FAPI PET Not Yet Validated for Thyroid Cancer Response

A narrative review published in Frontiers in Endocrinology concludes that fibroblast activation protein inhibitor (FAPI) positron emission tomography (PET) is not yet a validated biomarker for assessing treatment response in radioiodine-refractory differentiated thyroid cancer (RAIR-DTC). The study, led by researchers at the Cancer Hospital of Shantou University Medical College in China, critically evaluates the potential of FAPI PET to monitor patients receiving antiangiogenic tyrosine kinase inhibitors (TKIs). While the imaging technique shows promise in detecting lesions, the authors emphasize that current evidence is insufficient to support its use for guiding clinical decisions regarding therapy continuation or discontinuation. The review highlights a significant gap between diagnostic detectability and longitudinal response validation, urging the scientific community to prioritize rigorous, staged validation before clinical implementation.

The central challenge in managing RAIR-DTC is that standard response assessment tools, such as contrast-enhanced computed tomography (CT) evaluated via RECIST 1.1 criteria and serum thyroglobulin levels, may not fully capture early biological changes in the tumor microenvironment. Antiangiogenic TKIs, including lenvatinib, sorafenib, and cabozantinib, alter tumor perfusion, vascular permeability, and stromal signaling in ways that can diverge from anatomical shrinkage. This temporal decoupling suggests that molecular imaging could offer earlier insights into treatment efficacy. However, the review argues that plausibility alone does not translate into clinical utility, requiring a biomarker to demonstrate solid reproducibility and a clear correlation with long-term patient outcomes beyond standard evaluation tools.

FAPI PET targets fibroblast activation protein (FAP), which is abundant in activated tumor-associated stroma, particularly in cancer-associated fibroblasts. The theoretical advantage of this approach is its ability to visualize stromal changes that may precede structural tumor shrinkage. Yet, the measured uptake in FAPI PET is not a purely quantitative assay of FAP expression. It is influenced by tracer delivery, blood flow, vascular permeability, ligand kinetics, and non-malignant processes such as fibrosis, wound healing, and inflammatory repair. This complexity means that a change in FAPI uptake cannot be automatically equated with tumor control or progression, as antiangiogenic therapy directly alters the very physiological parameters that affect tracer accumulation.

The most direct evidence supporting the use of FAPI PET during TKI therapy in RAIR-DTC is limited to a single case report involving one patient. This 60-year-old man with papillary thyroid carcinoma and pulmonary metastases was treated with lenvatinib. After ten months, while CT and FDG PET/CT showed stable disease, FAPI uptake decreased markedly or resolved in several lesions. Although this divergence is biologically intriguing, the authors note that a single observation with a long scan interval cannot establish FAPI PET as an early response marker. It does not define quantitative cutoffs, prove an association with progression-free survival, or justify altering clinical management in a broader population.

Broader thyroid-specific cohorts confirm that FAPI uptake is common in RAIR-DTC, but tracer performance varies substantially depending on the specific ligand, lesion site, and cohort selection. For instance, a prospective cohort of 24 patients found that 68Ga-DOTA-FAPI-04 PET/CT detected disease in 87.5% of patients, with lesion SUVmax associated with subsequent growth. However, this study addressed diagnostic performance rather than longitudinal response. Other studies using different tracers, such as 68Ga-FAPI-46 and 18F-FAPI-74, showed variable detection rates and contrast levels compared to FDG PET. These inconsistencies underscore that findings with one FAPI compound cannot be assumed to apply quantitatively to another, preventing the establishment of universal response thresholds.

Cross-cancer studies provide indirect evidence that treatment can alter FAPI signal, but their relevance to RAIR-DTC remains limited. In breast, pancreatic, and ovarian cancers, longitudinal FAPI studies have shown that uptake changes can correlate with pathologic response and tissue FAP expression. For example, a prospective study of 22 breast cancer patients found that early changes in SUVmax were associated with pathologic response. However, these studies involve cytotoxic chemotherapy or chemoradiotherapy, which act via mechanisms entirely distinct from antiangiogenic TKIs. The tumor stroma in these cancers also differs markedly from that of thyroid carcinoma, meaning these studies serve primarily as methodological blueprints rather than direct validation for thyroid cancer.

Measurement reliability presents another significant hurdle. Emerging repeatability data indicate that FAPI PET has measurable test-retest behavior, but the coefficients of variation are substantial. A 2026 test-retest study in pancreatobiliary malignancies found repeatability coefficients of approximately 24% for SUVmean and SUVpeak, and 30% for SUVmax. These values are not RAIR-DTC-specific and cannot be adopted directly as response criteria. For a proposed response percentage to be meaningful, it must exceed this technical noise. Additionally, reader-related variability affects interpretation, with less experienced readers showing lower agreement in staging categories. This necessitates prespecified reader training, blinding, and adjudication in any future response studies.

The review outlines a staged validation pathway that must be completed before FAPI PET can be considered for clinical use. The first stage is baseline evaluability, determining if the intended TKI-treated population can be measured reliably without enrichment for known FAPI avidity. The second stage focuses on measurement reliability, establishing repeatability coefficients and minimum detectable change for each quantitative metric. The third stage involves clinical validity, assessing whether early FAPI changes relate to later clinically relevant outcomes such as progression-free survival. Only after these gates are passed can incremental value and clinical utility be evaluated, determining if acting on FAPI information actually improves patient outcomes or reduces harm.

Interpreting longitudinal FAPI changes under antiangiogenic therapy is complex due to the drug’s effect on tumor perfusion. A reduction in FAPI uptake could result from reduced tracer delivery due to perfusion changes, rather than a decrease in viable tumor burden. Conversely, treatment-induced stromal remodeling or inflammation could sustain or elevate tracer uptake even if the tumor is structurally controlled. The review emphasizes that these mechanisms are biologically plausible but have not been causally demonstrated in serial RAIR-DTC FAPI studies. Therefore, discordant FAPI findings should prompt careful multimodal review rather than automatic continuation, discontinuation, or switching of TKI therapy.

The authors recommend that future research prioritize prospective cohort studies enrolling consecutive patients initiating antiangiogenic TKIs for standard clinical indications. These studies should not mandate baseline FAPI positivity, as pre-selecting FAPI-avid patients would artificially inflate performance and obscure real-world evaluability. Incorporating a nested test-retest sub-study prior to TKI initiation will help establish disease-specific measurement noise. Protocols must lock in standardized scan windows, track TKI dose intensity meticulously, and capture parallel RECIST 1.1 measurements, TSH-adjusted thyroglobulin kinetics, and long-term survival endpoints. Biostatistical plans should account for intra-patient lesion clustering and utilize landmark or time-dependent models to avoid biases when linking early imaging changes to progression-free survival.

In conclusion, while FAPI PET has research value for longitudinal assessment in RAIR-DTC, diagnostic detectability cannot substitute for response-biomarker validation. Current evidence does not establish that treatment-associated FAPI change is a validated efficacy measure, predictive biomarker, or surrogate endpoint. The field must first define RAIR-DTC-specific repeatability and minimum detectable change, clarify the biological meaning of signal change under antiangiogenic therapy, and demonstrate incremental clinical value beyond standard assessment. Until these standards are met, FAPI uptake change alone should not be used to stop or switch TKI therapy or to confirm progression. A staged program from measurement validation to clinical utility is the most direct route to determining whether FAPI PET can ultimately support treatment decisions in this challenging cancer population.

Beyond biological interpretation, the review addresses the technical complexities of quantitative image analysis. Different metrics introduce distinct trade-offs: while maximum standardized uptake value (SUVmax) is straightforward to extract, it remains vulnerable to statistical noise, whereas peak SUV (SUVpeak) depends heavily on region-of-interest definitions, and mean SUV (SUVmean) demands reproducible contouring. Target-to-background ratios introduce an additional layer of instability if antiangiogenic treatment alters the designated reference tissue. Volumetric metrics like FAPI-avid tumor volume also display high sensitivity to segmentation thresholds and edge blurring. Compounding these quantitative challenges, tracer chemistry strongly shapes observability. For example, a 2026 prospective comparison of covalent-targeted 68Ga-CTR-FAPI against 68Ga-FAPI-04 across 40 patients with radioiodine-refractory disease demonstrated superior lesion detection and uptake for the covalent compound, proving that quantitative response thresholds derived from one radiopharmaceutical cannot be applied to another.

To navigate this heterogeneous landscape, the authors structured their review using SANRA guidelines and established a four-tiered evidence hierarchy separating direct longitudinal studies from diagnostic series, non-thyroid trials, and mechanistic data. They intentionally narrowed their primary scope by excluding selective RET, NTRK, and BRAF/MEK inhibitors because those targeted therapies alter tumor cells and stroma through biological pathways distinct from VEGFR inhibition. The synthesis notes that pooled diagnostic evidence, including a 2026 systematic review summarizing eight studies involving 167 patients, cannot be repurposed to justify response assessment. Diagnostic cohorts often recruit patients with thyroglobulin-elevated negative iodine scintigraphy (TENIS), a setting that does not correspond clinically to the progressive, high-burden disease that warrants initiating systemic antiangiogenic therapy.

Finally, the review warns that proving incremental clinical value requires rigorous statistical modeling rather than simple univariable correlations between tracer percentage changes and progression-free survival. Future validation protocols must compare standard multivariable clinical models against extended models incorporating molecular imaging parameters to measure tangible gains in discrimination, calibration, and net benefit via decision-curve analysis. Investigators must also implement landmark or time-dependent analyses to safeguard against immortal-time and selection biases, ensuring that patients who experience early disease progression, death, or treatment discontinuation before scheduled imaging windows are not systematically excluded from evaluations.

Subject of Research: Evaluation of FAPI PET as a response biomarker in radioiodine-refractory differentiated thyroid cancer

Article Title: PI PET for response assessment during antiangiogenic TKI therapy in radioiodine-refractory differentiated thyroid cancer: evidence and research priorities

Article References: Niu, C., Yao, H., Lin, P., Chen, Z., & Wen, H. (2026). PI PET for response assessment during antiangiogenic TKI therapy in radioiodine-refractory differentiated thyroid cancer: evidence and research priorities. Frontiers in Endocrinology, 17, Article 1983089. https://doi.org/10.3389/fendo.2026.1983089

Image Credits: AI Generated

DOI: 10.3389/fendo.2026.1983089

Keywords: FAPI PET, Radioiodine-refractory differentiated thyroid cancer, Tyrosine kinase inhibitors, Lenvatinib, Fibroblast activation protein, Imaging biomarker, RECIST 1.1, Thyroglobulin, Tumor microenvironment, Progression-free survival, Clinical validation, Nuclear medicine, Thyroid cancer, Stromal remodeling, Diagnostic imaging