The research, published in Frontiers in Nuclear Medicine, focused on neuroendocrine neoplasms, a heterogeneous group of tumors that most commonly arise in the gastrointestinal tract and lungs. These tumors range from slow-growing, well-differentiated neuroendocrine tumors to highly aggressive neuroendocrine carcinomas. The study included 23 patients with aggressive disease, including G2 and G3 neuroendocrine tumors, neuroendocrine carcinomas, and mixed neuroendocrine-non-neuroendocrine neoplasms, all of whom underwent both FDG and FAPI PET/CT scans along with a biopsy within a clinically narrow time window.

Dual-tracer PET/CT works by exploiting two different biological properties of tumors. FDG, a radioactive glucose analog, accumulates in cells with high metabolic activity, which typically reflects rapid proliferation and dedifferentiation. FAPI, on the other hand, targets fibroblast activation protein, which is overexpressed by cancer-associated fibroblasts in the tumor stroma. These stromal cells play critical roles in tumor invasion, proliferation, metastasis, and angiogenesis. When a lesion shows FDG uptake but no corresponding FAPI signal, it suggests a tumor subclone that has shifted toward an aggressive, dedifferentiated phenotype with low stromal activation but high glucose metabolism.

The researchers identified these discordant FDG-positive, FAPI-negative lesions in 9 of the 23 patients. Most frequently, these lesions appeared in patients with neuroendocrine carcinomas, followed by those with G3 tumors. The discordant lesions were located in the liver, lymph nodes, and skeleton. Notably, among the nine patients with discordant lesions, biopsies were obtained from the same organ in only two cases, and even then, lesion-level correspondence could not be confirmed in one instance.

Quantitative analysis revealed that the immunoreactive score for FAP expression correlated significantly with tumor volume and total lesion uptake for both radiotracers, but showed no significant correlation with standardized uptake value parameters such as SUVmax or SUVpeak. The researchers suggest this may be because voxel-level readouts capture only the hottest voxels, which may reflect heterogeneous outlier clusters unrepresentative of true stromal density, whereas global metrics like tumor volume and total lesion uptake average out these outliers and better parallel overall FAP expression.

Crucially, the presence of discordant lesions was associated with significantly shorter progression-free survival. Patients with FDG-positive, FAPI-negative lesions had a median progression-free survival of 4 months compared with 10 months for those without such lesions. This association remained independently significant in multivariate Cox regression after adjustment for age, with a hazard ratio of 8.269. In contrast, histological FAP expression showed no statistically significant influence on progression-free survival.

Of the PET parameters, only FDG-derived tumor volume and total lesion uptake were significantly associated with shorter progression-free survival, whereas FAPI-derived metrics were not. These results align with previous reports highlighting the prognostic value of FDG uptake in aggressive neuroendocrine neoplasms and suggest that tumor metabolism remains a more reliable predictor of outcome than stromal activity in this patient population.

The study has important implications for how clinicians approach risk stratification in these rare tumors. Because aggressive neuroendocrine neoplasms are known for substantial intrapatient heterogeneity, a single biopsy may not adequately capture the spatial diversity of stromal activation across the whole-body tumor burden. Dual-tracer PET/CT, as a whole-body diagnostic tool, provides insight into tumor phenotype across all lesions simultaneously and may offer superior risk stratification compared to immunohistochemical examination of FAP expression from a single tumor site.

The researchers acknowledge several limitations, including the retrospective single-center design, the modest sample size, and the spatial mismatch between biopsy sites and imaging-defined discordant lesions. They emphasize that future prospective studies with spatially matched biopsies are essential to validate the tissue-imaging relationship at the lesion level, especially given growing interest in stroma-targeted therapies and FAP-directed radioligand theranostics. Nevertheless, the highly significant association between discordant lesions and shorter survival supports its role as a prognostically relevant imaging biomarker that could guide individualized treatment decisions in precision oncology.

The distinction between the two tracers reflects fundamentally different tumor biology. Somatostatin receptor expression, which underpins DOTATATE-based imaging and peptide receptor radionuclide therapy, is typically retained in well-differentiated neuroendocrine tumors but lost as tumors dedifferentiate. FDG uptake, conversely, increases with proliferative activity and loss of differentiation, which is why it has long served as a marker of aggressive biology in this disease spectrum. FAPI imaging adds a third dimension by visualizing the tumor microenvironment rather than the tumor cells themselves, since the fibroblast activation protein is expressed mainly by activated fibroblasts within the stroma. A lesion that is metabolically active on FDG but lacks stromal FAP signal therefore represents a phenotype in which tumor cells drive glucose consumption while the supporting stromal compartment remains quiescent, a combination that appears to mark particularly adverse biology.

The quantitative findings of the study also carry methodological interest. The immunoreactive score correlated with volumetric measures such as tumor volume and total lesion uptake, but not with intensity-based measures like SUVmax, SUVmean, or SUVpeak. This pattern suggests that the sheer amount of metabolically active or stroma-rich tumor tissue, rather than the intensity of uptake in any single spot, mirrors the histological FAP signal best. Voxel-level maxima can be driven by small clusters of intensely avid cells or by partial volume effects, whereas volume-based parameters integrate signal across the entire lesion burden. For researchers designing future correlative imaging-pathology studies, this argues for prioritizing volumetric and total-lesion metrics when comparing whole-body imaging with tissue-based readouts.

The prognostic separation observed in the cohort was striking given its size. A median progression-free survival of 4 months versus 10 months, with a hazard ratio above 3 in univariate analysis and an even higher adjusted estimate in multivariate modeling, indicates that the discordant-lesion phenotype is not a subtle imaging curiosity but a clinically meaningful divider. It is worth noting that FDG-derived tumor volume and total lesion uptake also tracked with progression-free survival, reinforcing the established role of metabolic tumor burden as a prognostic parameter. FAPI-derived volumetric metrics, by contrast, did not reach significance for outcome prediction, which the authors interpret as further evidence that stromal activation, at least as captured by FAPI uptake, is less informative about aggressiveness than glucose metabolism in these tumors.

The spatial mismatch between biopsy and imaging deserves particular emphasis. Immunohistochemistry from a single site provides a snapshot of one lesion at one moment, while dual-tracer PET surveys the entire body. In a disease characterized by substantial inter- and intralesional heterogeneity, a biopsy from a liver metastasis may show abundant FAP expression while an occult lymph node or bone lesion behaves entirely differently. The finding that biopsies corresponded to the same organ as discordant lesions in only a minority of cases illustrates how easily tissue-imaging correlations can be obscured by sampling geography. This has practical consequences: if a biopsy is being considered to guide therapy selection, imaging can help identify which lesion is most likely to yield biologically representative tissue, or whether re-biopsy of a metabolically aggressive site is warranted.

From a therapeutic standpoint, the results intersect with two evolving areas. First, FDG-positive lesions in neuroendocrine neoplasms often indicate dedifferentiated subclones that may no longer respond to somatostatin receptor-directed approaches, making their detection a trigger for treatment escalation or alternative systemic regimens. Second, FAP is increasingly explored as a theranostic target, with radiolabeled FAP inhibitors under investigation for both imaging and endoradiotherapy. If FDG-positive, FAPI-negative lesions predict poor outcomes, they may also identify patients in whom FAP-directed theranostics would be unlikely to cover the full tumor burden, informing patient selection for such trials. Conversely, patients whose lesions are concordantly FAPI-positive might represent better candidates for stroma-targeted interventions.

The study design also illustrates the value of close temporal alignment between imaging and tissue sampling. By requiring that the two PET scans be performed within three months of each other and that no tumor-specific therapy intervene between scans and biopsy, the investigators reduced the confounding that treatment effects and disease progression would otherwise introduce. Even so, the maximum allowable interval of twelve months between imaging and biopsy remains a potential source of discordance, since tumor biology can evolve over such periods, particularly under selective pressure from prior therapies.

Looking forward, the central question raised by these findings is whether lesion-level tissue validation can confirm that FDG-positive, FAPI-negative lesions truly represent dedifferentiated, stroma-poor subclones, or whether technical factors such as tracer kinetics, lesion size, and resolution thresholds contribute to the apparent discordance. Prospective studies pairing targeted biopsies of imaging-defined discordant lesions with quantitative immunohistochemistry, and ideally with spatial transcriptomics or other high-resolution molecular methods, would clarify the biological substrate of the phenotype. Until such data are available, the combination of FDG and FAPI PET/CT offers clinicians a practical, noninvasive window into tumor heterogeneity that a single biopsy cannot match, and the presence of discordant lesions should prompt consideration of more aggressive disease biology when treatment decisions are being weighed.

Subject of Research: Prognostic value of discordant FDG-positive FAPI-negative lesions on dual-tracer PET/CT compared to histological FAP expression in aggressive neuroendocrine neoplasms

Article Title: Prognostic impact of discordant lesions on [18F]FDG and [68Ga]Ga-FAPI-04 PET/CT compared to histological FAP expression in neuroendocrine neoplasms

Article References: Jedamzik, T., Kosmala, A., Kircher, S., Heinrich, M., Schlötelburg, W., Buck, A. K., Meining, A., Werner, R. A., Weich, A., & Michalski, K. (2026). Prognostic impact of discordant lesions on [18F]FDG and [68Ga]Ga-FAPI-04 PET/CT compared to histological FAP expression in neuroendocrine neoplasms. Frontiers in Nuclear Medicine, 6, Article 1777541. https://doi.org/10.3389/fnume.2026.1777541

Image Credits: AI Generated

DOI: 10.3389/fnume.2026.1777541

Keywords: neuroendocrine neoplasms, dual-tracer PET/CT, FDG, FAPI, fibroblast activation protein, discordant lesions, progression-free survival, tumor heterogeneity, immunohistochemistry, risk stratification, nuclear medicine, precision oncology