…

The use of zirconium-89 as a radiolabel for monoclonal antibodies reflects a deliberate match between the physical properties of the radionuclide and the biological behaviour of the targeting molecule. Zirconium-89 has a half-life of approximately 78 hours, which is well suited to antibodies such as trastuzumab, whose intact immunoglobulin G structure circulates in the bloodstream for days and accumulates in target tissue slowly through binding to tumour-associated antigens. Shorter-lived positron emitters such as fluorine-18, with a half-life of roughly two hours, are poorly matched to intact antibodies because imaging would need to occur before adequate tumour localisation has taken place, resulting in unfavourable background activity. The 72-hour imaging window used in this patient, following the intravenous administration of 31.7 MBq of the tracer conjugated to 50 mg of the biosimilar trastuzumab Ontruzant, therefore represents a compromise that allows blood-pool activity to decline while retaining sufficient signal for detection. The DFO chelator, desferrioxamine, is the most widely used bifunctional chelator for zirconium-89 labelling of antibodies, forming a stable complex that holds the radiometal within the antibody construct as it distributes through the body. This pairing of radionuclide, chelator and antibody has become a standard framework for immuno-PET, and the present case extends that framework into a gynaecological malignancy where such imaging has previously been explored mainly in preclinical models.

The semi-quantitative measure reported in this case, the SUVratio, deserves particular attention when interpreting the observed uptake pattern. Standardised uptake values normalise measured radioactivity concentrations to the injected dose and body weight, but in immuno-PET the absolute values are influenced by additional factors including circulating antibody levels, antigen density on the cell surface, internalisation kinetics of the antibody-receptor complex, perfusion of the tumour tissue, and the time elapsed between injection and imaging. The ratio-based approach used here, comparing lesion uptake to a reference background region, partially mitigates some of these confounders and allows comparison across metastatic sites within the same patient. The striking gradient observed, from an SUVratio of 13.8 in the primary ovarian tumour through 3.6 in abdominal wall metastases down to 1.6 in hepatic lesions, spans a range that would be difficult to reconcile with a uniform HER2 status across all disease sites. This pattern is consistent with the biological reality of intrapatient tumour heterogeneity, a phenomenon that single-cell RNA sequencing and spatial transcriptomics studies have shown to be extensive in high-grade serous ovarian cancer, with distinct clones and microenvironmental contexts coexisting within a single tumour mass and across metastatic deposits.

The discordance between the intense uptake in the primary tumour, which corresponded to HER2 3+ expression on immunohistochemistry, and the near-background uptake in hepatic lesions raises important questions about the reliability of single-biopsy HER2 testing as a basis for treatment selection. Conventional HER2 assessment by immunohistochemistry and fluorescence in situ hybridisation examines only the tissue sampled, typically from the primary tumour or from a single metastatic site obtained at a single time point. In a disease characterised by peritoneal dissemination and frequent hepatic involvement, this sampling strategy may systematically miss sites where the target is absent or downregulated. If the hepatic lesions in this patient indeed expressed little or no HER2, they would be expected to derive little benefit from trastuzumab deruxtecan, whose cytotoxic payload, deruxtecan, acts on cells to which the antibody-drug conjugate delivers it, although a bystander effect on neighbouring antigen-negative cells has been proposed as a mechanism that may partially compensate for heterogeneous target expression. The mixed response pattern observed on the follow-up contrast-enhanced CT at 20 weeks, with some lesions regressing and others not, is compatible with such a scenario, although the authors of the report note that no clear relationship could be established between pre-treatment tracer uptake and the subsequent radiological behaviour of individual lesions.

This absence of a straightforward relationship between tracer avidity and response warrants careful consideration. Several factors may contribute. First, the pharmacokinetics of antibody-based therapeutics differ from those of the imaging tracer in ways that matter for prediction; trastuzumab deruxtecan has a modified antibody backbone and a cleavable linker-payload system, and its delivery to tumour cells depends not only on antigen binding but also on tumour perfusion, interstitial pressure and drug-release dynamics. Second, the timing of the imaging relative to treatment matters, and HER2 expression itself is dynamic, potentially changing under the selective pressure of prior chemotherapy lines. Third, response in ovarian cancer is often assessed against disease that is not measurable by RECIST criteria, as was the case here, where the qualitative regression of tracer-avid lesions on follow-up CT could not be captured by standard volumetric response categories. The CA125 decline from 676 to 254 U/mL over six weeks provided an early biochemical signal consistent with biological activity, and this biomarker, long established in the monitoring of epithelial ovarian cancer, complemented the imaging findings in a way that RECIST alone could not.

The concept of using paired tracers, one measuring a molecular target and the other measuring metabolic activity, is well illustrated by the ZEPHIR trial experience in breast cancer, where [89Zr]Zr-trastuzumab PET was combined with [18F]FDG PET to characterise HER2 expression and glucose metabolism simultaneously. That study demonstrated that lesions with low tracer uptake were enriched for genes involved in extracellular matrix organisation and glycosylphosphatidylinositol synthesis, while lesions with high uptake were associated with immune-related processes and were more likely to show metabolic response to trastuzumab emtansine. This transcriptomic dimension suggests that immuno-PET signal reflects not merely antigen density but a broader tumour biological context, including the microenvironmental composition of the lesion. In the oesophagogastric pilot study, [89Zr]Zr-trastuzumab PET detected fewer lesions overall than CT or FDG PET, yet it provided complementary lesion-level information, notably identifying a cerebellar metastasis with an SUV of 2.6 that was not visible on FDG PET and was subsequently confirmed by magnetic resonance imaging. These findings from other tumour types underscore that HER2-targeted immuno-PET is not a replacement for conventional imaging but a layer of molecular information that can be superimposed on anatomical and metabolic assessments.

The prevalence figures for HER2 overexpression in ovarian cancer, reported across studies to range from 6.6% to 39.5%, illustrate the challenge of patient selection for HER2-directed therapy in this disease. This wide range reflects variability in scoring criteria, antibody clones, cut-off thresholds and tissue sampling across studies, and it contrasts with the more uniform HER2 positivity seen in breast cancer, where trastuzumab-based therapy is firmly established. The DESTINY-PanTumor02 Phase II trial reported an objective response rate of 45% among patients with solid tumours showing HER2 overexpression, providing a rationale for extending trastuzumab deruxtecan beyond its approved indications into tumour-agnostic territory where HER2 expression, rather than tumour origin, drives eligibility. The ongoing DESTINY-Ovarian01 trial, evaluating trastuzumab deruxtecan combined with bevacizumab against bevacizumab alone as first-line maintenance therapy in HER2-expressing ovarian cancer, may further clarify the role of this agent in the gynaecological setting. Against this backdrop, a non-invasive method for confirming and mapping HER2 expression across all disease sites before committing a patient to an expensive and potentially toxic antibody-drug conjugate becomes clinically meaningful.

The patient described here had several features that constrained her treatment options and illustrate the complexity of real-world decision-making in platinum-resistant disease. Severe pre-existing hypertension precluded the use of bevacizumab, an anti-angiogenic agent commonly incorporated into ovarian cancer treatment algorithms. The extensive disease burden ruled out cytoreductive surgery, and the absence of a germline BRCA mutation removed the option of PARP inhibitor maintenance, while folate receptor alpha overexpression was also absent, excluding folate receptor-targeted approaches. This convergence of limitations left HER2, identified as 3+ on immunohistochemistry with a gastric scoring pattern, as the principal actionable target, and it was this finding that prompted the immuno-PET investigation. The case therefore demonstrates how molecular imaging can be deployed at a genuine decision point in a patient’s treatment trajectory, rather than as a purely investigational exercise detached from clinical consequence.

From a broader perspective, the integration of immuno-PET into the clinical pathway for ovarian cancer raises questions about workflow, cost and evidence generation that extend beyond the technical performance of the tracer itself. Zirconium-89 production requires a cyclotron and specialised radiochemistry infrastructure, and the 78-hour half-life, while advantageous for antibody imaging, means that patients must return for imaging on a separate day from injection, adding logistical complexity. The radiation dose associated with zirconium-89, while acceptable for oncological imaging in patients with life-limiting disease, is higher per unit of administered activity than that of fluorine-18, and this must be weighed against the informational value obtained. Furthermore, the interpretation of immuno-PET in a tumour type where the target is heterogeneously expressed requires careful attention to how uptake thresholds are defined and how partial volume effects in small peritoneal implants may distort apparent signal. The present case, with its combination of intense, moderate and near-background uptake across different sites, provides a useful illustration of the range of signal that may be encountered and the interpretive challenges that follow.

The observation that qualitative regression of tracer-avid lesions occurred on follow-up CT, despite the disease being non-measurable by RECIST, points to a potential role for molecular imaging in assessing response in situations where anatomical criteria fall short. Peritoneal disease in ovarian cancer is notoriously difficult to measure radiologically, and biomarkers such as CA125, while sensitive to change, do not localise the biological effect to specific lesions. Immuno-PET, by providing a spatially resolved map of target expression before treatment and potentially after treatment as well, could in principle identify which lesions are being reached by the therapeutic agent and which are escaping its effect. Serial immuno-PET, performed before and early after the initiation of targeted therapy, has been proposed as a means of pharmacodynamic assessment, and the present case, although limited to a single pre-treatment scan, lays the groundwork for such longitudinal applications in ovarian cancer.

The heterogeneity observed in this patient also connects to a wider theme in the molecular profiling of high-grade serous ovarian cancer, where single-cell and spatial transcriptomic approaches have revealed that tumour clones and microenvironmental components vary not only between patients but within individual tumour deposits. This biological reality means that any single-biopsy molecular test provides an incomplete picture, and it explains why patients with apparently similar HER2 status by immunohistochemistry may respond differently to the same targeted therapy. Immuno-PET, by surveying the whole body in a single examination, offers a way to capture some of this heterogeneity non-invasively, and the lesion-level information it provides may eventually be combined with genomic and transcriptomic data to build a more complete picture of each patient’s disease. The present report, while a single case, contributes to the growing evidence base suggesting that such multimodal integration, combining molecular imaging with conventional radiology, immunohistochemistry and biochemical markers, has a legitimate place in the personalised management of advanced ovarian cancer, particularly at the therapeutic crossroads represented by platinum-resistant disease.

Subject of Research: …

Article Title: Case Report: Non-invasive [89Zr]Zr-DFO-trastuzumab positron emission tomography in the multimodal characterisation of platinum-resistant high-grade serous ovarian cancer

Article References: Ferreira Machado, J. D. M., Winzer, R., Apolle, R., Forberger, A., Zarschler, K., Knieß, T., Kreller, M., Kopka, K., Kiß, O. C., Kuhlmann, J. D., Wimberger, P., Miederer, M., & Link, T. (2026). Case Report: Non-invasive [89Zr]Zr-DFO-trastuzumab positron emission tomography in the multimodal characterisation of platinum-resistant high-grade serous ovarian cancer. Frontiers in Nuclear Medicine, 6, Article 1832138. https://doi.org/10.3389/fnume.2026.1832138

Image Credits: AI Generated

DOI: 10.3389/fnume.2026.1832138

Keywords: Case, Report, Non-invasive, Zr-DFO-trastuzumab, positron, emission, tomography, multimodal, characterisation, platinum-resistant, high-grade, serous