A Phase 0 Imaging Trial of [²⁰³Pb]Pb-VMT-α-NET to Enable Dosimetry and Treatment Planning for Refractory or Relapsed Metastatic Neuroendocrine Tumors with [²¹²Pb]Pb-VMT-α-NET

Visual Abstract

Abstract

Peptide receptor radionuclide therapy (PRRT) using α-particle emitters has potential to provide improved patient outcomes over those achieved with β-particle PRRT. A promising candidate radiopharmaceutical pair, PSC-PEG₂-TOC (VMT-α-NET) conjugated to ²⁰³Pb for SPECT/CT imaging or ²¹²Pb for α-PRRT, is currently in early-phase clinical trials for patients with neuroendocrine tumors (NETs). Here, we present the imaging and dosimetry characteristics of this theranostic approach. Methods: A phase 0 imaging trial of [²⁰³Pb]Pb-VMT-α-NET was conducted between January and December of 2023. Ten participants with somatostatin receptor type 2 (SSTR2)–positive NETs underwent SPECT/CT and blood sampling at 1, 4, 24, and 48 h after intravenous infusion of approximately 185 MBq of [²⁰³Pb]Pb-VMT-α-NET. The diagnostic performance of [²⁰³Pb]Pb-VMT-α-NET was evaluated through lesion-by-lesion comparison against baseline SSTR2 PET/CT. A subset of lesions was further analyzed for signal-to-noise ratio to determine the optimal diagnostic imaging time point after [²⁰³Pb]Pb-VMT-α-NET administration. Patient-specific dosimetry of [²¹²Pb]Pb-VMT-α-NET was derived from ²⁰³Pb imaging and performed assuming local α- and β-particle energy deposition in tumors and normal organs. Effects of daughter ion relocation were considered using a whole-body pharmacokinetic model on the basis of parameters published by the International Commission on Radiological Protection. Results: Of the 162 total lesions identified on SSTR2 PET/CT scans, only 97 were detected on [²⁰³Pb]Pb-VMT-α-NET SPECT/CT. The highest signal-to-noise ratio for lesions occurred 4 h after [²⁰³Pb]Pb-VMT-α-NET administration. Sensitivity was 94% for lesions larger than 1 cm versus 35% for lesions no larger than 1 cm or nonmeasurable lesions. The effective dose associated with [²⁰³Pb]Pb-VMT-α-NET administration was 0.038 mSv/MBq (1.40 mSv/mCi). Estimated dosimetry for [²¹²Pb]Pb-VMT-α-NET (mean ± SD, not adjusted for relative biologic effectiveness) based on [²⁰³Pb]Pb-VMT-α-NET SPECT/CT was 15 ± 4.7 mGy/MBq for the kidneys, 8.4 ± 4.1 mGy/MBq for the spleen, 2.5 ± 0.8 mGy/MBq for the liver, 0.324 ± 0.108 mGy/MBq for the blood, 0.270 ± 0.081 mGy/MBq for the whole body, and 29.6 ± 25.8 mGy/MBq for tumors. Renal absorbed dose projections for ²¹²Pb were estimated to carry an overall standard uncertainty (k = 1) of 15.3%. Conclusion: [²⁰³Pb]Pb-VMT-α-NET appears to be a safe and effective SPECT/CT tracer for imaging NETs larger than 1 cm and for normal organ and tumor radiation dosimetry. The chemically matched ²⁰³/²¹²Pb theranostic pair offers the potential for a dosimetry-driven personalized treatment paradigm.

Neuroendocrine tumors (NETs) are classified as an orphan disease in the United States, yet the incidence of NETs has increased 6-fold over the past 4 decades (1). Although the disease is characteristically slow to progress in the early stages, the 5-y survival rate plummets to less than 30% once metastasis occurs (2). With few exceptions (e.g., neuroblastoma), NETs respond poorly to chemotherapy, with surgery providing the only potentially curative therapy. Approved therapies may relieve symptoms, but objective responses are rare. It is widely recognized that newer, more effective forms of therapy for these patients are urgently needed.

Radiopharmaceutical therapies for NETs, such as [¹⁷⁷Lu]Lu-DOTATATE (Lutathera; Novartis AG) and ¹³¹I-MIBG (Azedra; Lantheus Medical), have been approved by the Food and Drug administration (FDA) within the past 10 y (3–7). Studies with these agents have demonstrated improvement in progression-free survival when compared with standard therapy with octreotide; however, the objective response rate remains low (7,8). Although important survival benefits have been reported with [¹⁷⁷Lu]Lu-DOTATATE and ¹³¹I-MIBG, the long-term outlook for patients with NETs remains poor (9,10).

Emerging evidence suggests that α-particle peptide receptor radionuclide therapy (PRRT) has the potential to improve both response rates and survival (11,12). In particular, patients with metastatic NETs whose disease has progressed or relapsed after treatment with the FDA-approved β⁻-emitting therapeutic, [¹⁷⁷Lu]Lu-DOTATATE, stand to benefit from alternative treatment options that may delay progression and prolong life. α-emitting radiopharmaceutical therapy shows immense promise for meeting this need and for potentially improving response rates among patients who are radiopharmaceutical treatment–naïve. Recent studies have explored the use of [²²⁵Ac]Ac-DOTATOC, [²²⁵Ac]Ac-DOTATATE, and [²¹²Pb]Pb-DOTAMTATE, with some positive patient outcomes and preliminary dosimetry data being reported (13–17).

Previously, we reported on the preclinical development of Pb-PSC-PEG₂-TOC (Pb-VMT-α-NET), a promising candidate radiopharmaceutical for α-PRRT of NETs (18), which can be radiolabeled either by the therapeutic α-emitting radionuclide ²¹²Pb (half-life, 10.64 h) or by the imageable photon-emitting ²⁰³Pb (half-life, 51.92 h; $E_{\mathrm{\gamma }},$ 278 keV; 81% intensity). Considering the preclinical success of this radiopharmaceutical agent and the attractive nature of chemically matched radiopharmaceuticals for imaging and therapy, we initiated a phase 0 clinical trial to evaluate the diagnostic and potential therapeutic efficacy of [^203/212Pb]PSC-PEG₂-TOC. In this work, we present results from the first 10 participants imaged, including 3 who were subsequently treated using a personalized dosimetry paradigm. To our knowledge, this work constitutes the first prospective dosimetry-guided α-particle therapy in humans.

MATERIALS AND METHODS

Study Design and Patient Selection

A research investigational new drug application for [²⁰³Pb]Pb-VMT-α-NET (Fig. 1) was approved by the FDA. This trial was approved by University of Iowa’s biomedical institutional review board and was registered on ClinicalTrials.gov (NCT05111509) before enrollment of the first patient. Written informed consent was provided by participants; consent was confirmed by a monitor assigned by the data and safety monitoring committee. Protected health information was sequestered in compliance with the Health Insurance Portability and Accountability Act. Individual participant data will be shared, for those participants who opted-in to sharing, after a data usage authorization is fully executed.

FIGURE 1.

Chemical structure of [Pb]PSC-PEG₂-TOC (VMT-α-NET).

This phase 0 imaging trial of [²⁰³Pb]Pb-VMT-α-NET was conducted at the University of Iowa between January and December 2023. Ten patients were enrolled and met the following eligibility criteria: established NETs (grade 1 or 2, primary location known or believed to be foregut or midgut), at least 1 positive primary metastatic lesion on somatostatin receptor type 2 (SSTR2) PET imaging within 12 mo of providing consent with a measurable size of at least 2 cm in any dimension on CT or MRI, age of 18 y or older, and adequate performance status (Eastern Cooperative Oncology Group status of 0 or 1; Karnofsky Performance Status of ≥70). Patients were excluded from consideration if they were pregnant or lactating, declined to use effective contraception during the study, or had received a therapeutic investigational drug within 4 wk. Patients for whom discontinuation of somatostatin analog therapy represented a health risk (in the opinion of their treating physician) were also excluded from the study. Further criteria can be found at https://www.clinicaltrials.gov/study/NCT05111509.

Radiopharmaceutical Preparation and Administration

[²⁰³Pb]Pb-VMT-α-NET and [²¹²Pb]Pb-VMT-α-NET were provided by Perspective Therapeutics for purposes of this study. Production of [^203/212Pb]Pb-VMT-α-NET was achieved by radiolabeling the peptide-based ligand VMT-α-NET with ²⁰³Pb (Medical Isotope and Cyclotron Facility, University of Alberta) or ²¹²Pb from a ²²⁴Ra/²¹²Pb generator (Perspective Therapeutics). The radiopharmaceutical was prepared in accordance with Good Manufacturing Practice standards, using an automated synthesis module and methods derived from those introduced by Li et al. (19). The product was authorized for clinical use upon conformance with quality control specifications.

The radiopharmaceutical was received by the University of Iowa in a 1-mL vial. ²⁰³Pb activity was measured using Capintec dose calibrators, with dial settings of 318 for the CRC-55tR and 324 for the CRC-15R, in accordance with National Institute of Standards and Technology–traceable, high-purity germanium γ-spectrometric measurements. The appropriate quantity of radiopharmaceutical received was dispensed into a 30-mL vial, and the product volume was increased to 25 mL by the addition of sterile saline. The resulting radiopharmaceutical was administered to each participant using standard protocols developed and maintained at the University of Iowa. Briefly, peripheral intravenous access was obtained and normal saline initiated (250 mL at 30 mL/h). Next, ondansetron (8 mg) was administered intravenously (for 3 patients receiving an amino acid coinfusion), with an amino acid infusion (25 g of L-lysine and 25 g of L-arginine in 1,000 mL) initiated 30 min later at a rate of 250 mL/h. Thirty minutes after the amino acid infusion was initiated, the radiopharmaceutical was administered via infusion pump, followed by a saline flush. Participants were monitored for any adverse events, including extravasation, until imaging was completed (48 h after injection).

Imaging and Pharmacokinetic Sampling

An Intevo 2 SPECT/CT system (Siemens Healthineers) was commissioned for use in this study through determination of the system-specific sensitivity and mean concentration recovery coefficients for [²⁰³Pb]Pb-VMT-α-NET. To determine system sensitivity, a low-density polyethylene bottle was filled with water and a known quantity of ²⁰³Pb and placed within a water-equivalent plastic cylinder phantom (20 cm in diameter and length; CIRS), as previously described (20). For recovery coefficient determination, an International Electrotechnical Commission National Electrical Manufacturers Association body phantom was modified to accommodate 48- and 60-mm fillable spheres by removing the 13- and 22-mm spheres and the lung insert. The spheres were filled with ^99mTc, with no background activity in the phantom. Both phantoms were scanned and reconstructed in accordance with the imaging protocol, with adjustment of the energy window for ^99mTc. Image acquisition was performed using the following settings: 120 views (60 views per head), 20 s per view, ×1.0 zoom, 128 × 128 matrix, noncircular continuous rotation orbit, medium-energy low-penetration collimator, 20% window centered on the 279 keV γ-peak with 10% upper and lower scatter windows, and 2 SPECT fields of view covering the chest, abdomen, and pelvis. Images were reconstructed in 3 dimensions using an ordered-subset expectation maximization (8 iterations, 12 subsets), with corrections for attenuation, scatter, and collimator detector response effects (Siemens Flash3D), and a postreconstruction 5-mm gaussian filter. This technique, using ^99mTc with a medium-energy collimator as a “resolution surrogate” for ²⁰³Pb, has been preliminarily shown to be appropriate for well-resolved photopeak imaging when there is minimal septal penetration (i.e., similar collimator detector response characteristics) (20).

SPECT/CT imaging was performed at 1–2, 4, 24, and 48 h (except for participant 10, for whom the 48-h time point was omitted) after intravenous infusion of approximately 185 MBq of [²⁰³Pb]Pb-VMT-α-NET. At each time point, whole-body conjugate imaging was also performed (12 cm/min), and blood was collected for activity measurements using a calibrated auto γ-counter (Wizard 2470; Perkin Elmer). For all imaging, a ²⁰³Pb imaging standard (50–100 μCi in ∼20 mL of saline in a 30-mL glass vial) was placed in the imaging field of view to enable verification of SPECT quantitation accuracy. For the 3 participants who received amino acids, the amino acid infusion was continued normally during the first imaging time point.

[²⁰³Pb]Pb-VMT-α-NET Diagnostic Performance Analysis

To assess the diagnostic performance of [²⁰³Pb]Pb-VMT-α-NET, a lesion-by-lesion analysis was conducted comparing tumor uptake on [²⁰³Pb]Pb-VMT-α-NET SPECT/CT with that of an FDA-approved SSTR2 PET agent. Images from [²⁰³Pb]Pb-VMT-α-NET were compared side by side with corresponding SSTR2 PET/CT images. Up to 5 representative lesions positive on SSTR2 PET larger than 1 cm were identified for each region for comparison with [²⁰³Pb]Pb-VMT-α-NET SPECT/CT. Measurements were obtained in any dimension from diagnostic CT or MRI; attenuation CT of PET/CT was used if a contemporaneous diagnostic CT or MRI of the area of interest was unavailable. In addition, up to 5 positive lesions on SSTR2 PET per region were identified, measuring no more than 1 cm or not visible or measurable on cross-sectional imaging. The body was divided into 9 regions: brain, lungs and pleura, liver, pancreas, bowel and peritoneum, lymph nodes of the neck and chest, lymph nodes of the abdomen and pelvis, axial skeleton of the spine and pelvis, and remaining skeleton. Tumor uptake was considered positive if the uptake was equal to or higher than liver uptake. Concordance between [²⁰³Pb]Pb-VMT-α-NET and SSTR2 PET was critical for further development of the theranostic pair. The 4-h images were used for the final sensitivity analysis, taking into consideration the results of the lesion detectability analysis performed in a subgroup of lesions.

A subset of lesions was further analyzed to determine the optimal diagnostic imaging time point after [²⁰³Pb]Pb-VMT-α-NET administration. For each lesion identified, an anatomic contour was created using the CT scan, followed by morphologic expansion of 10 mm to create a functional contour (thereby encompassing partial-volume effects). An additional hollow-shell “ring” contour was created beyond the functional contour to determine local background concentration and estimate noise quality. This ring contour was manually modified to minimize any overlap with nearby physiologic uptake. A 5-cm sphere was placed in normal liver to evaluate liver uptake. Tumor uptake, tumor-to-liver ratio, estimated signal-to-noise ratio (eSNR), and estimated contrast-to-noise ratio (eCNR) were calculated to assess the optimal imaging time point as follows:

$$\mathrm{eSNR}=\frac{(\mathrm{total}\mathrm{ }\mathrm{number}\mathrm{ }\mathrm{of}\mathrm{ }\mathrm{counts}\mathrm{ }\mathrm{in}\mathrm{ }\mathrm{expanded}\mathrm{ }\mathrm{anatomic}\mathrm{ }\mathrm{contour})}{(\mathrm{SD}\mathrm{ }\mathrm{of}\mathrm{ }\mathrm{counts}\mathrm{ }\mathrm{per}\mathrm{ }\mathrm{voxel}\mathrm{ }\mathrm{in}\mathrm{ }\mathrm{background}\mathrm{ }\mathrm{region})}$$ Eq. 1

and

$$\mathrm{eCNR}=\frac{\left[\left(\mathrm{counts}\mathrm{ }\mathrm{per}\mathrm{ }\mathrm{voxel}\mathrm{ }\mathrm{in}\mathrm{ }\mathrm{anatomic}\mathrm{ }\mathrm{contour}\right)-\left(\mathrm{counts}\mathrm{ }\mathrm{per}\mathrm{ }\mathrm{voxel}\mathrm{ }\mathrm{in}\mathrm{ }\mathrm{background}\right)\right]}{(\mathrm{SD}\mathrm{ }\mathrm{of}\mathrm{ }\mathrm{counts}\mathrm{ }\mathrm{per}\mathrm{ }\mathrm{voxel}\mathrm{ }\mathrm{in}\mathrm{ }\mathrm{background}\mathrm{ }\mathrm{region})(\mathrm{counts}\mathrm{ }\mathrm{per}\mathrm{ }\mathrm{voxel}\mathrm{ }\mathrm{in}\mathrm{ }\mathrm{background})}.$$ Eq. 2

Dosimetry Calculations

Tissue segmentation for dosimetry was performed manually on the 4-h SPECT/CT in MIM Protégé (MIM Software) for organs of interest (kidneys, liver, bladder, spleen, bladder) and lesions that could be identified on both SPECT and anatomic images (fused CT acquired alongside SPECT or recent diagnostic PET/CT, CT, or MRI). Anatomic contours for lesions, kidneys, and spleen were expanded by 10 mm to account for partial-volume effects, followed by manual region revision to minimize overlap with adjacent physiologic uptake. These expanded regions of interest (ROIs) were rigidly translated in the x-, y-, and z-directions to correct for any misregistration between the CT and SPECT images. For liver lesions with low contrast (less than ∼5 times the surrounding liver), phantom-derived recovery coefficients were used for partial-volume correction rather than the expanded ROI technique. ROIs were propagated to other imaging time points (1, 24, and 48 h) by SPECT–SPECT rigid registration, followed by manual translation of the expanded ROIs to ensure accurate alignment. Tissue masses, in grams, were obtained from the anatomic contours using the CT scan, using $\frac{\mathrm{H}{\mathrm{U}}_{\mathrm{mean}}+ 1000}{1000}\ast V,$ where HU is Hounsfield units and V is volume in milliliters (21). A time–activity curve for each tissue was derived from the expanded. ROIs (or partial-volume–corrected anatomic regions) from the sequential quantitative SPECT images.

Time–activity curves for each tissue structure were fit by least-squares regression using various analytic models to determine the appropriate functional form for the observed pharmacokinetics. It was determined that a biexponential fit of the form $A\left(t\right)=A_1{e}^{-k_{1}t}(1-{A_{2}e}^{-k_{2}t})$ was most appropriate for tumors. For kidneys, a monoexponential fit described the data well. A biexponential fit of $A\left(t\right)=A_1{e}^{-k_{1}t}+A_2{e}^{-k_{2}t}$ was most appropriate for the bladder and spleen.

Time-integrated activity coefficients (TIACs) for [²⁰³Pb]Pb-VMT-α-NET were obtained directly from the time–activity regression fits by analytic integration to $t=\mathrm{\infty }$, whereas TIACs for the ²¹²Pb therapy were obtained by scaling the activity measurements by the ratio of decay constants ($e^{-{\mathrm{\lambda }}_{\mathrm{Pb}212}t}/e^{-{\mathrm{\lambda }}_{\mathrm{Pb}203}t}$), followed by least-squares regression and TIAC determination by analytic integration to $t=\mathrm{\infty }$. TIACs for blood (dosimetric surrogate for marrow) were obtained by performing a biexponential fit to the sample-derived time–activity curve. Total blood volume was assumed to be 9% of the patient’s total body mass.

[²⁰³PbVMT-α-NET TIACs were used in conjunction with OLINDA/EXM version 2.2 to calculate the effective dose (22) and effective dose equivalent (23) of the imaging procedure for each subject. The average of the effective dose values obtained for all 10 participants was used to assign an effective dose estimate for the [²⁰³Pb]Pb-VMT-α-NET imaging radiopharmaceutical.

A whole-body pharmacokinetic model was used to determine the dosimetric effects of daughter ²¹²Bi ions released from the Pb-specific chelator (PSC) after the initial β-decay of ²¹²Pb. Prior reports with similar macrocycle chelators have indicated that roughly 36% of daughter ions are released after the initial decay of ²¹²Pb (24). However, the VMT-α-NET structure includes a chelator that was specifically designed for ²¹²Pb and has been reported to reduce ²¹²Bi decoupling to less than 5% in formulation (24 h); therefore, at the time of injection, the percentage of free ²¹²Bi is expected to be dosimetrically negligible (25). In addition, the contribution of the high-energy γ-emitting daughter radioprogeny ²⁰⁸Tl to whole-body dosimetry has been shown to be negligible (26). Methods for pharmacokinetic modeling and dosimetry corrections have been described previously (27). Briefly, the pharmacokinetic model for radiobismuth that appeared in International Commission on Radiological Protection publication 137 (28) was reproduced in MATLAB (MathWorks), using the compartment transfer coefficients provided in Table 10.3 of the publication. The model was run one time for each source compartment, simulating the release of 100,000,000 bismuth ions, tallying the subsequent compartment of decay for each atom. These results were used to construct a daughter-translocation matrix of source and destination compartments, with $\mathrm{\psi }\left(r_{\mathrm{i}}\leftarrow r_{\mathrm{s}}\right)$ denoting the fraction of ion decays (ψ) occurring in the destination compartment ($r_{\mathrm{i}}$) with initial release in the source compartment ($r_{\mathrm{s}}$). Daughter translocation was incorporated into the ²¹²Pb dosimetry by generating corrected TIACs $\tilde{A}\left(r_{\mathrm{i}}\right)$ from the imaging-derived TIACs ${\tilde{A}}_{\mathrm{raw}}\left(r_{\mathrm{i}}\right)$ using the following equation:

$$\tilde{A}\left(r_{\mathrm{i}}\right)=\left(1-f\right){\tilde{A}}_{\mathrm{raw}}\left(r_{\mathrm{i}}\right)+f\sum _{r_{\mathrm{s}}}{\tilde{A}}_{\mathrm{raw}}\left(r_{\mathrm{s}}\right)\mathrm{\psi }(r_{\mathrm{i}}\leftarrow r_{\mathrm{s}}),$$ Eq. 3

where f is the assumed release fraction of daughter bismuth ions, herein taken to be 0.36 (24).

Using the corrected ²¹²Pb TIACs, the absorbed dose to each tissue structure was calculated assuming local energy deposition with a total ²¹²Pb decay chain energy of 8.6015 MeV (7.8063 MeV of ${\mathrm{\alpha }}^{2+}$ and 0.7952 MeV of ${\mathrm{\beta }}^{-}$). Photon contributions were not included. No weighting for relative biologic effectiveness was performed.

Efforts were made to estimate the uncertainty associated with ²⁰³Pb-derived predictions of [²¹²Pb]Pb-VMT-α-NET dosimetry. Factors considered included organ segmentation and activity quantitation reproducibility, CT organ distortion caused by respiratory motion, CT HU–to–density calibration, time–activity curve fitting and integration, uncertainty in published nuclear data (half-life and particle emission energies), SPECT sensitivity calibration and dead-time effects, radionuclide calibrator dial settings for ²⁰³Pb and ²¹²Pb, potential bias associated with assuming local energy deposition for β-particles, bias from not including photon cross-irradiation, uncertainty in the ²¹²Bi daughter release fraction, and the time interval between imaging and radiopharmaceutical administrations. An accounting of these factors is provided in the supplemental materials, available at http://jnm.snmjournals.org (29–32).

RESULTS

Diagnostic Performance of [²⁰³Pb]Pb-VMT-α-NET SPECT/CT and Optimal Imaging Time Point

Ten patients underwent [²⁰³Pb]Pb-VMT-α-NET SPECT/CT, SSTR2 PET/CT, and morphologic imaging with CT or MRI. SPECT/CT imaging was completed after an average of 113 ± 72 d after SSTR2 PET/CT imaging and 53 ± 38 d after morphologic imaging. For SSTR2 PET/CT imaging, 7 patients were imaged with [⁶⁸Ga]Ga-DOTATOC, 2 patients with [⁶⁴Cu]Cu-DOTATATE, and 1 patient with [⁶⁸Ga]Ga-DOTATATE.

The diagnostic sensitivity of ²⁰³Pb SPECT/CT versus SSTR2 PET/CT is summarized in Table 1. There was at least 1 positive lesion in each of the [²⁰³Pb]Pb-VMT-α-NET SPECT/CT scans. Of the 162 lesions identified on SSTR2 PET/CT scans, 97 were positive on [²⁰³Pb]Pb-VMT-α-NET SPECT/CT, resulting in a sensitivity of 60% for [²⁰³Pb]Pb-VMT-α-NET. In the 3 most common sites of involvement, [²⁰³Pb]Pb-VMT-α-NET visualized 75% of liver lesions (40/53), 56% of lymph node lesions (32/57), and 33% (9/27) of bone lesions seen on SSTR2 PET. Of the 54 SSTR2 PET–positive lesions no larger than 1 cm, 15 were identified with [²⁰³Pb]Pb-VMT-α-NET SPECT/CT. Of the 40 nonmeasurable or nonvisible lesions on CT/MRI, which were predominantly in the bone and gastrointestinal tract/peritoneum, 18 were identified on [²⁰³Pb]Pb-VMT-α-NET SPECT/CT. Overall, 33 of 94 lesions (35%) no larger than 1 cm or nonmeasurable lesions were visualized on [²⁰³Pb]Pb-VMT-α-NET SPECT/CT scans. The sensitivity of [²⁰³Pb]Pb-VMT-α-NET was significantly higher for lesions larger than 1 cm, with 94% (64/68) visualized. Four target lesions larger than 1 cm that were localized in the liver, pancreas, and lymph nodes were not identified on [²⁰³Pb]Pb-VMT-α-NET SPECT/CT. Figure 2 shows [²⁰³Pb]Pb-VMT-α-NET SPECT, [⁶⁸Ga]Ga-DOTATOC PET, and diagnostic CT images of a subject with a metastatic ileal NET, demonstrating tumor uptake of [²⁰³Pb]Pb-VMT-α-NET but limited visualization of small tumors (≤1 cm).

TABLE 1.

Diagnostic Sensitivity of ²⁰³Pb SPECT/CT vs. SSTR2 PET/CT

	No. lesions
Lesion size	203Pb SPECT/CT	SSTR2 PET/CT	Sensitivity (%)
n	97	162	60
>1 cm	64	68	94
≤1 cm in longest diameter or nonmeasurable	33	94	35

FIGURE 2.

Participant with grade 2 metastatic ileal NET. Columns left to right show [²⁰³Pb]Pb-VMT-α-NET maximum-intensity projection whole-body image (A), [⁶⁸Ga]Ga-DOTATOC maximum-intensity projection whole-body image (B), [²⁰³Pb]Pb-VMT-α-NET transaxial SPECT images through liver and lower abdomen (C), [⁶⁸Ga]Ga-DOTATOC transaxial PET images through same level of liver and lower abdomen (D), and contrast-enhanced CT images through same level of liver and lower abdomen (E). [²⁰³Pb]Pb-VMT-α-NET images show clear visualization of liver metastases measuring up to 2.6 cm (top arrow, E). Although some mesenteric deposits are visualized on [²⁰³Pb]Pb-VMT-α-NET SPECT, many subcentimeter [⁶⁸Ga]Ga-DOTATOC–positive mesenteric lesions (bottom arrows, E) are not visualized on [²⁰³Pb]Pb-VMT-α-NET SPECT.

To determine the best imaging time point to visualize lesions with [²⁰³Pb]Pb-VMT-α-NET, 48 lesions from 9 patients were analyzed (Fig. 3). Tumor uptake (total imaging counts in functional contour) was rapid, reaching a maximum at approximately 4 h after injection and decreasing thereafter (98% of maximum at 1 h; 31% of maximum at 24 h; 16% of maximum at 48 h). The tumor-to-liver concentration ratio increased steadily over time, with values of 15.5, 17.5, 32.9, and 45.9 at 1, 4, 24, and 48 h, respectively. Lesion detectability, as measured by eSNR and eCNR, reached a maximum at 4 h, with mean eSNRs of 224, 236, 142, and 80 at 1, 4, 24, and 48 h, respectively. The corresponding eCNRs for these time points were 151, 168, 107, and 54, respectively.

FIGURE 3.

Lesion uptake normalized to first imaging time point (A), partial-volume–corrected tumor-to-liver concentration ratio (B), eSNR (C), and eCNR (D) for 48 lesions identified in 10 subjects as function of imaging time point after [²⁰³Pb]Pb-VMT-α-NET SPECT/CT.

Safety and Adverse Events

The adverse-event observation window was 10 times the physical half-life of ²⁰³Pb (i.e., 22 d). No adverse events were associated with administration of [²⁰³Pb]Pb-VMT-α-NET. Safety data and patient charts were independently monitored by the data and safety monitoring committee.

Whole-Body Excretion Kinetics, Tumor Uptake, and Retention

Radiopharmaceutical accumulation in tumors, kidneys, liver, and spleen was rapid, reaching a maximum concentration within 1–2 h after administration. The effective elimination half-lives of [²¹²Pb]Pb-VMT-α-NET in tissues of interest were 3.9 ± 1.0 h (range, 3.0–6.25 h) for the whole body, 7.2 ± 1.4 h (range, 5.1–9.8 h) for the kidneys, 6.3 ± 0.7 h (range, 4.8–7.3 h) for the spleen, 6.7 ± 0.8 h (range, 5.2–8.0 h) for the liver, 1.7 ± 1.0 h (range, 0.69–3.59 h) for the bladder, and 7.6 ± 1.5 h (range, 3.2–10.8 h) for tumors. Overall, tissue clearance kinetics were approximately exponential, with the exception of those for the bladder and blood, which exhibited biexponential clearance behavior. This is ostensibly attributable to the short physical half-life of ²¹²Pb, leading to the later portion of a typical biphasic biologic clearance pattern being suppressed by physical decay. Measured whole-body, blood, and tumor time–activity curves for [²¹²Pb]Pb-VMT-α-NET are summarized in Figure 4.

FIGURE 4.

[²¹²Pb]Pb-VMT-α-NET time–activity curves in whole body (A), whole blood samples (B), and tumors (C). Curves represent experimental [²⁰³Pb]Pb-VMT-α-NET measured data, corrected for differences in physical half-life between ²⁰³Pb and ²¹²Pb.

Dosimetry of [²⁰³Pb]Pb-VMT-α-NET

Organs receiving the highest radiation exposure from [²⁰³Pb]Pb-VMT-α-NET included the urinary bladder wall (0.47 ± 0.18 mGy/MBq), kidneys (0.22 ± 0.12 mGy/MBq), spleen (0.13 ± 0.07 mGy/MBq), and liver (0.067 ± 0.041 mGy/MBq). The mean whole-body dose was 0.24 ± 0.23 mGy/MBq. The effective dose equivalent (23) and effective dose (22) were 0.066 mSv/MBq (2.46 mSv/mCi) and 0.038 mSv/MBq (1.40 mSv/mCi), respectively. The effective dose from the 185-MBq dose was 7.0 mSv, with an additional 2–5 mSv of effective dose per time point attributable to CT imaging.

Predicted Dosimetry of [²¹²Pb]Pb-VMT-α-NET

Patient-specific organ doses calculated for [²¹²Pb]Pb-VMT-α-NET on the basis of ²⁰³Pb imaging data are summarized in Table 2. The highest normal organ dose was in the kidneys (15.0 ± 5.2 mGy/MBq), followed by the spleen (8.4 ± 4.1 mGy/MBq). An accounting of our dosimetry uncertainty estimation is included in the supplemental materials. Excluding the effects of ²¹²Bi translocation modeling and the effect of time delay between dosimetry and treatment, our estimated type A and type B uncertainties (specified as 1σ) are 4.5% and 6.7%, respectively, where type A–evaluated uncertainties were determined by experimental statistical methods and type B–evaluated uncertainties were determined by other methods. The inclusion of translocation and time-delay effects between imaging and therapy increased our type B estimated uncertainty to 14.5%. Therefore, our combined standard uncertainty was approximately 15.2% for renal absorbed dose specifications for [²¹²Pb]Pb-VMT-α-NET on the basis of dosimetry derived from [²⁰³Pb]Pb-VMT-α-NET imaging. This specification is based on the standard root-sum-of-squares method in combining sources of uncertainty (including type A– and type B–evaluated uncertainties), as discussed in the National Institutes for Standards and Technology and International Organization for Standardization guides to the expression of measurement uncertainty (33,34).

TABLE 2.

²¹²Pb-VMT-α-NET Dosimetry Results

	Dose (mGy/MBq), by patient no.
Tissue	1	2	3	4	5	6	7	8*	9*	10*	Mean ± SD dose (mGy/MBq)
R kidney	16.6	10.8	15.9	12.0	19.2	25.4	14.6	19.4	12.1	7.3	15.3 ± 5.2
L kidney	13.8	11.2	17.1	11.6	16.5	23.5	13.3	17.6	14.2	7.9	14.7 ± 4.3
Kidneys	15.3	11.0	16.5	11.8	17.8	24.4	13.9	18.5	13.2	7.6	15.0 ± 4.7
Spleen	12.70	11.92	8.41	5.84	7.14	4.22	6.24	15.6	3.97	—	8.43 ± 4.10
Liver	3.73	3.00	3.16	1.41	1.97	1.76	2.35	3.51	2.03	1.68	2.46 ± 0.84
Blood	0.38	0.22	0.24	0.30	0.41	0.41	0.32	0.38	0.51	0.22	0.32 ± 0.11
Whole blood	0.30	0.27	0.30	0.22	0.24	0.27	0.27	0.46	0.30	0.19	0.27 ± 0.08
All tumors	—	—	—	—	—	—	—	—	—	—	29.6 ± 25.8
Tumor 1	11.8	9.6	23.8	—	14.7	6.6	30.8	43.1	12.2	99.7	—
Tumor 2	6.2	10.7	10.1	—	10.7	0.89	39.0	46.2	2.54	55.9	—
Tumor 3	9.3	13.0	89.5	—	12.6	17.8	30.7	24.4	4.0	102.3	—
Tumor 4	53.6	9.7	—	—	10.8	32.3	30.2	75.7	5.6	77.0	—
Tumor 5	36.6	9.8	—	—	30.8	19.2	19.9	90.1	5.3	35.5	—
Tumor 6	43.4	11.6	—	—	40.7	—	38.0	—	13.8	49.8	—
Tumor 7	16.4	12.4	—	—	46.6	—	21.0	—	6.8	45.6	—

^*Patients 8–10 received coinfusion of amino acids.

DISCUSSION

The elementally equivalent ²⁰³Pb/²¹²Pb theranostic pair offers significant advantages for personalized dosimetry-based image-guided therapy. For NETs, [²¹²Pb]Pb-DOTAMTATE has shown great promise in a phase 1 trial, with objective responses seen in 8 of 10 participants (14). Molecular modifications to peptide structures, including chelators, can significantly change not only the binding affinities of radiopeptides to targeted receptors but also the stability of the radiometal coupling. We previously reported improved receptor binding and a higher tumor-to-kidney uptake ratio of [²⁰³Pb]Pb-VMT-α-NET compared with [²⁰³Pb]Pb-DOTATOC (25). In preparation for our study of [²¹²Pb]Pb-VMT-α-NET with dose escalation based on renal absorbed dose rather than administered activity, we conducted this imaging and dosimetry study in 10 participants to assess the tumor-targeting ability of [²⁰³Pb]Pb-VMT-α-NET and measure the absorbed doses to the organs and tumors expected from the therapeutic agent [²¹²Pb]Pb-VMT-α-NET.

Our study conclusively demonstrated that SSTR2-positive lesions are targeted with [²⁰³Pb]Pb-VMT-α-NET. The sensitivity of [²⁰³Pb]Pb-VMT-α-NET SPECT/CT imaging among lesions larger than 1 cm was 94%. The visualization of lesions was significantly lower, at 35% for tumors 1 cm or smaller or those lesions not measurable on cross-sectional imaging. The lower sensitivity of [²⁰³Pb]Pb-VMT-α-NET in lesions smaller than 1 cm is consistent with our phantom data, which showed a significant drop in sensitivity for smaller lesions, likely reflecting the lower resolution of SPECT when compared with PET. The sensitivity of [²⁰³Pb]Pb-VMT-α-NET appeared lower for bone lesions, which were considered nonmeasurable on CT. We believe this likely reflects the small size of these lesions rather than inherent lower binding of the radiopharmaceutical agent, as the sensitivity for all nonmeasurable lesions was similar to that of lesions smaller than 1 cm.

Our findings demonstrate that, among the time points evaluated (1, 4, 24, and 48 h), imaging at 4 h with [²⁰³Pb]Pb-VMT-α-NET provided the highest lesion signal-to-noise ratio, followed closely by imaging at 1 h. Although the tumor-to-liver ratio continued to increase at later time points beyond 4 h, the noise penalty of lower count rates reduces lesion detectability beyond the first day. Of note, the performance of [²⁰³Pb]Pb-VMT-α-NET SPECT/CT in this study was to evaluate the targeting of lesions and not to determine the diagnostic accuracy of the imaging agent in a masked setting. We do not expect that [²⁰³Pb]Pb-VMT-α-NET can replace SSTR2 PET imaging agents for diagnostic purposes because of the inherent lower sensitivity of SPECT. However, given its high sensitivity in tumor targeting and its expected similar biodistribution to [²¹²Pb]Pb-VMT-α-NET, our study shows that [²⁰³Pb]Pb-VMT-α-NET is an excellent imaging and dosimetry analog for theranostic treatment with [²¹²Pb]Pb-VMT-α-NET.

The calculated whole-body clearance half-life of [²¹²Pb]Pb-VMT-α-NET was 3.9 ± 0.97 h. Interestingly the elimination rate of [²¹²Pb]Pb-VMT-α-NET appears similar to the FDA-approved PRRT agent, [¹⁷⁷Lu]Lu-DOTATATE, which has an overall elimination half-life of 3.5 ± 1.4 h and an initial rapid clearance that is followed by a slow clearance phase (35). Normal organ dosimetry predictions for [²¹²Pb]Pb-VMT-α-NET from this study seem to indicate that the renal absorbed dose will be treatment-limiting in most patients. Given a nominal expected relative biologic effectiveness of 3–7 (36), our measured renal dose of 15 ± 5 mGy/MBq, and a cumulative renal dose limit of approximately 38 Gy biological effective dose for β-PRRT (37), we expect that cumulative administered activities of 300–1,300 MBq (8–35 mCi) will be optimal for most patients, depending on individual patient’s renal dosimetry results. The lower end of this range is calculated on the basis of high renal dose per administered activity (∼19.7 mGy/MBq; 0.73 Gy/mCi) and high relative biologic effectiveness (∼7), with the higher end of the range corresponding to low renal dose per administered activity (∼10.3 mGy/MBq; 0.38 Gy/mCi) and low relative biologic effectiveness (∼3). With that said, our study included some participants who received an amino acid coinfusion (n = 3) and some participants who did not receive an amino acid (n = 7) coinfusion with [²⁰³Pb]Pb-VMT-α-NET. The decision to administer amino acid coinfusion for dosimetry was based on whether patients were intended to receive subsequent therapy with [²¹²Pb]Pb-VMT-α-NET. As such, the true population-average renal dose with [²¹²Pb]Pb-VMT-α-NET would likely be slightly lower than what we have reported here, given that patients will receive amino acid coinfusion during therapy.

One limitation of this imaging and dosimetry study is lack of exact knowledge of the true ²¹²Bi release fraction after the initial β-decay of ²¹²Pb. Mirzadeh et al. experimentally measured a release fraction of 36 ± 2% for the DOTA macrocycle (24), which is consistent with the 36.9% yield of K-shell conversion and Auger electrons (38), leaving the residual ²¹²Bi atom in a highly ionized state. Unlike ²²⁵Ac, for which daughter ion release is expected for 100% of initial decays because of the α-particle recoil energy, it is generally believed that, for ²¹²Pb, daughter ion release is the result of electron chemistry interactions between the chelator and the ionized daughter bismuth atom. Li et al. have published stability data for PSC in solutions of 1, 3, and 5 μM peptide in saline, demonstrating a concentration-dependent increase in ²¹²Bi-PSC chelate retention at 24 h (>95% with 5 μM) (25). This is an important improvement over ²¹²Pb-DOTA formulations because it significantly reduces the bolus of free ²¹²Bi that is expected to rapidly accumulate in the kidneys. Additional experiments are ongoing by our group, using methods similar to those of Mirzadeh et al. to determine the exact daughter release fraction for ²¹²Pb-PSC (24).

As of February 2025, 6 participants with gastroenteropancreatic NETs have undergone [²⁰³Pb]Pb-VMT-α-NET dosimetry-guided therapy with ²¹²Pb-VMT-α-NET within the phase 1 renal absorbed dose–escalation study at the University of Iowa (NCT06148636). Detailed results of this investigator-initiated clinical trial will be presented separately. Other future work will include investigation of the feasibility of direct ²¹²Pb-based imaging and dosimetry, as well as exploration of the utility of single-time-point dosimetry techniques for [^203/212Pb]Pb-VMT-α-NET.

CONCLUSION

[²⁰³Pb]Pb-VMT-α-NET appears safe and effective for single-photon imaging of SSTR2-positive NETs larger than 1 cm, as well as for purposes of normal organ and tumor radiation dosimetry. The chemically matched ²⁰³Pb/²¹²Pb theranostic pair offers the potential for a dosimetry-driven personalized treatment paradigm. To our knowledge, this work constitutes the first published human imaging and dosimetry study using ²⁰³Pb as a surrogate for a ²¹²Pb-based therapeutic radiopharmaceutical. The imaging and dosimetry findings suggest that [^203/212Pb]Pb-VMT-α-NET shows promise as a theranostic pair for imaging and therapy of SSTR2-positive gastroenteropancreatic NETs.

DISCLOSURE

This work was supported by grants from the National Institutes for Health (NCI R01CA167632 and NCI P50CA174521). Radiopharmaceutical support was received from Perspective Therapeutics. Michael Schultz is an employee of Perspective Therapeutics. Stephen Graves is a medical physics consultant to Perspective Therapeutics. No other potential conflict of interest relevant to this article was reported.

KEY POINTS

QUESTION: What are the imaging and dosimetric characteristics of [²⁰³Pb]Pb-VMT-α-NET and [²¹²Pb]Pb-VMT-α-NET?

PERTINENT FINDINGS: Optimal SPECT/CT diagnostic performance, in terms of lesion sensitivity, was observed approximately 4 h after administration of [²⁰³Pb]Pb-VMT-α-NET. Dosimetry indicates that the renal absorbed dose is likely to be treatment-limiting with [²¹²Pb]Pb-VMT-α-NET. Because of the short physical half-life of [²¹²Pb]Pb-VMT-α-NET, approximately 24 h of [²⁰³Pb]Pb-VMT-α-NET imaging data are required to accurately ascertain normal organ and tumor dosimetry.

IMPLICATIONS FOR PATIENT CARE: Dosimetry-guided α-particle PRRT with [^203/212Pb]Pb-VMT-α-NET for NETs appears feasible and promising.