Radiochemical Quality Control Methods for Radium-223 and Thorium-227 Radiotherapies

Abstract

Background:

The majority of radiopharmaceuticals for use in disease detection and targeted treatment undergo a single radioactive transition (decay) to reach a stable ground state. Complex emitters, which produce a series of daughter radionuclides, are emerging as novel radiopharmaceuticals. The need for validation of chemical and radiopurity with such agents using common quality control instrumentation is an area of active investigation. Here, we demonstrate novel methods to characterize ²²⁷Th and ²²³Ra.

Materials and Methods:

A radio-TLC scanner and a γ-counter, two common and widely accessible technologies, as well as a solid-state α-particle spectral imaging camera were evaluated for their ability to characterize and distinguish ²²⁷Th and ²²³Ra. We verified these results through purity evaluation of a novel ²²⁷Th-labeled protein construct.

Results:

The γ-counter and α-camera distinguished ²²⁷Th from ²²³Ra, enabling rapid and quantitative determination of radionuclidic purity. The radio-TLC showed limited ability to describe purity, although use under α-particle-specific settings enhanced resolution. All three methods were able to distinguish a pure from impure ²²⁷Th-labeled protein.

Conclusions:

The presented quality control evaluation for ²²⁷Th and ²²³Ra on three different instruments can be applied to both research and clinical settings as new alpha particle therapies are developed.

Introduction

Radionuclides have long been used as tools to image and treat cancer. When systemically administered, specific radiopharmaceuticals may localize to sites of disease for diagnostic or therapeutic purposes. Presently, there is a growing interest in α particle emitting radiotherapy for cancer treatment. ²²³RaCl₂ (tradename Xofigo) is the first α particle therapy approved by the U.S. Food and Drug Administration for the treatment of bone metastatic castrate-resistant prostate cancer in men.^1,2

Radium-223 (²²³Ra) is a bone seeker, and it accumulates in sites of active bone turnover, as existing at metastatic bone tumors.^3,4 The α particles emitted by ²²³Ra and its daughters are characterized by a very short track range, irradiating bone metastatic prostate cancer cells and the surrounding tumor microenvironment, while largely sparing healthy tissue from fatal irreversible DNA damage.⁵

Thorium-227 (²²⁷Th) is also currently under investigation in clinical trials (NCT03507452, NCT03724747) for its potential in targeted α therapy. This radionuclide is the parent ion of ²²³Ra and can be readily chelated and conjugated to various targeting vectors (Fig. 1).^6,7 This enables specific targeting of ²²⁷Th to sites of soft tissue disease with molecular precision. In addition, its long half-life allows advantages for preparation, transportation, and administration.⁸

FIG. 1.

(A) Decay chain of ²²⁷Ac, which decays into ²²⁷Th and ²²³Ra. (B) Depiction of the anion exchange column used for the purification of ²²³Ra and ²²⁷Th. (C) Antibody radiolabeling with ²²⁷Th. ²²⁷Th is chelated to an antibody and purified using size exclusion chromatography.

²²³Ra and ²²⁷Th are produced from an ²²⁷Ac (21.7 years half-life) source and purified using an ion-exchange resin. Radionuclidic purity may be evaluated by γ-spectroscopy for subsequent labeling and radiopharmaceutical formulation.⁹ However, the assortment of daughters emitted through concatenated decay of these radionuclides leads to complicated chelation status analyses and characterization of the products, which are essential for reliable quality control of these radiopharmaceutical products. In addition, the complex (γ, β-, and α-particle) decay profiles of ²²⁷Th and ²²³Ra have overlapping emission spectra, further complicating individual radionuclide distinction.¹⁰

Although measurement by high purity Germanium detector (HPGe) is a simple method for quality control, and ²²⁷Th and ²²³Ra are easily distinguished, many community care centers and hospitals do not possess such equipment. At present, manufacturer determination of purity, specific activity, and activity is performed at a central manufacturing or compounding site and subsequently transported to medical centers for use.

A growing consideration for the field is how to properly validate novel radiopharmaceuticals for clinical application, specifically radionuclides that produce multiple daughters and long half-lives.^11,12 In this study, a series of characterizations on technologies typically accessible for radiopharmaceutical identification and determination of purity was undertaken, including measures from radio-thin layer chromatography (TLC), γ counting, and a solid-state α-particle detector (Table 1).

Table 1.

Most Significant α Particle Energies (Intensity >3%) of the Radioisotopes in the Decay Chain of Thorium-227 and Their Corresponding Intensities

227Th		223Ra		219Rn		211Bi		215Po		211Po
Energy (keV)	Intensity (%)	Energy (keV)	Intensity (%)	Energy (keV)	Intensity (%)	Energy (keV)	Intensity (%)	Energy (keV)	Intensity (%)	Energy (keV)	Intensity (%)
5700.8	3.6	5433.6	1.8	6425.0	7.5	6278.2	16.2	7386.1	100	7450.3	98.9
5708.8	8.3	5539.8	8.9	6552.6	12.9	6622.9	83.5
5713.2	4.9	5606.7	25.0	6819.1	79.4
5756.9	20.4	5716.2	51.2
5866.6	2.42	5747.0	8.9
5959.7	3.0
5977.7	23.5
6008.8	2.9
6038.0	24.2

All energies are recorded in keV.

²²⁷Th and ²²³Ra identification and quantification were demonstrated utilizing γ spectrometry and γ-counting and also an α-particle imaging spectrometer was evaluated. Novel use of a standard radio-TLC scanner for α-particle specific imaging was also evaluated. Through these methods, a quality control protocol for a ²²⁷Th-radiolabeled protein construct was developed.

Materials and Methods

All reagents and solvents were purchased from Sigma-Aldrich, unless otherwise noted. ²²⁷Th was supplied by the department of energy (Oak Ridge). All buffered solutions were prepared from metal-free water purified using Chelex 100 resin (BioRad). The handling of radioactivity requires specific caution and protection, and certification and protocol approval were provided by the Environmental Health and Safety Department at Washington University in St. Louis (Radioactive Materials Authorization 1169-01). Additional details for methods are found in the Supplementary Data, where indicated.

Characterization of ²²³Ra and ²²⁷Th using radio-TLC

Radium-223 and Thorium-227 were produced from a microgenerator, as previously described.¹³ Additional details are found in the Supplementary Data. The Bioscan AR-2000 Imaging Scanner (Eckert and Ziegler Radiopharma, Inc.) was used to measure the decay of ²²³Ra, ²²⁷Th, and a mixture of both isotopes. Purified ²²³Ra (0.025–0.1 μCi, 925–3700 Bq, 2–5 μL) was spotted in triplicate on 2.5 × 2.5 cm squares of Whatman paper and was not migrated (3 mm, Schleicher and Schuell).

A similar process was conducted for pure ²²⁷Th and a cospotted mixture of ²²³Ra and ²²⁷Th. Each paper strip was covered in cellophane wrap and read on the radio-TLC for 1 min at a high voltage setting (1500 V) and for 5 min at a low voltage setting (1000 V). Placing each square at a similar position on the scanner, these measurements were completed daily covering an entire half-life of ²²³Ra (11.4 days) or ²²⁷Th (18.7 days), with the initial reading occurring immediately after purification.

The counts per minute (CPM) data were exported from the Bioscan software (version 3.13), and the total area under the curve was calculated using Riemann sums for each time point. The average sum was plotted and fit to a one-phase decay curve using Graphpad Prism.

Quantitative readings of ²²³Ra and ²²⁷Th using γ counting

The Wizard² Gamma Counter (PerkinElmer) was used to measure the decay of ²²³Ra and ²²⁷Th over an entire half-life. In brief, purified ²²³Ra, ²²⁷Th, and a cospotted mixture (0.03–0.1 μCi, 1110–3700 Bq, 1–5 μL) were pipetted on 1 × 1 cm squares of 3 mm Whatman paper in triplicate. Each paper was then inserted into a 12 × 75 mm disposable culture tube and read on the γ counter for 60 s daily. The samples were counted over the full energy window from 100 to 2000 keV.

For quantification of activity amounts, the γ-well counter and HPGe system were calibrated using a National Institute of Standards and Technology (NIST) traceable standard of ²²³Ra. Defined volumes of the standard were aliquoted to form a dilution series, and each volume was weighed using a precision microbalance for accuracy (XP205; Mettler Toledo). The standards ranged from 0 to 9 kBq, and the full spectrum of γ-emissions was captured (2048 channels). CPM were plotted over activity, and linearity at each measured time point was r² = 0.992. The same undertaking was achieved using quantitative HPGe measurements using the 154 keV spectral line for ²²³Ra, producing a linear response of r² = 0.997.

A method was developed to separate ²²³Ra and ²²⁷Th activity amounts using the automated sample-changing γ-counter (Fig. 2). Spectral measurements of ²²³Ra (at secular equilibrium with its daughters) and freshly purified ²²⁷Th (<10 min) were acquired daily for 25 days, providing time for ingrowth of daughter and attendant emissions. Quantitation of the ²²⁷Th portion to absolute (Bq) activity amounts was made using the absolute NIST quantification of the ²²³Ra component and the known relative amounts of each radioisotope in the mixed spectra, calculated using the Bateman equations.¹⁴

FIG. 2.

(A) National Institute of Standards and Technology calibrated dilution series for ²²³Ra standards. (B) Quantitation of activity from traceable ²²³Ra source summed across channels. (C) Measurements of a ²²³Ra source at equilibrium over time, days indicated. (D) Measurements of ²²⁷Th over time. Ingrowth of ²²³Ra and daughters is noted. CPM, counts per minute.

Taken together, these measurements allowed two basis functions (for ²²⁷Th and ²²³Ra with daughters) to be defined, each in units of counts per spectral bin per Bq. Additional details are found in the Supplementary Data.

²²³Ra and ²²⁷Th readings using Minipix α spectral detection

The Minipix spectral imaging camera (300 μm silicon; ADVACAM) was used to characterize the α decay of ²²³Ra and ²²⁷Th. Approximately 0.025 μCi (925 Bq) of purified ²²³Ra or ²²⁷Th was spotted on 3 × 3 cm squares of 3 mm Whatman paper. In addition, ∼0.025 μCi (925 Bq) of purified ²²³Ra or ²²⁷Th was spotted on a 3 × 3 cm square of a solid aluminum support and allowed to air dry. The clustering tool on the PIXET PRO software (ADVACAM, version 1.4.10.704) was used to collect the global cluster volume of each detection event. For the measurement of ²²³Ra and ²²⁷Th α decay, samples were placed on top of the α particle imaging spectrometer, ∼0.7 cm from the detector, and measured daily over the period of one half-life for ∼15 min. The data were exported from PIXET PRO and analyzed as counts per energy bin (keV) using Graphpad Prism.

Radionuclide separation using N,N,N′,N′-tetra-n-octyldiglycolamide-coated chromatographic paper

²²⁷Th (0.012 μCi, 444 Bq) and ²²³Ra (0.007 μCi, 259 Bq) sources were cospotted on a 100 × 10 mm strip of chromatography paper impregnated with N,N,N′,N′-tetra-n-octyldiglycolamide (DGA; TrisKem International) and migrated in a mobile phase of 1 M HNO₃. The migrated strip was wrapped in cellophane and read using the radio-TLC under two high-voltage settings (1500 and 1000 V). Strips were measured at 5 min, 1 h, and 3 h postmigration, with an acquisition time of 5 min at 1500 V and 10 min at 1000 V. A duplicate DGA migration was separated into three equal sections and measured on the HPGe detector immediately after separation for radioisotopic identification.

²²⁷Th-radiolabeled protein characterization

A protein conjugated to a ²²⁷Th chelator of the hydroxypyridinone ligand (HOPO) class was purified by size exclusion spin centrifugation (Zeba column; ThermoFisher Scientific). Chelation status was analyzed by TLC before and after purification. Both purified and unpurified samples of the ²²⁷Th-labeled protein were spotted in triplicate on silica chromatographic paper and migrated in a mobile phase of 50 mM diethylenetriamine pentaacetate and assessed by TLC scanner at high-voltage settings of 1000 and 1500 V. Following this reading, the strips were cut in half, and both sets were measured using the α particle imaging spectrometer. Each half was then measured on the γ counter for 60 s.

Results

The overall goal was to develop quality control methods for characterization of targeted α-particle therapy constructs. For each source purified by the microgenerator column, radioisotopic purity was evaluated using the HPGe (Supplementary Fig. S1).¹³ Measurement by HPGe is a straightforward method for distinguishing ²²⁷Th from ²²³Ra, but many secondary and tertiary clinics, including academic medical centers, do not routinely use such equipment. The authors, therefore, tested the capacity of more readily available technologies to measure and report purity of an antibody-conjugated radiotherapeutic.

Radio-TLC scanner

The authors first evaluated a TLC scanner, the most widely deployed tool for radiolabeled agents, to measure the decay of ²²³Ra and ²²⁷Th using a novel α-particle specific setting. The differences in CPM acquisition for ²²⁷Th and ²²³Ra between the default high-voltage setting (1500 V) and a secondary setting (1000 V) were examined to report mixed ionizing radiation or α-particle specific emissions, respectively. A ²²⁷Th source measured at 1500 V showed an increase in CPM over time, gradually stagnating in slope near its transient equilibrium (Fig. 3A). This increase in CPM over time can be attributed to the ingrowth of its daughter isotopes, specifically ²²³Ra.

FIG. 3.

Readings at 1500 V (A–C) and at 1000 V (D–F) of each radioisotope and a mixed source spotted on cellulose support, measured over time. (A, D) ²²⁷Th CPM were plotted and fit according to a one-phase exponential decay. (B, E) ²²³Ra CPM were plotted and fit to a one-phase exponential decay. (C, F) Mixed source of ²²³Ra and ²²⁷Th. For all curves, the initial point is excluded from the exponential fitting to allow the system to reach equilibrium. Error bars indicate standard deviation between triplicates. CPM, counts per minute.

Measured under similar parameters, ²²³Ra showed a profile moderately fitted over a standard one-phase decay (r² = 0.5810; Fig. 3B). This poor fit is most likely due to the ingrowth of ²²³Ra daughters. The cospotted mixture of ²²⁷Th and ²²³Ra produced a fitted curve with minimal changes in CPM over the half-life of ²²⁷Th (Fig. 3C), again demonstrating the limitations of the system in measuring an isotopic mixture.

At lower bias voltage, β-particles will not produce detectable ionization events, whereas higher energy α-particles will produce electrons suitable for signal amplification, allowing for α-particle specific readings at 1000 V (Fig. 3D–F). The observed curve profiles were similar at 1000 and 1500 V for the mixture of ²²⁷Th/²²³Ra and for ²²⁷Th alone. However, the count plot for ²²³Ra matched a more accurate one-phase decay despite the presence of daughters (r² = 0.8611).

The authors were interested to see how these results would compare with measurements of a migrated TLC strip separating ²²³Ra from ²²⁷Th using the default and α particle specific settings. DGA-impregnated TLC paper was used to separate ²²⁷Th from ²²³Ra by ionic discrimination. Immediately after migration, three distinct peaks were recorded at the 1500 V setting (Fig. 4A), whereas two peaks were present when measured at 1000 V (Fig. 4D). Approximately 1 h later, the middle peak had decreased significantly (Fig. 4B), and by 3 h postmigration, this peak disappeared (Fig. 4C). The peaks at the front and base of the strip were not observed to decrease over this timeframe.

FIG. 4.

A strip of DGA-coated chromatographic iTLC paper was spotted with a mixture of ²²³Ra/²²⁷Th at 10 mm and migrated with a mobile phase of 1 M HNO₃ up to 100 mm. TLC readings over high voltage 1500 V at 20 min (A), 1 h (B), and 3 h postmigration (C). TLC readings of the same mixture at 1000 V at 5 min (D), 1 h (E), and 3 h postmigration (F). The middle peak seen at both high-voltage settings immediately after purification gradually decreases until it completely disappears 3 h postmigration. CPM, counts per minute; DGA, N,N,N′,N′-tetra-n-octyldiglycolamide; iTLC, instant thin layer chromatography; TLC, thin layer chromatography.

To verify the radioisotopic identity of each peak, a second strip was migrated, and each section was measured using the HPGe detector (Supplementary Fig. S1D–F). These measurements suggested that [²²⁷Th]Th⁴⁺ remained at the bottom of the strip, [²²³Ra]Ra²⁺ migrated to the top, and ²¹¹Bi (3+ charge) (half-life = 2.14 min) was present in the middle. Based on these results, the radio-TLC has a limited ability to quantitatively describe ²²⁷Th and ²²³Ra, although the 1000 V setting enhances resolution and increases the accuracy of measurements.

Sodium iodide γ counting

Another widely available instrument for radiopharmaceutical quality control is the γ-counter. Although highly sensitive to γ emissions, distinction of ²²⁷Th and ²²³Ra is challenging due to overlapping emissions (Supplementary Fig. S2A). The CPM acquired over a full window of energy spectra (100–2000 keV) for ²²⁷Th and ²²³Ra across all channels (open window) presents similar measurement profiles to those acquired under the radio-TLC scanner (Fig. 5A, B). The mixture of the two radioisotopes on the γ-counter shows a trend that is somewhat similar to that of ²²⁷Th (Fig. 5C). This is because a higher activity of ²²⁷Th than ²²³Ra was initially spotted on the paper.

FIG. 5.

Total CPM acquired in an open window channel measured for ²²⁷Th (A), ²²³Ra (B), and a mixture (C); Deconvoluted activity (Bq) calculated for a pure ²²⁷Th sample (D) showing ²²⁷Th decay fitting a single-phase decay (R² = 0.9980), as well as ²²³Ra ingrowth. (E) Activity (Bq) for ²²³Ra fit with a single-phase decay (R² = 0.9986). (F) ²²⁷Th and ²²³Ra mixed sample activity (Bq) plot. CPM, counts per minute.

To use this ubiquitous tool for radiopharmaceutical evaluation of ²²⁷Th and ²²³Ra, a method to deconvolute each energy spectrum was developed and the counts to report activities for each radioisotope were calibrated, enabling absolute quantification (Supplementary Fig. S2B). The method was tested by measuring an initially pure ²²⁷Th sample. In this study, a highly accurate one-phase decay curve (r² > 0.99) was observed while separately recording the ingrowth of ²²³Ra (Fig. 5D).

As expected, pure ²²³Ra reveals the predicted one-phase decay (r² > 0.99; Fig. 5E). The profiles of ²²⁷Th and ²²³Ra in the cospotted mixture showed the expected exponential decay for ²²⁷Th while recording ²²³Ra ingrowth (²²⁷Th: r² = 0.9294; Fig. 5F). Together, this evaluation demonstrates the utility of this method to reliably separate the ²²³Ra and ²²⁷Th γ spectra to quantitate activities accurately.

α Particle spectral detection

²²⁷Th and ²²³Ra emit α particles at characteristic energies, and isotope identification based upon these high-energy particles can be used for radiopharmaceutical evaluation. The authors were interested to use the Minipix spectral imaging camera for this purpose, a robust and mobile system for localized α-particle detection (Table 2). The α spectra of pure ²²⁷Th and ²²³Ra were evaluated and it was found that they were significantly different, allowing for differentiation between the two (Fig. 6F).

Table 2.

Specifications for Instrumentation Examined in These Methods

	Bioscan—1500 V	Bioscan—1000 V	γ Counter	Minipix
Detected radiation	All ionizing radiation	α particles	γ	All ionizing radiation
Detection mechanism	P-10 ionizing gas	P-10 ionizing gas	Well-type NaI detector	Si solid-state semiconductor (256 × 256 pixels)
Detection timea	1 min	5 min	1 min	15 min
Detection resolution	3 mmb	3 mmb	<30%b	Spatial: 9 lp/mm Energy: 0.8 keV (THL) and 2 keV (ToT)
Software	WinScan	WinScan	Wizard2 data analyzer	PixetPRO
Portable	No	No	No	Yes

^a For samples >0.025 μCi.

^b For most γ emitters.

THL, threshold level; ToT, time-over-threshold.

FIG. 6.

(A) A ²⁴¹Am/²³⁷Np/²⁴⁴Cm standard source read using Minipix spectral imaging camera. From left to right, the three peaks correspond to the α-particles emitted by ²³⁷Np, ²⁴¹Am, and ²⁴⁴Cm. (B, C) ²²³Ra and ²²⁷Th imaging spectra over one complete half-life. (D, E) Plotted counts acquired at 5.43 and 7.45 MeV for pure ²²⁷Th (D) and ²²³Ra (E). 5.43 MeV depicts the most significant α particle contribution emitted by ²²³Ra, and 7.45 MeV corresponds to the α particle released by ²¹¹Po and relates to the second peak seen in the graphs. Both increase over time for ²²⁷Th and decrease for ²²³Ra. (F) For the mixed spectra, the day 0 measurements of the radionuclides from (B) and (C) were normalized and plotted together. (G) The α spectra of ²²⁷Th and ²²³Ra plus daughters when spotted on an aluminum support.

Qualitatively, ²²³Ra shows two very spread peaks at ∼5500 and 7500 keV (Fig. 6B). The peak seen at 5500 keV corresponds to the α particles that are released by ²²³Ra (Table 2). The higher energy α-particles are most likely contributed by the daughters of ²²³Ra, specifically ²¹⁵Po (7386.1 keV) and ²¹⁰Po (7450.3 keV) and energies are collated in Table 2. Over 13 days, total counts in the ²²³Ra sample peaks decreased by approximately one half, corresponding to its half-life (Fig. 6E).

Unlike ²²³Ra, pure ²²⁷Th showed a single distinct peak at ∼5500 keV (Fig. 6C). Over time, the counts of this peak increase (Fig. 6D), due to ²²³Ra ingrowth, which emits α particles close in energy (Table 2). The emergence of a second peak at 7500 keV again corresponds to the high energy α particles emitted by the daughters of ²²³Ra.

The activity spotted on Whatman paper demonstrated low-energy resolution of individual α-particle spectra. This may be due to attenuation in air, by the required chromatography support and angular dispersion. Acquisitions were not collimated, as it produced no significant reduction in the energy distribution of individual peaks (not shown). When a pure ²²³Ra or ²²⁷Th source was spotted on a solid aluminum support, however, the energy resolution increased significantly (Fig. 6G). Here, comparative assessment of α-spectral detection permitted the differentiation of ²²⁷Th and ²²³Ra.

²²⁷Th-radiolabeled protein quality control

The previously presented analytical methods were applied to characterize a purified and nonpurified ²²⁷Th-radiolabeled protein. After migration on silica-coated chromatography paper, samples were first measured by radio-TLC. For the purified sample, only a single peak at the base of the strip was read, while the unpurified sample presented two peaks, suggesting the presence of free radioisotopes (Fig. 7A, B, E, F). These results demonstrate the value of the radio-TLC to evaluate the purity of a sample after it is migrated by TLC, but also its inability to identify specific radionuclides.

FIG. 7.

²²⁷Th-labeled antibody migrated on silica-coated chromatographic iTLC with DTPA (10 mM). Purified (A–D) and nonpurified (E–H) material were compared using radio-TLC scanner (A, B, E, F), γ counting (C, G), and α spectral imaging (D, H). The TLC scanner of purified and nonpurified ²²⁷Th-labeled protein was acquired immediately after migration at a high-voltage setting of 1000 V (A, E) and 1500 V (B, F), showing a main peak at the spotted area characteristic of ²²⁷Th-labeled material and migration of unchelated ²²³Ra at the top of the strip. (C, G) Purified (C) and unpurified (G) ²²⁷Th-labeled protein measured using γ counting. ²²⁷Th was found primarily nonmigrated at the bottom of the strip, confirming protein labeling. ²²³Ra was detected at the top of the strip for the unpurified material (G); (D, H) α spectral imaging of the purified (D) and nonpurified (H) ²²⁷Th-labeled antibody showed ²²⁷Th α emission at the bottom of the strip for both materials and ²²³Ra contribution at the top of the strip for the nonpurified material. CPM, counts per minute; DTPA, diethylenetriamine pentaacetate; iTLC, instant thin layer chromatography; TLC, thin layer chromatography.

The authors next measured the ²²⁷Th-radiolabeled protein using the γ counting protocol on a bisected sample. The purified protein showed only ²²⁷Th, which was present primarily at the bottom of the TLC strip; this contrasts to the unpurified material for which the majority of activity at the front was identified as ²²³Ra (Fig. 7C, G). The radiolabeling efficiency (92%) and radiochemical purity (>98%) for the purified compound were significant. The samples were also read on the Minipix spectral imaging camera. The purified sample depicts ²²⁷Th at the origin with nothing detected at the solvent front. Conversely, the unpurified sample reveals ²²³Ra at the front of the strip (Fig. 7D, H).

Discussion

The ability to accurately analyze a radiolabeled pharmaceutical is critical in both development and translation of new radiopharmaceuticals. With increasing interest in the field of radiopharmaceuticals for therapeutic use, the classical instrumentation utilized for radiotracer evaluation may be insufficient to assess the purity of agents that involve concatenated decay products with complex emission profiles.^3,13 The evaluation of several technologies was conducted to determine the capacity of widely available or affordable systems for these purposes. These results demonstrate the ability to quantitatively discriminate between ²²⁷Th and ²²³Ra utilizing a γ-counter, as well as the practical applications of these methods for the quality assessment of radiolabeled antibodies.

One common quality control approach for radiopharmaceutical evaluation is TLC migration read using a P-10 gas imaging scanner. The system is, by manufacturer default, operated at a setting of 1500 V. These results show that at 1500 V, the imaging scanner presents low sensitivity for measuring the decay of complex emitters. The measured decay of ²²³Ra presented an r² of only 0.5810 to an exponential fit. This is likely explained by the presence of its daughters, which emit X-rays, γ-rays, electrons, and α-particles for which the system has variable sensitivities.

The same observation is noted for pure ²²⁷Th, which did not follow a one-phase decay. This increase in signal before equilibrium can be explained by the ingrowth of ²²³Ra, which presents a comparable half-life (11.4 days) and strong emitted γ rays (324 keV, 3.64%; 269 keV, 13.30%; 154 keV, 6.02%; 144 keV, 3.47%) and has increased detection through ionization of the P-10 gas source. When lowering the high-voltage setting to 1000 V, the authors found that the accuracy of measurement increased substantially for ²²³Ra (r² = 0.8611).

A different migration profile was recorded for a mixture of ²²³Ra and ²²⁷Th when the scanner was used at a setting of 1500 V as compared to at 1000 V. At the lower voltage setting, the spatial resolution of peaks was enhanced, at the cost of longer acquisition duration. As both time and resolution are important considerations, the application of the radio-TLC will determine which high-voltage setting is appropriate. These observations suggest the limitations of the imaging scanner to quantitatively describe complex emitting isotopes, but the utility is apparent in qualitatively determining impurities in a migrated sample.

The total counts obtained from γ-counting readings under an open energy window setting (100–2000 keV) showed similar results to those of the radio-TLC. However, ²²³Ra values demonstrated an accurate one-phase decay (r² = 0.9993). This increase in accuracy when read on the γ-counter is due to the selective measurement of γ-rays. The γ-energy spectra of ²²⁷Th and ²²³Ra share significant regions of overlap and are, therefore, challenging to distinguish. Measurements of each pure radioisotope presented classical one-phase decay with time, with an r² > 0.99 for both ²²⁷Th and ²²³Ra sources over the course of 20 days. Both the mixture and pure ²²⁷Th capture the ingrowth of ²²³Ra.

The absolute activity values first begin to plateau ∼15 days postpurification, consistent with the transient equilibrium (19 days). The deconvolution method employed here enhances the quantification capabilities of the γ-counter, enabling the distinction of ²²⁷Th from ²²³Ra and can be applied in research or clinical settings for purity assessment as well as an exact activity amount of each isotope in a mixture. This method may be useful for clinical quality control applications, which currently employ high-resolution γ-spectroscopy.¹⁵

As another analytical method for measuring the emitted α-particles from ²²⁷Th and ²²³Ra, detection using the Minipix solid-state semiconductor was investigated. The results show that this method was also able to differentiate between ²²⁷Th and ²²³Ra. Pure ²²⁷Th can be isolated by a single peak close to 6000 keV, which is consistent with the reported α particles (Table 2). Over time, the main peak observed shifts slightly downward in energy and increases in counts as ²²³Ra grows into the mixture, and a second peak closer to 8000 keV appears.

This second high-energy peak becomes apparent in as little as a day postpurification, demonstrating the utility of this method to differentiate between pure and impure ²²⁷Th. Pure ²²³Ra measured over time shows two peaks, with the lower peak corresponding to ²²³Ra and the upper peak to its higher energy α-emitting daughters. Both peaks decrease in overall counts over the course of one half-life to approximately half that of the day 0 measurement, consistent with the natural decay of the radioisotope. As demonstrated here, it is possible to qualitatively measure and distinguish ²²⁷Th and ²²³Ra by α-spectrometry.

The authors did observe significant spreading of individual peaks when measuring both ²²⁷Th and ²²³Ra. The high number of emitted α-particles (Table 2) and the distance that the α-particles must travel through the air before reaching the detector (0.7 cm) contribute to this observation. Using an aluminum support in place of the filter paper improved the energy resolution of ²²³Ra and ²²⁷Th samples. This implies the dispersion of α-particle energies was due to absorption within the paper. From these results, the authors note that a solid support may provide improved α-particle spectrometry-based assessment of radiopharmaceuticals.

Finally, the practicality of these methods using a ²²⁷Th-labeled protein drug was shown, and the important clinical applications of this study for other radiopharmaceutical products incorporating complex emitters were demonstrated, such as ²²³Ra and ²²⁵Ac, among others. The authors showed the ability to distinguish between the purified and nonpurified construct utilizing all three previously described methods. The radio-TLC at both high-voltage settings clearly demonstrates the presence of two chemically different species in the unpurified sample, but the authors were unable to identify individual radioisotopes by TLC evaluation.

To further characterize each sample, the upper and lower sections of the TLC strip were measured using the γ-counter and the counts per channel of energy were processed using the deconvolution method. The activity of ²²⁷Th and ²²³Ra in each sample was quantified, and the authors determined that the majority of the impurity that migrated to the front of the strip in the unpurified sample was free ²²³Ra. This is a validation of the γ-counting method to quickly and accurately determine the purity of a prepared radiopharmaceutical. The final results were confirmed by α-spectral analysis and HPGe measurement.

A final consideration is the improvements that these methods can make to current quality control practices. Currently, measurement by HPGe is the gold standard for discrimination between ²²⁷Th and ²²³Ra. However, this equipment is expensive and not accessible at all research and care facilities, and these methods propose three instruments that are widely available. With the exception of the γ-counter, the instruments are unable to provide exact quantification of activity amounts but provide a measurement of qualitative purity assessment. The γ-counter can provide quantitative activity amounts for the radioisotope of interest and the impurity, and is also broadly available, leading to the possibility for more accurate and accessible quality control of complex-emitting radiopharmaceuticals.

Conclusion

The development and clinical evaluation of ²²⁷Th and ²²³Ra therapies place new demands on radiochemists and pharmacists to ensure safe and effective use. As the field evolves to regulate and handle these potent agents, the authors have investigated widely deployed, robust, and affordable methods for their quality control.¹⁶ The data demonstrate an accessible method for assessing ²²⁷Th and ²²³Ra radiopharmaceutical purity using a γ-counter after discrimination of each isotope.

Radio-TLC and α-particle spectrometer imaging may prove useful for evaluating constituents and purity of a sample, but they may not be ideally suited for evaluation of complex mixtures of isotopes. Overall, this evaluation of two emerging radionuclides for α-therapy treatment underlines technological developments toward improved quality assurance for radiopharmaceuticals with concatenated decays and complex emission profiles.

Authors' Contributions

A.H. contributed to writing (lead), conceptualization (lead), formal analysis (lead), methodology development (equal), review, and editing (equal). W.J. was involved in writing (support), formal analysis (support), methodology development (support), review, and editing (equal). N.B. and L.S. carried out methodology development (support). P.L. carried out methodology development (equal). M.S.L. was involved in conceptualization (support), methodology development (equal), review, and editing (equal). B.J.B. carried out methodology development (equal), review, and editing (equal). H.Z. was in charge of conceptualization (support) and methodology development (support). R.L.W. carried out conceptualization (support), review, and editing (support). D.S.A. took care of writing (equal), conceptualization (lead), formal analysis (equal), methodology development (equal), review, and editing (equal). D.L.J.T. was in charge of writing (equal), conceptualization (equal), formal analysis (equal), methodology development (equal), review, and editing (equal).

Disclosure Statement

There are no existing financial conflicts.

Funding Information

The study herein was funded in part by the National Cancer Institute at the National United States Institutes of Health R01CA229893 (D.L.J.T.), R01CA201035 (D.L.J.T.) and R01CA240711 (D.L.J.T.). Isotopes used in this research were supplied in part by the United States Department of Energy Office of Science, Isotope Program, Office of Nuclear Physics.

Supplementary Material

Supplementary Data

Supplementary Figure S1

Supplementary Figure S2