Abstract
Objectives
Thyroglobulin-elevated negative iodine scintigraphy (TENIS) syndrome presents a significant diagnostic challenge in the management of differentiated thyroid carcinoma (DTC). Although 18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) is the established imaging modality for detecting presence of structural disease, Gallium-68 (68Ga)-fibroblast activation protein inhibitor (FAPI) PET/CT, which targets the tumor stroma, represents a promising new alternative. This study directly compared the lesion detection capabilities of both tracers in patients with TENIS.
Methods
In this prospective single-center study (2021-2024), 22 patients with TENIS syndrome underwent both 18F-FDG and 68Ga-FAPI PET/CT scans within a 7-day interval of each other. Lesion detectability and maximum standardized uptake values (SUVmax) were compared between the two modalities. Lesions were validated using histopathology or cytopathology, biochemical response to empirical 131I therapy, imaging correlation or progression, and clinical progression methods.
Results
Among 22 patients (median age 43.5 years, 90% with papillary carcinoma), 18F-FDG was positive in 23/33 cervical lymph nodes, whereas 68Ga-FAPI was positive in 14/33 lymph nodes. Fourteen lymphnodes were positive on both modalities with comparable SUVmax (median 18F-FDG 3.3 vs. 68Ga-FAPI 3.6, p=0.689). All local recurrences in thyroid bed (n=4) were 18F-FDG-positive, whereas only 3 of the local recurrences were 68Ga-FAPI-positive. Distant metastases (lung and bone) showed similar detection rates.
Conclusion
18F-FDG PET/CT demonstrates superior lesion detectability, particularly for cervical lymph nodes, in TENIS syndrome. 68Ga-FAPI PET/CT provides complementary information with comparable semiquantitative uptake in detected lesions but with lower overall detection. 18F-FDG PET/CT remains the preferred imaging modality for comprehensive disease evaluation in patients with radioactive iodine-refractory DTC.
Introduction
Thyroid cancer (TC) is the most prevalent endocrine malignancy, accounting for nearly 95% of endocrine neoplasms worldwide (1). Most cases are classified as differentiated thyroid carcinoma (DTC), including papillary and follicular subtypes (2). The standard management of DTC consists of total or near-total thyroidectomy, often followed by radioactive iodine (RAI) ablation and thyroid-stimulating hormone (TSH) suppression in intermediate- or high-risk patients, as recommended by the American Thyroid Association (ATA) guidelines (3, 4). With appropriate management, patients with DTC frequently achieve a favorable prognosis, with reported 10-year disease-specific survival rates approaching 98% (5, 6).
Nevertheless, a subset of patients encounters persistent or recurrent disease, manifested biochemically by elevated serum thyroglobulin (Tg ≥1 ng/mL while suppressed or ≥ 10 ng/mL after stimulation), with or without positive anti- Tg antibodies, even in the absence of structural evidence of disease (6, 7). Approximately 20% of these patients eventually develop structural recurrence, which is associated with a worse prognosis (6). Moreover, between 5% and 15% of DTC and up to 50% of metastatic lesions evolve into radioiodine-refractory (RAIR) disease, characterized by loss of radioiodine avidity and more aggressive clinical behavior (8, 9).
A clinically significant manifestation of RAIR-DTC is thyroglobulin-elevated, negative iodine scintigraphy (TENIS) syndrome, which is encountered in up to 27% of post-treatment DTC cases (10). Patients with TENIS have elevated Tg levels despite negative radioiodine whole-body scans, implying iodine-non-avid, active disease (11). This scenario presents a diagnostic and management challenge because conventional radioiodine imaging fails to localize the disease (10, 11).
The ATA guidelines recommend 2-18F-Fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) for patients with TENIS to localize metabolically active sites of recurrence/metastasis (3, 4). The underlying rationale is the “flip-flop phenomenon,” i.e., the inverse relationship between iodine and glucose avidity in TC, in which dedifferentiated tumour cells lose iodine transport capacity but acquire increased glucose metabolism, rendering them detectable by 18F-FDG PET/CT (12, 13). In well-differentiated, iodine-avid tumors, 18F-FDG uptake is low, but it increases in dedifferentiated and aggressive lesions (12, 14). The sensitivity of 18F-FDG PET/CT in TENIS has been reported to be 68.8-87%, with false-negative rates of 8-21% depending on the disease burden and metabolic characteristics (12, 15).
Recent advances have introduced fibroblast activation protein inhibitor (FAPI) labelled radiotracers that target cancer-associated fibroblasts abundant in the tumor stroma. Multiple FAPI compounds, such as FAPI-02, FAPI-04, and FAPI-46, are available, with Gallium-68 (68Ga) labelling providing compatibility with clinical PET imaging (16). 68Ga-FAPI PET/CT can detect lesions, irrespective of cellular differentiation or iodine avidity (16, 17). Importantly, stroma and FAP expression are major features of papillary and follicular thyroid carcinomas, particularly in dedifferentiated or aggressive disease, providing a rationale for 68Ga-FAPI PET application in this setting (18, 19).
While several studies have demonstrated strong 68Ga-FAPI uptake in metastatic or thyroid bed recurrent RAIR-DTC lesions, others have suggested variable results, especially in well-differentiated or low-stroma tumors (20). A recent head-to-head comparison in RAIR-DTC showed that 68Ga-FAPI detected additional lesions in approximately 20% of cases missed by 18F-FDG, with higher target-to-background ratios (21). However, comprehensive comparative data on TENIS syndrome remain limited. Therefore, the comparison of 68Ga-FAPI and 18F-FDG PET/CT in TENIS remains an active area of clinical investigation.
Materials and Methods
Study Design and Patient Selection
This prospective, single-center study was conducted at a tertiary referral university-affiliated teaching hospital between 2021 and 2024. The study protocol was approved by the Amrita Institute of Medical Sciences and Research Institutional Clinical Research Ethics Committee, and written informed consent was obtained from all participating patients. This study conformed to the ethical guidelines of the 2013 revision of the Declaration of Helsinki.
Eligible patients were adults aged ≥18 years with histopathologically confirmed DTC who had previously undergone total or near-total thyroidectomy, received adjuvant 131I ablation, and subsequently developed TENIS syndrome. The inclusion criteria required documented elevated serum Tg levels (stimulated Tg ≥10 ng/ml or further rising trend of stimulated Tg level in patients with previously documented stimulated Tg ≥2 ng/ml), and a negative diagnostic 131I scan. The exclusion criteria were the presence of a second primary malignancy or refusal to provide informed consent.
PET/CT Acquisition Protocol
All PET/CT examinations were performed using a Siemens Biograph Horizon PET/CT system (Siemens Healthineers, Germany) equipped with lutetium oxyorthosilicate detectors and time-of-flight (TOF) reconstruction. Each patient underwent both 18F-FDG and 68Ga-FAPI PET/CT within seven days (median interval: 2 days). Patients fasted for at least six hours before 18F-FDG PET/CT. Blood glucose was measured prior to radiotracer injection and was required to be <150 mg/dl. Intravenous 18F-FDG was administered at a dose of 3.7 MBq/kg , followed by whole-body images (skull base to mid-thigh) acquired 45-60 min post-injection, with one minute per bed position. For FAPI PET/CT, the patients received an intravenous injection of 1.8-2.2 MBq/kg 68Ga-FAPI-04. Imaging was initiated 30-45 min post-injection, using the same anatomical range but with a two-minute acquisition per bed position. A low-dose, non-contrast CT scan (130 kV, 110-130 mAs, 3 mm slice thickness) was acquired for attenuation correction and anatomical localization, followed by a three-dimensional PET acquisition. Images were reconstructed using the ordered-subsets expectation-maximization + TOF algorithm (3 iterations × 10 subsets) with CT-based attenuation and scatter correction. The data were reviewed using the syngo MI Workplace software suite (Siemens Healthineers).
Image Analysis
The images were independently reviewed by two board-certified, experienced nuclear medicine physicians blinded to the clinical and laboratory information and each other’s interpretations. Discrepancies were resolved through consensus. Lesions were considered positive on PET if they showed focal 18F-FDG or 68Ga-FAPI uptake exceeding the background without a physiological explanation. Semi-quantitative analysis involved manual delineation of regions of interest around each lesion to obtain the maximum standardized uptake value (SUVmax). Lesions on CT without PET positivity were considered pathological if they fulfilled CT criteria for disease involvement.
Lesion Confirmation and Validation
We employed a hierarchical multimodal approach for lesion validation. Whenever feasible, lesions were confirmed by surgical excision with histopathological examination, or by fine-needle aspiration cytology. This represents the gold standard for confirmation. For lesions that could not be biopsied, the biochemical response to empirical 131I therapy was used as a surrogate marker. Decreases in stimulated Tg levels following therapy are considered supportive evidence of both disease presence and treatment response. Recurrent lesions in the thyroid bed were corroborated by correlative ultrasound imaging, which demonstrated the characteristic features of recurrent TC. Imaging features suggestive of metastatic disease were used to identify distant lesions. Serial follow-up imaging (PET/CT, CT, or magnetic resonance imaging) demonstrating interval growth or increased metabolic activity of the lesions was also considered confirmatory evidence. Clinical deterioration with new or worsening symptoms (e.g., bone pain and respiratory symptoms), corroborated by imaging findings, was used to validate lesion significance. This multimodal confirmation strategy provided validation despite the practical limitations of performing histopathological sampling in all patients with advanced diseases.
Data Collection
Clinical and laboratory data were systematically collected from electronic hospital medical records and included: patient demographics (age and sex), histopathological subtype and grade, initial surgical treatment, number of 131I therapy cycles, cumulative radioiodine dose, and serial stimulated serum Tg and anti-Tg antibody (TgAb) measurements (measured at initial diagnosis, at TENIS detection, and post-therapy). The imaging data included lesion location, lesion size on anatomical CT, and SUVmax on both 18F-FDG and 68Ga-FAPI PET/CT. Follow-up data included subsequent management (surgery, empirical radioiodine therapy), histopathological confirmation results, serial imaging findings, biochemical response assessments, and clinical outcomes.
Statistical Analysis
Continuous variables are presented as medians (ranges), and categorical variables are presented as counts (percentages). Paired comparisons of SUVmax in lesions positive on both modalities were performed using the Wilcoxon signed-rank test. Statistical significance was set at p<0.05.
Results
Patient Characteristics
This study included 22 TENIS patients. The median age was 43.5 years (range, 21-79 years), and 63% were female. Most cases (90%) were identified as papillary thyroid carcinoma, while follicular and poorly differentiated/microcarcinoma accounted for the remaining 10%. Most patients (59%) underwent total thyroidectomy, while 37% also underwent lymph node dissection. Eleven patients (50%) received one cycle of 131I therapy, and the remaining 11 patients (50%) received two cycles; the median cumulative radioiodine dose was 118.5 mCi (range, 60-215 mCi). The initial median Tg and anti-TgAb levels were 50.48 ng/ml and 28.07 IU/ml, respectively. Most patients (86%) had Tg levels above 10 ng/mL, and most were at intermediate risk on initial assessment. Patient characteristics are shown in Table 1.
Lesion Confirmation
Lesion validation followed a hierarchical multimodal approach. The validation methods by lesion type are summarized in Table 2. Cervical lymph node involvement was confirmed by histopathology or cytopathology in 75% (9 of 12 patients), providing high confidence in nodal disease findings. Local recurrences and distant metastatic lesions were confirmed primarily by biochemical response, imaging findings, and clinical progression, reflecting the practical challenges of invasive sampling at these anatomical sites.
Local thyroid bed recurrent lesions
CT imaging identified local recurrence in 4 of 22 patients. In all four patients, the lesions were 18F-FDG-avid, whereas in three patients the lesions were 68Ga-FAPI-positive (Figure 1). One 18F-FDG-avid lesion (SUVmax 3.2) was negative on 68Ga-FAPI PET/CT imaging. The median 18F-FDG SUVmax for local recurrence was 4.45 (range, 3.2-6.8), and the median 68Ga-FAPI SUVmax was 2.9 (range, 2.4-10.5). The SUV metrics are presented in Table 3.
Lymph Node Metastases
Among the 33 lymph nodes identified on CT imaging, 18F-FDG was positive in 23 nodes (69.7%), whereas 68Ga-FAPI was positive in 14 nodes (42.4%) (Figure 2). Fourteen nodes were positive on both modalities and analyzed in paired comparisons of SUVmax. The median SUVmax was 3.3 (range 2.2-16.8) for 18F-FDG and 3.6 (range 2.2-9.8) for 68Ga-FAPI (Figure 3). The Wilcoxon signed-rank test showed no significant difference in SUVmax between 18F-FDG and 68Ga-FAPI in paired lymph-nodal lesions (p=0.689). 18F-FDG PET/CT detected more lymph nodes overall than 68Ga-FAPI PET/CT. The detailed metrics are presented in Table 4.
Distant Metastases
Five pulmonary lesions were identified on CT. 18F-FDG was positive in all five lesions, whereas 68Ga-FAPI was positive in four lesions (Figure 4). The median SUVmax was 3.3 for 18F-FDG and 5.05 for 68Ga-FAPI, respectively. Mean SUVmax was 5.72±5.75 for 18F-FDG and 5.95±4.13 for 68Ga-FAPI. Two bone lesions were identified, both detected by18F-FDG and 68Ga-FAPI PET/CT. The median SUVmax values were 4.75 for 18F-FDG and 3.5 for 68Ga-FAPI. The mean SUVmax was 4.75±0.64 for 18F-FDG and 3.5±0.42 for 68Ga-FAPI. The Wilcoxon signed-rank test comparing paired lung and bone lesions (n=6) showed no statistically significant difference in SUVmax (p=1.0). The detailed metrics are presented in Table 5.
Discussion
This prospective comparative study of 18F-FDG and 68Ga-FAPI PET/CT in a representative TENIS cohort demonstrated important differences in tracer performance, especially in nodal detection. 18F-FDG was positive in a significantly higher number of lymph node metastases than 68Ga-FAPI (23 vs. 14 nodes, 69.7% vs. 42.4% detection rate), reaffirming its established role as the preferred modality for the evaluation of TENIS (2, 3, 4). These findings are consistent with large meta-analyses and practice guidelines emphasizing the sensitivity of 18F-FDG PET/CT for metabolically active, dedifferentiated recurrent disease (5, 6).
Clinical Significance of Nodal Detection in Papillary Thyroid Carcinoma
Papillary thyroid carcinoma characteristically metastasizes to the cervical lymph nodes, with nodal involvement occurring in 30-80% of cases, depending on tumor size and histologic features (22, 23). Accurate nodal detection is critical for staging, prognostication, and treatment planning. In this context, the superior nodal detectability of 18F-FDG PET/CT (23 nodes detected vs. 14 by 68Ga-FAPI) represents a clinically significant advantage that directly impacts patient management decisions regarding surgical intervention, empirical radioiodine therapy, or watchful waiting (24, 25).
Lesion Confirmation and Validation
The multimodal lesion confirmation approach employed in this study reflects the real-world challenges of validating imaging findings in patients with advanced TC. While 75% (9 of 12) of cervical lymph node cases were confirmed by histopathology or cytopathology, which provided robust validation, the remaining nodal lesions and most distant lesions relied on biochemical response, imaging progression, or clinical correlation. This hierarchical validation strategy, although pragmatic and clinically appropriate, limits the ability to definitively establish false-negative rates. The high proportion of histopathologically confirmed cervical nodes strengthens confidence in our finding of superior nodal detectability with 18F-FDG, as these represent the most rigorously validated lesions in our cohort. For distant lesions, reliance on imaging progression and clinical correlation may introduce uncertainty but remains consistent with accepted clinical practice in managing advanced DTC, where biopsy is often impractical (26, 27).
Comparative Performance of 18F-FDG and 68Ga-FAPI PET/CT
68Ga-FAPI PET/CT, targeting fibroblast activation in the tumor microenvironment, demonstrated comparable SUVmax values to those of 18F-FDG in paired local, nodal, and distant lesions, but failed to detect several cervical lymph nodes with metastases. This is consistent with recent reports highlighting that, while 68Ga-FAPI excels in certain stroma-rich or desmoplastic lesions (especially at the local recurrence site or in some distant metastases), it is less sensitive for varied nodal disease burden (16, 18, 28). The mechanism underlying this differential performance is likely related to heterogeneity in the composition of the tumor microenvironment. Lymph node metastases may have lower stromal FAP expression relative to their metabolic activity, which favors 18F-FDG detection (29, 30).
In our analysis, distant lesions in the lungs and bones were detected by both tracers at similar detection rates. One pulmonary lesion was negative for 68Ga-FAPI. The SUVmax values showed no significant differences (p=1.0). However, these two tracers may have different strengths. 68Ga-FAPI can outperform 18F-FDG in fibrotic or cancer-associated fibroblast-rich metastatic foci, as shown in recent clinical trials (21); however, it is not uniformly superior across all lesion types. The occasional higher 68Ga-FAPI SUVmax in some locally recurrent lesions (maximum 10.5 vs. 6.8 for 18F-FDG) suggests that 68Ga-FAPI may identify lesions with prominent stromal reactions that could be underestimated by FDG alone (31, 32).
Semiquantitative Analysis
The lack of a significant difference in uptake intensity (SUVmax) between 18F-FDG and 68Ga-FAPI in paired lesions (p=0.689 for lymph nodes, p=1.0 for distant metastases) suggests that both tracers detect lesions and provide a comparable semiquantitative assessment. This finding indicates that 68Ga-FAPI can be reliably used to characterize lesions. However, the clinical impact of missing lesions—particularly the nine additional lymph nodes detected only by18F-FDG (27% of all nodes) in our cohort—underscores the greater utility of 18F-FDG for comprehensive restaging and treatment planning. Missed nodal disease can lead to understaging and inadequate treatment selection, potentially affecting long-term outcomes (33, 34).
Comparison with Existing Literature
Our findings align with and extend those of recent comparative studies of 18F-FDG and 68Ga-FAPI PET/CT in TC. A recent study by Alevroudis et al. (21) reported that 68Ga-FAPI detected additional lesions in approximately 20% of patients with RAIR-DTC with higher target-to-background ratios. However, that study included a broader RAIR-DTC population rather than focusing specifically on TENIS syndrome. Our study specifically addresses the TENIS population and demonstrates that, while 68Ga-FAPI may offer complementary information in select cases, 18F-FDG provides superior overall lesion detection, particularly for nodal disease.
Clinical Implications and Future Directions
Based on our findings, 18F-FDG PET/CT should remain the first-line imaging modality for TENIS syndrome evaluation. 68Ga-FAPI PET/CT may serve as a complementary tool in specific scenarios, including when 18F-FDG PET/CT is negative or equivocal despite high Tg levels; for evaluating local recurrence where stromal reaction may be prominent; in patients with poorly differentiated or dedifferentiated TC with suspected high stromal content; and for treatment planning when targeting stromal components with novel therapies.
Future research should focus on larger multicenter prospective studies to validate these findings and to establish clear clinical algorithms; to assess the correlation of 68Ga-FAPI uptake with histopathologic FAP expression and stromal density; to explore 68Ga-FAPI PET/CT for response assessment following targeted therapies; and to develop radiomic and artificial intelligence approaches that integrate multi-tracer PET data for improved diagnostic accuracy and prognostication.
Study Limitations
This study has a few limitations. First, the modest sample size (n=22) from a single center limits the generalizability and statistical power of the subgroup analyses. Technical factors, including different acquisition protocols for 18F-FDG and 68Ga-FAPI (1 vs. 2 min per bed position), may influence image quality and lesion detection, although these protocols reflect standard clinical practice for each tracer. The study was also limited by the heterogeneity of the validation methods. Finally, we did not perform a detailed analysis of the target-to-background ratios or assess the impact of lesion size on detectability, which could provide additional insights into the relative tracer performance.
Conclusion
18F-FDG PET/CT remains the reference imaging modality for detecting structural disease in TENIS syndrome, while 68Ga-FAPI PET/CT is a promising, currently supplementary tool when conventional imaging is inconclusive. The superior nodal detectability of 18F-FDG PET/CT has direct clinical implications for the accurate staging and treatment planning of papillary thyroid carcinoma, the predominant histologic type in TENIS syndrome. The integration of both modalities, potentially in a dual-tracer protocol, could be explored to further enhance diagnostic yield and therapeutic guidance in challenging DTC cases.
