The study, published in the European Journal of Nuclear Medicine and Molecular Imaging, was motivated by a stubborn clinical problem. Immune checkpoint inhibitors targeting the PD-1/PD-L1 axis have transformed the treatment of melanoma and lung cancer, producing durable responses and long-term survival in subsets of patients. Prostate cancer, however, has largely resisted this revolution. Only small fractions of men with mCRPC benefit from agents such as pembrolizumab or nivolumab, a failure attributed to the disease’s low tumor mutational burden and its profoundly immunosuppressive microenvironment. Prostate tumors are, in immunological terms, cold: they present few recognizable targets to T cells and actively suppress the immune cells that do infiltrate them.

Radiation has long been proposed as a way to heat those cold tumors up. External beam radiotherapy can damage cancer cells in ways that release tumor antigens, trigger inflammatory cytokine secretion, recruit dendritic cells, and activate anti-tumor T cells. In rare, striking cases, irradiating one tumor site has even shrunk untreated lesions elsewhere in the body, the so-called abscopal effect. Clinical trials combining radiation with checkpoint inhibitors have shown hints of enhanced activity in genitourinary cancers. Radium-223, an intravenously administered alpha emitter approved for mCRPC patients with predominantly bone metastases, seemed like an elegant systemic version of this strategy. In the pivotal phase III ALSYMPCA trial it extended overall survival by 3.6 months, and because it circulates through the bloodstream it can in principle irradiate every bone lesion at once.

Earlier laboratory work had fed optimism. In vitro experiments showed that sublethal exposure to radium-223 enhanced the ability of cytotoxic T lymphocytes to kill prostate cancer cells by inducing MHC class I and calreticulin, proteins central to antigen presentation. Studies of patient blood revealed that radium-223 enriched DNA damage repair pathways in plasma exosomes and raised exosomal levels of PD-L1, LAG-3 and IDO, immune checkpoints associated with immunosuppression and poorer survival. A mouse model of prostate cancer even suggested that combining radium-223 with anti-PD-1 and anti-CTLA-4 antibodies produced greater tumor regression than either immunotherapy alone. The clinical rationale for combining the radionuclide with checkpoint blockade appeared, on the surface, compelling.

To test that rationale rigorously in humans, the research team designed a biomarker discovery study in which patients received up to six injections of radium-223 at four-week intervals, while undergoing an unusually demanding battery of measurements. The centerpiece was sequential PET imaging with zirconium-89 labeled atezolizumab, a radiolabeled version of the approved PD-L1 antibody atezolizumab. Produced under Good Manufacturing Practice at the University Medical Center Groningen and injected at a tracer dose of 37 MBq, roughly 11 milligrams of antibody and less than one percent of a therapeutic dose, the tracer allowed the team to quantify whole-body PD-L1 expression in vivo, something a single biopsy can never do. Gallium-68 PSMA PET/CT served as the anatomical and molecular reference, pinpointing viable prostate cancer lesions in the skeleton against which atezolizumab uptake could be measured.

The imaging results were striking. At baseline, uptake of the PD-L1 tracer in PSMA-positive bone metastases was low overall and highly heterogeneous, both between patients and between individual lesions within the same patient, with a median standardized uptake value of 2.5 across the total volume of PSMA-positive bone disease. After three cycles of radium-223, that median fell to 0.9, a statistically significant median decrease of 33 percent, while PSMA uptake in the same lesions remained essentially stable. Crucially, the decline was independent of whether a patient’s disease responded on PSMA PET: good responders, defined as those remaining on radium-223 for at least 24 weeks, and poor responders showed comparable drops. Some lesions even lost PD-L1 signal while their PSMA signal held steady, and new PSMA-avid disease appearing at the borders of treated lesions showed lower tracer uptake than the original metastases had at baseline.

Tissue analysis told a consistent story. Guided by PSMA PET, the team obtained bone metastasis biopsies at baseline in 25 patients, with tumor cells present in 76 percent of samples. Multiplex immunohistochemistry revealed low PD-L1 expression, found mainly on tumor cells but also on non-tumor cells, mirroring the modest and variable PET signal. In the small number of patients who provided sequential biopsies, the number of lymphocytes in both tumor tissue and surrounding bone marrow decreased after three cycles of radium-223, and one on-treatment biopsy showed necrosis consistent with radiation effect. Flow cytometry of blood samples reinforced the picture: circulating lymphocyte counts fell during therapy, and good responders showed a significant rise in TIM3-positive T cells, a marker of T cell exhaustion, after treatment.

The authors argue that the drop in atezolizumab uptake most likely reflects a treatment-induced loss of immune cells within the tumor microenvironment, and possibly a reduction in PD-L1 expression per cell, rather than simply fewer tumor cells. Because PSMA uptake did not fall in parallel, and because new PSMA-positive disease did not carry elevated PD-L1 signal, the immune compartment appears to be the dominant source of the tracer signal and the main casualty of the alpha radiation. The team acknowledges limitations, including the possibility of confounding by changes in tumor vascularization or receptor saturation, though the stability of tracer uptake in bone marrow and cortex over time supports a genuine treatment effect. Sequential biopsies were feasible in only a handful of patients, and decalcification of bone samples may have masked PD-L1 staining with standard clinical antibodies.

The clinical implications are sobering but clarifying. The findings align with the disappointing results of phase Ib and phase II trials that added radium-223 to atezolizumab, pembrolizumab or enzalutamide and found no benefit for mCRPC patients. Far from sensitizing prostate cancer to PD-L1 blockade, alpha radiation appears to thin out the very immune cells that checkpoint inhibitors rely upon, potentially undermining immunogenic effects of irradiation through sheer depletion of the lymphocyte pool. The researchers suggest that future combination strategies might instead turn to radionuclides with lower linear energy transfer delivered at lower doses, such as lutetium-177 PSMA-targeted therapy, which could induce immunogenic signaling without devastating the immune infiltrate. For now, the study stands as a powerful demonstration of what whole-body immune PET can reveal: a molecular narrative of treatment response that biopsies alone could never tell, and a cautionary lesson about assuming that radiation and immunotherapy are natural allies in every tumor type.

Subject of Research: The effect of radium-223 alpha radiation on PD-L1 expression and the immune microenvironment of bone metastases in metastatic castration-resistant prostate cancer, measured with zirconium-89 atezolizumab PET imaging and immunohistochemistry.

Article Title: Radium-223 decreases PD-L1 expression in bone metastases: A study with [89Zr]Zr-atezolizumab PET and immunohistological correlation

Article References: Ling, S. W., de Jong, A., Gorris, M., Segbers, M., Graven, L., van Driel, M., van Leenders, G., Isebia, K., de Hooge, M. L., Hamberg, P., Mehra, N., van Brakel, M., de Wit, R., Debets, R., de Vries, E., Brabander, T., de Vries, J., & van der Veldt, A. (2026). Radium-223 decreases PD-L1 expression in bone metastases: A study with [89Zr]Zr-atezolizumab PET and immunohistological correlation. European Journal of Nuclear Medicine and Molecular Imaging. https://doi.org/10.1007/s00259-026-08175-9

Image Credits: AI Generated

DOI: 10.1007/s00259-026-08175-9

Keywords: radium-223, prostate cancer, PD-L1, atezolizumab, PET imaging, bone metastases, immune checkpoint inhibitors, PSMA, immunotherapy, alpha radiation, tumor microenvironment, mCRPC