The promise of therapeutic radiopharmaceuticals
Denis R Beckford-Vera, Head of Radiopharmacology at Champions Oncology explores how radiopharmaceuticals can benefit from the use of patient-derived xenograft models.
The rising promise of radiopharmaceuticals
Targeted radionuclide therapies (TRTs) are rapidly becoming a major force in oncology drug development. These agents combine a tumour-targeting molecule with a radioactive isotope, delivering cytotoxic radiation directly to cancer cells while sparing most healthy tissue.
In recent years, we have seen clinical successes with agents targeting prostate-specific membrane antigen (PSMA), HER2, and somatostatin receptors, alongside a pipeline of new constructs targeting diverse tumour antigens. The same molecular designs can also be paired with imaging isotopes, enabling PET or SPECT scans to map drug distribution and guide therapy, the basis of “theranostics.”
Yet, developing these complex agents is challenging. Their success depends not only on chemistry and targeting strategy, but also on the preclinical models used to evaluate them. Choosing a model that reflects the true biology of human tumours is essential to predicting clinical performance. Highly characterized patient-derived xenograft (PDX) models, which maintain the heterogeneity and microenvironment of patient tumours, offer a powerful platform to accelerate radiopharmaceutical development and reduce the risk of late-stage failure.
Core concepts of radiopharmaceutical development
At the heart of every TRT is what appears to be a “simple architecture”: a targeting ligand, such as a monoclonal antibody, small molecule, or peptide, linked to a radioactive isotope. While this design seems straightforward, in practice it is more complex. Each component, the choice of isotope, the stability of the chelator, the linker chemistry, and the pharmacokinetics of the targeting ligand must be precisely aligned to ensure the therapy reaches the tumor effectively and delivers radiation with the desired therapeutic index.
Therapeutic isotopes fall into two main categories. Alpha emitters, such as Ac-225, Pb-212, and At-211, deliver high-energy particles with short path lengths, creating intense, localized damage that is ideal for micrometastases and isolated tumor cells. Beta emitters, such as Lu-177, Y-90, and Tb-161, have longer penetration, making them well suited to bulky or heterogeneous tumors. Imaging isotopes, including Zr-89, Ga-68, and In-111, enable companion diagnostic development, patient selection, and dosimetry studies. From a radiochemistry standpoint, chelation secures the isotope to the targeting ligand, linker stability ensures it stays attached until it reaches the tumor, and isotope half-life must be matched to the ligand’s pharmacokinetics. While these chemical considerations are critical, the choice of biological model often has a greater impact on whether preclinical data will translate to patients.
The role of preclinical models in TRT development
TRTs present unique preclinical challenges. Developers must demonstrate that the therapy specifically targets tumors while minimising uptake in critical organs. They also need to characterise biodistribution and dosimetry to estimate radiation dose delivered to tumour versus normal tissues, and assess pharmacokinetics and pharmacodynamics in a physiologically relevant system. Finally, efficacy must be shown against tumor types and disease stages that reflect the intended clinical population. Inadequate models risk generating misleading data, such as high uptake in an artificial tumour environment that would never be achieved in a human patient. This is where model selection becomes a decisive factor in translational success.
Why PDX models supercharge radiopharmaceutical R&D
PDX models are created by implanting fragments of a patient’s tumour directly into immunodeficient mice. Unlike cell-line–derived xenografts (CDX), PDX tumours maintain the architecture, stromal composition, and genetic heterogeneity of the original patient sample. Many PDX collections also include tumours from patients previously treated with chemotherapy, targeted therapy, or immunotherapy, providing models with clinically relevant resistance mechanisms.
For TRTs, this fidelity is invaluable. Human-derived biology ensures antigen expression levels, vascular architecture, and tumour microenvironment are clinically relevant. Molecular characterisation through genomics, transcriptomics, and proteomics allows researchers to stratify models based on target expression and mutational status. This level of clinical realism increases the likelihood that observed biodistribution, retention, and therapeutic index in PDX will predict patient outcomes. The result is more accurate dosimetry, better go/no-go decisions, and a stronger foundation for regulatory submissions.
Case example: biodistribution and efficacy in PDX
A typical workflow for evaluating a TRT in PDX might involve radiolabelling the candidate molecule with a therapeutic and/or imaging isotope, followed by dosing into a cohort of PDX models selected for high, moderate, and low target expression. Ex-vivo tissue cut-and-count or imaging with PET or SPECT can then quantify tumour uptake and assess off-target accumulation in organs such as the liver, kidneys, or bone marrow. These results inform dosimetry calculations to estimate absorbed radiation doses to tumor and normal tissues. Finally, therapeutic studies measure tumor growth delay, regression, or survival benefit. When imaging and therapy are integrated in the same models, researchers gain both mechanistic insight into where the drug goes and how long it stays, as well as efficacy data, enabling a more complete translational picture. To understand potential toxicities, a biologically relevant species should also be selected.
Applications across target classes: large molecule, small molecule, and beyond
Large molecules, such as monoclonal antibodies (mAbs), have been widely used in HER2 and EGFR targeting, with some applications in PSMA. However, for PSMA specifically, small molecules currently dominate clinical development. Radiolabeling extends the role of antibodies from growth inhibition to direct tumor cell killing, and PDX models reveal how antibody penetration and retention vary with tumour size, vascularity, and antigen heterogeneity. Antibody-drug conjugates (ADCs) can also be radiolabeled, allowing researchers to track both the monoclonal antibody and the payload depending on the position of the radiolabel. PDX studies capture real-world linker stability and payload release patterns, helping developers better anticipate clinical performance.
Small molecules, including enzyme inhibitors and receptor ligands such as PSMA, GRPR, and integrins, offer rapid tumour penetration and clearance from normal tissue. PDX models allow kinetic comparisons across tumours with different genetic drivers. Peptides, such as somatostatin and bombesin analogs, have shown success in neuroendocrine and prostate cancers, and PDX models help validate tumour-to-background ratios and inform patient selection.
Emerging modalities, such as bispecific and multi-specific binders, can be studied in PDX to evaluate simultaneous engagement of multiple tumor antigens and to measure efficacy in heterogeneous tumors. Similarly, hypoxia- or DNA damage-targeting ligands can be tested in PDX models with documented hypoxic gradients, providing realistic conditions for these innovative approaches.
Combination strategies are also advancing. TRTs combined with checkpoint inhibitors leverage the ability of radiation to increase antigen presentation, and humanised PDX models provide a powerful route for preclinical proof-of-concept. TRTs combined with DNA damage repair inhibitors also hold promise, with PDX enabling biomarker-guided studies to identify tumours most likely to respond.
Regulatory and translational advantage
PDX-generated data can directly strengthen the regulatory case for a TRT. Robust evidence of target engagement, biodistribution, and safety margins provides confidence in IND-enabling packages. Dosimetry from ex-vivo cut-and-count or imaging studies can inform first-in-human dose planning by predicting safe and effective dosing. Imaging isotopes validated in the same targeting scaffold can support companion diagnostic development to guide patient selection. Regulators increasingly expect data that demonstrates a therapy’s performance in clinically relevant models, and PDX can bridge the gap between concept and clinic.
Choosing the right model for the right question
While PDX models are often the gold standard for evaluating therapeutic radiopharmaceuticals, other preclinical systems can be better suited for certain stages of development or specific research needs. Cell-line–derived xenografts (CDX) are best suited for early proof-of-concept and high-throughput screening when speed, cost, and reproducibility are priorities. Their consistent tumor growth kinetics make them ideal for comparing multiple candidates quickly before moving to more complex models.
Organoids and patient-derived organoids (PDOs) are valuable for validating target expression, screening compound libraries, and studying tumor biology in a controlled environment. They enable rapid iteration on chemistry before committing to in vivo studies, particularly for target discovery or early ligand optimization.
Syngeneic and genetically engineered mouse models (GEMMs) are critical when immune system interactions are central to the research question, such as in immuno-radiotherapeutics or combinations with checkpoint inhibitors. Because these models maintain intact immune responses, they allow evaluation of immune-mediated effects of TRT.
In silico and computational models also play an important role, supporting dosimetry predictions and PK/PD modeling using preclinical datasets. They can refine study designs and help predict likely responders before in vivo testing. In many programs, these models are complementary rather than competitive. A rational development strategy often begins with simpler or higher-throughput systems to narrow the field, then advances into highly characterized PDX models for clinically predictive validation and regulatory confidence.
The next decade of TRT and PDX
Theranostics, pairing the same molecular targeting vector with both imaging and therapeutic isotopes, is likely to dominate the TRT landscape. Combination regimens will continue to expand, incorporating TRTs with DNA repair inhibitors, targeted agents, and immunotherapies. Advances in humanized PDX models, which reconstitute components of the human immune system, will also open new doors for evaluating immuno-radiotherapeutics. As these technologies evolve, the integration of precise radiochemistry with patient-derived biology will remain central to success.
The path to better oncology outcomes
Therapeutic radiopharmaceuticals offer unique potential to precisely target and destroy cancer cells. Realising this potential depends on the models we use to develop them. Highly characterized PDX models capture the complexity of human tumours, enabling more accurate biodistribution studies, more predictive efficacy data, and stronger regulatory packages. When combined with advances in radiochemistry and targeting design, they can accelerate development timelines and improve the odds of clinical success.
In the future, the marriage of cutting-edge radioactive therapeutics with the biological realism of PDX models will be one of the defining drivers in oncology drug discovery and development, bringing new hope to patients who need it most.
