers a single room-geometry and structural composition. Larger or smaller rooms may have different scattering properties, and the corresponding optimal shielding strategy may differ. In addition, the composition of the floor and ceiling are essential considerations in this type of calculation because those surfaces are responsible for most of the radiological scatter. Understanding the contribution to caregiver dose as a function of room size and building materials is beyond the scope of this paper but can be a topic of future investigation.
A significant limitation of this study lies in its scope, which only considered situations where the caregiver is positioned behind the shield while the patient remains in bed. This focus does not consider scenarios involving ambulatory patients who move around the room, a factor which could significantly influence radiation exposure patterns and levels for caregivers, and thus the optimum use and positioning of the shield. To manage radiation exposure and optimise safety, it is crucial to educate patients about ALARA principles. Reducing the time spent in close proximity to the source of radiation, maximizing the distance from it, and utilizing effective shielding can all significantly decrease radiation exposure. In terms of future research, more detailed investigation into scenarios involving ambulatory patients is required. There is a need for research that takes into account the changes in exposure patterns when patients move around and how to optimally position the shield under such circumstances. Furthermore, it is important to consider how patient education can be improved to help minimize doses when they are mobile.
Finally, the study estimated caregiver's dose rates using air-kerma, which has limitations in calculating organ doses and effective dose. The current approaches do not consider dose rate inhomogeneities at different receptor locations. For example, the caregiver's torso shielding factor may differ from the shielding factor at the ankles and feet because the lead shield does not extend to the floor. This report only considers torso shielding; thus, further research should investigate the inhomogeneities of the radiation field associated with a mobile lead shield in relation to a partially shielded receptor.

Conclusions
Monte Carlo methods were used to investigate the effect of the position and width of a single lead shield on the dose rate to staff and caregivers in a room with an I-131 patient. Several conclusions were reached based on these Monte Carlo calculations and the validation measurements: 1) When the lead shield is placed near the caregiver, dose rates are lower than when the lead shield is placed near the patient; however, this strategy shields a much smaller area of the room. This finding was confirmed using ionisation chamber measurements in a hospital room with an I-131 patient. 2) Dose rates to staff and caregivers in the caudal direction are lower than those in the LLAT orientation.
3) About 25% of the unshielded I-131 dose rate to the caregiver was found to be due to photon scatter from the ceiling, floor, and walls. 4) When considering optimizing caregiver dose rates, priority should be given to shielding the entire width of the patient over increased shield-thickness when the lead shield is at least 1.26 cm thick. 5) The dose rate at a distance of 1 meter from the torso of an I-131 patient was calculated to be 54% of the dose rate relative to an I-131 point source of the same activity.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).

Funding
This research was funded in part through the National Institutes of Health/National Cancer Institute Cancer Center Support Grant P30 CA008748.