Dosimetric implications of the potential radionuclidic impurities in 153Sm-DOTMP

Thehuman internal dosimetry of the radionuclidic impurities of samarium-153 in a new bone-seeking radiopharmaceutical, 153Sm-1,4,7,10tetraazacyclododecanetetramethylenephosponic acid (153Sm-DOTMP), has been estimated from preclinical data. The effective dose from the impurities in lower-specific-activity 153Sm is less than 17% of the effective dose from pure Sm-153. It has a background-equivalent radiation time for a dosage of 37 MBq/kg of less than one-half year.


Introduction
A new radiopharmaceutical, 153 Sm-labeled 1,4,7,10tetraazacyclododecanetetramethylenephosponic acid (PubChem, 2022) ( 153 Sm-DOTMP), is currently undergoing clinical trials for the treatment of bone cancer and bone cancer metastases. It is prepared using lower-specific-activity 153 Sm. An estimate of its human internal dosimetry that was based upon preclinical data and the properties of 153 Sm demonstrates that its biodistribution, which is primarily long-term uptake in the skeleton, imparts a radiation absorbed dose predominantly to the bone, the red marrow, and the urinary bladder (Simon et al., 2012). Samarium-153 is known to have long-lived radionuclidic impurities (Ma et al., 1996;Kalef-Ezra et al., 2015;Naseri et al., 2021). These arise from its production by neutron activation of enriched stable 152 Sm, which may still contain small amounts of several other isotopes and elements, as well as subsequent activation of its daughter, 153 Eu, and of the activation products themselves (Ma et al., 1996;Kalef-Ezra et al., 2015). The DOTMP chelant binds these impurities with the same high degree of efficiently as it does 153 Sm (Simon et al., 1991). Thus, their contributions to the internal dosimetry of 153 Sm-DOTMP are both an important consideration clinically, given their persistence in the skeleton, and a relatively straightforward question to answer, given estimates of the uptake and clearance of 153 Sm-DOTMP in various source organs.

Materials and methods
The significant radionuclidic impurities of reactor-produced 153 Sm are tabulated in Table 1 ( Kalef-Ezra et al., 2015;Naseri et al., 2021). The relative concentrations of the radionuclidic impurities in four sources of 153 Sm -activation of naturally abundant and of enriched 152 Sm as reported by Naseri et al. (Naseri et al., 2021), and activation of higher-specific-activity (40.8 GBq/mg) and of lower-specific-activity (8.77 GBq/mg) 153 Sm using 99% or more enriched 152 Sm by the University of Missouri Research Reactor (MURR) in Columbia, Missouri (Ma et al., 1996) as measured by the authors -are given in Table 2. The neutron flux and activation time for the Naseri samples are inferred from that same institution's earlier report of producing 153 Sm with a specific activity of 12.8 GBq/mg (Naseri et al., 2011).
A dose of 153 Sm-ethylenediamine tetra(methylene phosphonic acid) ( 153 Sm-EDTMP, tradename Quadramet) that had been made with higher-specific-activity 153 Sm from MURR and had been calibrated for 5.55 GBq (150 mCi) on January 7, 2009 with a stated expiration 56 h after calibration was analyzed 157 months after calibration. A sample of lower-specificactivity 153 Sm was obtained by the authors from MURR and allowed to age for 64 days (or 33 half-lives of 153 Sm) by which time the initial activity of the 153 Sm had decayed by ten orders of magnitude so that it would not interfere with the peaks of the long-lived impurities.
The impurities in the MURR samples were measured using high resolution gamma spectroscopy. This was performed with a germanium crystal detector [Canberra GC2519, Mirion Technologies, Meriden, CT] coupled to a multichannel analyzer [Easy-MCA, Ortec Advanced Measurement Technology, Oak Ridge, TN]. Gammavision software [Version 7.02.01, Ortec Advanced Measurement Technology, Oak Ridge, TN] was used to analyze the spectra. The system was calibrated for energy and efficiency with a NIST-traceable multi-nuclide source. The samples were then counted for 4800 s at a distance of 11.4 cm from the detector. The activities of 153 Gd and 156 Eu in the higher-specific-activity sample had to be estimated based upon the measured activities of the other isotopes of europium and the activation parameters for they had decayed to an undetectable level because of their relatively short half-lives compared to the age of the sample.
The standard Medical Internal Radiation Dose (MIRD) schema (Loevinger et al., 1991) was employed in this study. A preclinical investigation of 153 Sm-DOTMP in rats has previously been reported by the authors (Simon et al., 2012). That study was conducted with the approval of the IsoTherapeutics Group LLC Animal Care and Use Committee. The raw data from that study are decay-corrected, biological clearance data. Those data have been re-analyzed in the present study by applying the physical decay of each radionuclide and then fitting curves to the decayed data for each radionuclide. The parameters of these fits are given in the appendix. The time-integrated activity coefficients, or residence times, which represent the areas under the normalized time-activity curves, were converted to human values through Equation (8) in (Macey et al., 2001) using the murine organ masses from the earlier report and the organ masses of the ICRP 89 (International Commission on Radiological Protection, 2002) Adult Male and Adult Female models, which are the most up-to-date models that are incorporated into OLINDA/EXM Version 2.2 (Stabin et al., 2005;Stabin and Farmer, 2012). These time-integrated activity coefficients are given for each source organ and radionuclide in Table 3 for the Adult Male model and in Table 4 for the Adult Female model.
The present report takes a few departures from the original analysis of the raw data (Simon et al., 2012) in the determination of the time-integrated activity coefficients.
The time-activity curves were integrated for fifty years rather than to infinity in order to be consistent with the definition of committed dose equivalent by the US Nuclear Regulatory Commission (US Nuclear Regulatory Commission, 2021).
When the fits to the biological data that had been decayed with the physical half-life of a particular radionuclide yielded an effective half-life that exceeded the physical half-life of that radionuclide, its physical half-life was used as the effective half-life.
The biological data for the liver have a unique time course among the source organs. They suggest a rapid clearance combined with a slow uptake as shown in Fig. 1. This is consistent with the rapid clearance of 153 Sm-DOTMP from the blood along with the accumulation of unchelated 153 Sm by the liver (O'Mara et al., 1969;Goeckeler et al., 1987;Banerjee et al., 2005). Despite the assumption of the longest possible biological half-life from that last datum onward, the uptake in the liver is a miniscule fraction of the total administered activity. The asymptotic value of the superimposed uptake curve in Fig. 1 is 0.18% of the administered activity, which illustrates the high binding efficiency of the DOTMP chelator even at the very low chelant-to-metal ratio of 1.5:1 that was used in this preclinical study (Simon et al., 2012). The biological time-activity data from the liver were analyzed by decaying them for each radionuclide, calculating the area under the curve out to the last datum (that is, the 48-h measurement) by a trapezoidal fit, and assuming single exponential decay with the physical half-life of the radionuclide from that time onward.
The intestines and the stomach had no appreciable activity remaining in the 48-h datum, hence just the area under the trapezoidal fit was used for those two source organs.
The rat-to-human organ mass-based conversion of the time-integrated activity coefficients was done separately for the ICRP 89 Adult Male and Adult Female models. The timeintegrated activity coefficients for blood and muscle were combined into the remainder of the body term. It is not practical in rat studies to get true whole-body count data from which a whole-body time-integrated activity coefficient could be derived for the calculation of the time-integrated activity coefficient of the remainder of the body. The time-integrated activity coefficient of the large intestine was apportioned one-quarter each to the right colon and rectum and one-half to the left colon. The time-integrated activity coefficient of the skeleton was apportioned 38% to the cortical bone and 62% to the trabecular bone (Breitz et al., 2006).

Results
The time-integrated activity coefficients were analyzed with the OLINDA/EXM software [Version 2.2, Hermes Medical Solutions, Stockholm, Sweden] The resulting equivalent doses to the target organ for the ICRP 89 Adult Male and Adult Female models are given in Table 5 and Table 6. The equivalent doses to each target organ from 153 Sm including the impurities in each of the preparations and the equivalent dose to each target organ from just the impurities alone in each preparation are given in Table 7 and Table 8.

Discussion
Although the doses per unit activity from the impurities are often higher than those from 153 Sm, as shown in Tables 5 and 6, the small activities of the impurities compared to that of 153 Sm make their actual effects much weaker in practice, where they are often one or two orders of magnitude less, as shown in Tables 7 and 8. Three target organs are especially affected by the charged particle emissions of 153 Sm: the osteogenic cells, the red marrow, and the urinary bladder wall. This situation arises from the accumulation of about 40% of the administered activity in the skeleton with an effective half-life that is close to the physical half-life while the remainder of the administered activity is cleared rapidly through the urinary system. The doses from the impurities are highest in the two target organs that receive the highest doses from 153 Sm, namely the osteogenic cells and the red marrow. This is because of the long biological half-life of the uptake in the skeleton, which was treated as infinite.
The relatively short physical half-life of 153 Sm compared to those of the impurities means that the activity of the impurities will grow compared to that of 153 Sm as the preparation ages. The US Pharmacopeia limits the activity of 154 Eu to 93 ppm of that of 153 Sm (i. e., 0.093 μCi of 154 Eu per mCi of 153 Sm) in 153 Sm-EDTMP. It limits the sum of the activities of all other radionuclidic impurities to 0.1907% of that of the 153 Sm (US Pharmacopeia, 2013) The rationale for these values is unknown to the authors, but at these limits, the effective dose from 154 Eu would be 16.0% of that from pure 153 Sm in the Adult Male model and 16.6% of that from pure 153 Sm in the Adult Female model. The times to reach these limits for the four preparations of 153 Sm are given in Table 9. It is noteworthy that the stated expiration of the dose of 153 Sm-EDTMP that had been prepared with higher-specific-activity 153 Sm was 56 h after calibration, whereas the analysis of that sample suggests that it should have been only 28 h after calibration. The authors of this report postulate that the expiration time of commercially-prepared 153 Sm-EDTMP is a standard one that was derived from the analysis of some samples of higher-specific-activity 153 Sm that had been performed during the original development of the drug. The actual activities of the impurities are impossible to measure by any practical means during the preparation of a particular dose of a 153 Sm-labeled radiopharmaceutical.
The authors have found only one previous report of the measurement of 153 Gd in a sample of 153 Sm (Nuñez et al., 2015) although its presence is not unexpected (Zhang et al., 2007). The dominant photopeak of 153 Sm is at 103 keV with an abundance of 29.8%. Gadolinium-153 has a photopeak at 97.4 keV with an abundance of 29.0% and another at 103 keV with an abundance of 21.1% (International Commission on Radiological Protection, 2008). This spectral overlap has been recognized as an analytical problem in the production of 153 Gd when there is 152 Sm present in the 152 Gd target (Holden, 1986). In the case of a small amount of 153 Gd in a sample of 153 Sm, the photopeak of 153 Sm overwhelms those of 153 Gd, and thus 153 Gd at low concentrations cannot be measured with gamma ray spectroscopy until the 153 Sm has decayed to a negligible level compared to that of 153 Gd. In a spark-source mass spectroscopic analysis of a sample of 99.47% enriched 152 Sm that was commissioned by the authors, gadolinium was found with an upper limit of 200 ppm. This is the most likely source of the 153 Gd impurity in 153 Sm. The MURR higher-specific-activity sample was too old for any detectable 153 Gd to remain and thus the amount was estimated based upon the irradiation conditions and the amount of 153 Gd that had been measured in the lower-specific-activity 153 Sm from MURR.
Tables 5 and 6 can be used to estimate the dosimetry of other preparations of 153 Sm-DOTMP that might contain different concentrations of the impurities than the four sources of 153 Sm that were analyzed here. Based upon the relative doses of the impurities to the dose from pure 153 Sm, the preferred source for preparing 153 Sm-DOTMP would be a lower-specific-activity 153 Sm using a highly-enriched 152 Sm target that had been activated for a relatively short time in a relatively low neutron flux.
At a dosage of 37 MBq/kg of 153 Sm-DOTMP, i.e. 2.7 GBq for a 73 kg adult man and 2.2 GBq for a 60 kg adult woman, which is the standard dosage of 153 Sm-EDTMP for bone pain palliation, the effective dose from the impurities in the MURR lower-specific-activity 153 Sm to the adult man is 1.63 mSv or a background-equivalent radiation time (BERT) (Nickoloff et al., 2008) of about 160 days at sea level. For the adult woman, the effective dose from the impurities in the MURR lower-specific-activity 153 Sm is 1.54 mSv or a BERT of about 150 days at sea level. What is more, since most of this dose is imparted over the course of 50 years, the dose rate is well below the natural background dose rate.
Past clinical uses of the radiopharmaceutical 153 Sm-EDTMP, which must be prepared with higher-specific-activity 153 Sm, have included high dosages of up to 50 GBq for the ablation of bone marrow (Bartlett et al., 2002) and fifteen or more administrations of low dosages such as 18.5 MBq/kg (0.5 mCi/kg) over the course of four years or longer for durable bone pain palliation and tumor treatment (Sinzinger et al., 2009). If the newer radiopharmaceutical 153 Sm-DOTMP were to be used in similar fashions, the reduction in the doses to the target organs and the effective doses from the radionuclidic impurities in lower-specific-activity 153 Sm compared to higher-specific-activity 153 Sm would be even more favorable when the total administered activity of 153 Sm is large.

Conclusion
The effective dose from the radionuclidic impurities in the lower-specific-activity 153 Sm that is used to make 153 Sm-DOTMP is a small fraction of the overall effective dose.

Funding:
This work was supported in part by a grant to IsoTherapeutics Group, LLC from the National Cancer Institute, R43CA150601, a gift earmarked for the research of REW to The University of Texas MD Anderson Cancer Center by IsoTherapeutics Group, LLC, and a consulting agreement between QSAM Biosciences, Inc., and IsoTherapeutics Group, LLC. The authors thank Nicholas C. Xydas, CNMT, for his technical assistance. Fitting parameters for the biological data decayed by the different radionuclides. The fits were either to a single exponential, a bi-exponential, or a trapezoidal approximation to the data followed by a single exponential with the physical half-life of the radionuclide when the last datum was not nil. Y 0 and T eff are the parameters of a single exponential fit, Y 0 , T eff0 , Y 1 and T eff1 are the parameters of a bi-exponential fit, and A is the area under the trapezoidal fit to the data out to 48 h followed by an exponential with the initial amplitude Y 48 , which is the last decayed measured datum, and the physical half-life of the radionuclide. The intestines and the stomach had no appreciable activity in the 48-h datum, hence just the area under the trapezoidal fit was used.

Fig. 1.
The decay-corrected uptake of 153 Sm-DOTMP in the rat liver as a fraction of the administered activity (Simon et al., 2012). Those data are shown as crosses. A single exponential decay curve was constrained to approach zero asymptotically and was fit to the first two data, shown as squares. An uptake curve was constrained to have a nil value at time zero and was fit to the last two data shown as circles. These fitted curves illustrate the postulated rapid clearance of the blood pool that perfused the liver followed by the later accumulation in the liver of a small amount of free 153 Sm.  Table 2 Relative activities of the radionuclidic impurities from the four production sources of Sm-153. Naseri et al. (2021) do not report the flux of their neutron source or the length of time over which their targets were activated, so these were inferred from (Naseri et al., 2011).. (SA = specific activity, NR = not reported, ND = not detected, and underlined = estimated.)

Neutron Flux
Activ.  Table 4 Time-integrated activity coefficients (in hours) for the ICRP 89 Adult Female model.