Covalent FAPI Imaging–Guided Precision Surgery in Patients with Medullary Thyroid Carcinoma: A Phase II Clinical Trial

Abstract

Purpose:

Medullary thyroid carcinoma (MTC) is curable only by complete resection of all malignant lesions; however, biochemical cure rates remain suboptimal because of imprecise lesion localization. We previously developed a covalent targeted radioligand fibroblast activation protein inhibitor (CTR-FAPI-30) with superior MTC detection rate and accuracy. This study evaluated whether [⁶⁸Ga]Ga-CTR-FAPI-30 positron emission tomography–computed tomography (PET-CT)–guided surgery improves patient outcomes.

Patients and Methods:

In this prospective, open-label phase II clinical trial, 50 patients with MTC were enrolled and underwent [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT–guided surgery. Patients were stratified into three predefined arms: (i) newly diagnosed MTC, R0 resection; (ii) recurrent MTC, R0 resection; and (iii) unresectable disease or distant metastasis. The primary endpoint was the biochemical cure rate at 1 month postoperatively. Secondary endpoints included event-free survival, the diagnostic accuracy of [⁶⁸Ga]Ga-CTR-FAPI-30, and surgical plan modification rate.

Results:

The biochemical cure rates were favorable under [⁶⁸Ga]Ga-CTR-FAPI-30–guided surgery, with 84.2% [95% confidence interval (CI), 60.4%–96.6%] in arm 1 (newly diagnosed, R0 resection) and 46.7% (95% CI, 21.3%–73.4%) in arm 2 (recurrent, R0 resection), both of which exceeded historical data (P = 0.007–0.049). For 231 lesions with gold-standard pathology, [⁶⁸Ga]Ga-CTR-FAPI-30 demonstrated superior diagnostic accuracy (96.5% vs. 72.7%, P < 0.0001), sensitivity (98.5% vs. 81.7%, P < 0.0001), and specificity (85.3% vs. 20.6%, P < 0.0001) compared with conventional imaging. Surgical plans were modified in 46% of patients based on [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT compared with investigator-determined approaches, with 91% of these modifications histopathologically justified.

Conclusions:

[⁶⁸Ga]Ga-CTR-FAPI-30–guided surgery achieved favorable biochemical cure rates for both newly diagnosed MTC and recurrent MTC, enabling precision surgical resection through accurate lesion localization.

Translational Relevance.

Surgical management of medullary thyroid carcinoma (MTC) remains the only curative option but is hindered by the suboptimal lesion localization provided by conventional imaging and serum calcitonin levels. Building on our previous development of a microenvironment-targeting covalent targeted radioligand (CTR), this phase II trial demonstrated that [⁶⁸Ga]Ga-CTR-FAPI-30 positron emission tomography–computed tomography (PET-CT)–guided surgery achieved a biochemical cure rate of 84.2% in newly diagnosed MTC and 46.7% in recurrent MTC reaching R0 resection. Importantly, surgical plans were modified in 46% of patients based on PET-CT findings, with 91% of these modifications being histopathologically justified. These findings redefine MTC surgery by enabling precise lesion localization, reducing unnecessary dissections, and improving curative outcomes. A subsequent randomized phase III trial of [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT–guided surgery in newly diagnosed MTC is currently ongoing to validate these results.

Introduction

Medullary thyroid carcinoma (MTC) is a malignant neoplasm originating from the calcitonin-producing parafollicular cells (C cells) of the thyroid, representing 1%–4% of thyroid malignancies yet accounting for 13.4% of thyroid cancer mortality (1, 2). Surgical resection remains the sole potentially curative intervention, in which biochemical cure, defined as postoperative serum calcitonin normalization, constitutes the most rigorous therapeutic benchmark and correlates with significantly improved survival (3–5). Nevertheless, large-scale studies indicate persistently unsatisfactory biochemical cure rates: 57.9% in newly diagnosed MTC and 16.2% to 17.9% in recurrent disease following curative surgery (6–8).

Current guidelines lack consensus with regard to the optimal surgical extent in MTC because of the inability for precise lesion localization. Conventional imaging–guided approaches frequently prove insufficient, as evidenced by occult nodal metastases in 36% of radiologically negative lateral compartments (9). Calcitonin-directed prophylactic neck dissection has consequently emerged as an alternative strategy, as lymph node metastasis may be present in the ipsilateral neck (>20 pg/mL), contralateral neck (>200 pg/mL), and mediastinum (>500 pg/mL; ref. 10). This paradigm is endorsed by both American Thyroid Association (ATA) and British Thyroid Association guidelines (11, 12). However, retrospective analyses demonstrate no survival benefit from prophylactic lateral dissection (13–15), fueling ongoing debate between imaging-based and biomarker-driven surgical approaches. Additional proposed strategies include prophylactic lateral neck dissection if tumor diameter >1 cm or central lymph node metastasis [endorsed by National Comprehensive Cancer Network (NCCN) guidelines; ref. 16], yet no standardized approach has achieved universal acceptance.

Fibroblast activation protein inhibitor (FAPI) represents a class of small-molecule radiopharmaceuticals targeting fibroblast activation protein (FAP), a transmembrane serine protease overexpressed in cancer-associated fibroblasts (17–21). Given the excellent co-localization between MTC and FAP in tumor stroma (22), [⁶⁸Ga]Ga-FAPI positron emission tomography–computed tomography (PET-CT) has demonstrated superior detection of metastatic MTC lesions that are overlooked by conventional imaging modalities and alternative PET tracers (23, 24). However, the clinical utility of FAPI is constrained by rapid systemic clearance, resulting in suboptimal tumor uptake and retention that compromises sensitivity for low volume or modestly FAP-expressing lesions (25, 26).

To address these limitations, we developed a platform technology incorporating sulfur(VI)–fluoride exchange (SuFEx) chemistry-based linkers into radiopharmaceuticals for tumor-selective covalent ligation (27). Integration of this SuFEx-enabled covalent targeted radioligand (CTR) moiety into FAPI achieved >80% covalent conjugation with FAP tyrosine residues. Preclinical evaluation revealed 257% greater tumor uptake and 13-fold prolonged retention compared with the original FAPI-04 in animal models (27). In clinical translation, the first generation [⁶⁸Ga]Ga-CTR-FAPI-02 demonstrated improved patient-based detection rate (98%), tumor uptake (SUV_max 11.71 ± 9.16), and diagnostic accuracy (96.7%) compared with [¹⁸F]fluoro-2-deoxy-D-glucose (FDG) PET-CT in MTC (28). We further developed a second-generation covalent agent (denoted as CTR-FAPI-30) through strategic molecular modifications to improve the tumor-to-background ratio.

This prospective clinical trial was therefore designed to evaluate whether the superior lesion localization offered by [⁶⁸Ga]Ga-CTR-FAPI-30 translates into improved surgical precision and therapeutic outcomes. By assessing its impact on surgical planning, oncological outcomes, and patient management, this study aims to provide clinical evidence for implementing CTR to improve surgical management and outcomes for patients with tumor.

Patients and Methods

Study design and participants

This was a prospective, single-center, open-labeled phase II investigator-initiated clinical trial designed to evaluate whether [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT–guided surgery improved clinical outcomes for patients with MTC (NCT06277180). The study was conducted in accordance with the principles outlined in the Declaration of Helsinki and was approved by the institutional review board (23/378-4120). All participants provided written informed consent. The clinical protocol and notification to attending physicians are included as supplementary data.

Inclusion criteria for the current study were: (i) age 18–75 years; (ii) newly diagnosed or recurrent MTC scheduled for surgical intervention within 1 month; (iii) elevated serum calcitonin level to enable postoperative biochemical assessment; (iv) ≥1 resectable lesion according to Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 criteria; (v) no significant abnormalities in blood routine, liver function, kidney function, or recently onset of infections; and (vi) willingness to participant and ability to sign the informed consent form. Candidates were excluded if they had contraindications to surgery (extensive distant metastases/unresectable disease) or declined either surgical management or the protocol-specified PET-CT–guided approach.

[⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT and interpretation

All patients underwent traditional imaging comprising neck ultrasound and contrast-enhanced CT (including neck, thorax, and upper abdomen) as standard of care. Contrast-enhanced CT/MRI of the liver, bone scan/MRI, or PET-CT with other tracers might be performed for selected cases or collected if previously conducted at other institutes.

[⁶⁸Ga]Ga-CTR-FAPI-30 was locally synthesized and radiolabeled. Following intravenous administration of 1.8 to 3.7 MBq/kg [⁶⁸Ga]Ga-CTR-FAPI-30, PET-CT acquisitions commenced approximately 60 minutes after injection using a uEXPLORER total-body PET-CT scanner (United Imaging Healthcare) or PoleStar m660 PET-CT scanner (SinoUnion Healthcare). Image voxels underwent decay correction and body weight normalization to yield standardized uptake value (SUV) maps.

Two nuclear medicine physicians independently interpreted the [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT scans. Readers were masked to patient information, other readers’ assessments, and findings from additional imaging modalities. Anonymized data comprising PET-CT scan, patient history, and the recent calcitonin level were provided.

Readers were required to identify and delineate all potential MTC lesions for quantitative analysis through a standardized protocol: (i) Normal organ uptake quantification was performed by placing spherical reference region of interest (ROI_ref) with a diameter of 0.5 cm (lymph node), 1 cm (thyroid, if not resected, and bone), or 2 cm (lung, liver, and aorta), from which maximum and mean SUV values (denoted as N_max and N_mean, respectively) were calculated. (ii) Lesion segmentation was executed semiautomatically using a threshold of SUV >2.0 × N_max, followed by manual refinement. Areas exhibiting visually discernible uptake above background but below this threshold were also segmented as positive, whereas regions exceeding the threshold but strongly considered as non-MTC (e.g., adrenal pheochromocytoma) were excluded. Visual and quantitative assessments of the [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT from each nuclear medicine physician, together with the original image, were provided for surgeon’s reference.

Surgery and pathologic validation

Surgical extent was defined by the [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT findings. For newly diagnosed MTC cases, concurrent level VI dissection was mandated during resection of the primary thyroid tumor in accordance with the ATA and NCCN guidelines, irrespective of the imaging result, calcitonin level, or tumor diameter (11, 16). Surgeons prospectively documented (i) the initial operative strategy based on conventional imaging and serum calcitonin levels (at our institute, prophylactic ipsilateral neck dissection is planned for newly diagnosed MTC with serum calcitonin above the cutoff, cT1b or more advanced primary tumors, or cN1a disease and prophylactic contralateral neck dissection is planned for an extremely high level of calcitonin or significant ipsilateral neck metastasis); (ii) the PET-CT–guided surgical approach, prioritizing compartment-oriented lymph node dissection over lesion-specific excision when indicated, consistent with oncological principles for thyroid malignancies; and (iii) modifications attributable to PET-CT findings, with dramatic change (i.e., sparing an entire lateral neck dissection or including upper mediastinum dissection via median sternotomy) considered a modification of surgical extent whereas minor modification (i.e., inclusion/exclusion of level IIB/VB with an already planned lateral neck dissection) was not. Resection completeness (eradication of all [⁶⁸Ga]Ga-CTR-FAPI-30–avid lesions or incomplete resection) was recorded immediately after surgery.

Lesions meeting predefined validation criteria were separately obtained for imaging diagnostic accuracy assessment. Inclusion required (i) suspicion of malignancy on ≥1 imaging modality [[⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT, or contrast-enhanced CT, or ultrasound, or PET-CT with [¹⁸F]FDG or [¹⁸F]fluorodihydroxyphenylalanine ([¹⁸F]FDOPA) or [⁶⁸Ga]Ga-somatostatin analogs ([⁶⁸Ga]Ga-SSA) if completed at other institutes; ref. 29] and (ii) availability of gold-standard verification through either histopathology (confirming MTC/non-MTC status) or clinical follow-up (specifically for unresected non-MTC lesions demonstrating postoperative biochemical cure). Lesions uniformly classified as benign across all imaging modalities were excluded from point-to-point matching because of inherent limitations in correlating imaging findings with intraoperative observations.

Safety evaluation and follow-up

Patients underwent adverse event monitoring during and for 2 hours following radiotracer administration. Pre- and postinjection vital signs (body temperature, heart rate, and blood pressure) were systematically documented. For delayed adverse event surveillance, participants could report symptoms via phone call report or outpatient clinical evaluation.

Surgical complications, including, but not limited to, recurrent laryngeal nerve dysfunction, permanent hypoparathyroidism, lymphatic leakage, and surgical wound infection, were documented by the referring physician. Each complication was classified with regard to its attribution to surgical plan modifications guided by [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT findings.

In addition to preoperative and 1-month postoperative serum biomarker assessment for primary endpoint evaluation, longitudinal monitoring incorporated both serum biomarkers [calcitonin and carcinoembryonic antigen (CEA)] and traditional imaging at 6, 12, and 24 months postoperatively. This surveillance protocol mandated neck ultrasound at all intervals, supplemented by contrast-enhanced CT, liver CT/MRI, or bone scan/MRI when clinically indicated to detect biochemical and structural evidence of disease recurrence.

Outcomes and predetermined investigative arms

The primary endpoint was the biochemical cure rate at 1 month after surgery, defined by normalization of serum calcitonin levels. This parameter represents the strictest curative level for MTC and correlates with survival outcomes (3–5). In a nation-wide retrospective analysis involving 863 patients, the 10-year overall survival (OS) of biochemically cured MTC was 97.7%; in comparison, the 10-year OS of noncured MTC was 70.3% (3). The timepoint of 1 month after surgery was selected in consistent with the standard protocol at our institute, taking into account the serum half-life of calcitonin and timely evaluation of hormone replacement.

Secondary endpoints included (i) event-free survival (EFS), defined as time from intervention to the earliest occurrence of structural recurrence in R0-resected regions, progression of residual disease, or new disease in nondissected regions due to negative PET-CT. Assessment of EFS incorporated neck ultrasound, structural imaging, and surgical resection status. Notably, this EFS definition is relatively strict as recurrent or progressive lesions may not meet RECIST criteria because of MTC’s frequent presentation as subcentimeter metastases. (ii) Progression-free survival (PFS), spanning from intervention to systemic structural disease progression per RECIST criteria. (iii) Diagnostic accuracy of imaging modalities for MTC lesion identification. (iv) Proportion of patients changed surgical strategy. (v) Safety of [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT–guided surgical intervention.

Given the distinct treatment expectations among patients with differing MTC status (newly diagnosed or recurrent disease) and varying resection extents (curative or palliative surgery), patients were stratified into three arms for separate analysis:

Arm 1: newly diagnosed MTC achieving R0 resection, with the goal of attaining biochemical cure and long-term survival.

Arm 2: recurrent MTC achieving R0 resection, with the goal of reducing serum calcitonin to low levels (in which biochemical cure may be attainable in a subset of population).

Arm 3: palliative surgery due to distant metastasis or unresectable disease.

Statistical analysis

Based on preliminary data of our [⁶⁸Ga]Ga-CTR-FAPI-02 imaging trial and historical data, a prospective statistical power analysis was conducted to determine the sample size for each study arm. For arm 1, enrolling 22 participants was calculated to provide 80% power to detect a difference of 27.2% in biochemical cure rates (85.1% vs. 57.9%) at one-sided α of 0.025 (6, 28). For arm 2, a sample size of 14 participants was estimated to provide the same power to detect a difference of 32.1% to 33.8% (50% vs. 16.2%–17.9%) biochemical cure rates (7, 8, 28). To account for a 10% rate of distant metastasis at diagnosis, approximately 15% of patients with unresectable disease, and an estimated 15% dropout rate, the final total sample size was set at 50 patients.

To assess the postoperative treatment response, the biochemical cure rate (proportion of patients achieving biochemical cure) and the postoperative-to-preoperative serum calcitonin ratio were calculated. The biochemical cure rate was statistically compared against historical data using the one-sample t test. Baseline comparison with historical data was conducted using Wilcoxon signed-rank, χ², or Fisher exact tests, as appropriate.

To evaluate the diagnostic accuracy of [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT versus conventional imaging for MTC identification, the McNemar test (two-sided) was applied with Clopper–Pearson–derived 95% confidence intervals (CI). Lesion uptake quantification employed three PET parameters: SUV_max (maximum SUV within ROI), SUV_mean (mean SUV within ROI), and T/N ratio (ratio of lesion SUV_max to the corresponding regional N_max). Intergroup comparisons of mean ± standard deviation (SD) values were performed via independent t tests.

All PET-CT data processing utilized 3D Slicer (version 4.11, RRID: SCR_005619) and Python (version 3.8.5, RRID: SCR_024202), whereas statistical analyses were implemented in R (version 4.2.0, RRID: SCR_001905).

Results

Baseline characteristics

Between October 2023 and June 2025, 86 patients with MTC were recommended for surgery and screened for eligibility, with 64 patients receiving [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT, and 50 patients enrolled in the current study (Fig. 1). The cohort comprised 29 newly diagnosed MTC and 21 recurrent MTC. The median calcitonin levels were 501.5 pg/mL (IQR, 227.8–2,224.5; range, 16.4–31296) for the whole population, 695 pg/mL (IQR, 335–3,570; range, 22.1–31296) for the newly diagnosed MTC, and 345 pg/mL (IQR, 157–584; range, 16.4–3,774) for the recurrent disease. Among newly diagnosed cases, 65.5% (19/29) were classified as high-grade MTC (30). Interreader agreements for lesions in the thyroid bed, ipsilateral neck, contralateral neck, mediastinum, and distant metastasis were good, with Cohen κ performances ranging from 0.834–0.947 (Supplementary Table S1). The median interval between [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT and surgery was 11.5 days (IQR, 6.3–17.5). Baseline characteristics are presented in Table 1.

Figure 1.

Study profile. Schematic illustrating patient allocation and study design. A cohort of 50 patients was enrolled and stratified into three arms.

Table 1.

Baseline characteristics of the study cohort.

Characteristic	Whole population (n = 50)	Subgroup
		Arm 1 (n = 19)	Arm 2 (n = 15)	Arm 3 (n = 16)
Age (median and IQR)	52.5 (IQR, 38.3–59.8)	58.0 (IQR, 48–60.5)	50 (IQR, 37–59)	45.5 (IQR, 32.8–56.5)
Sex	​	​	​	​
Male	54% (27)	20% (10)	16% (8)	18% (9)
Female	46% (23)	18% (9)	14% (7)	14% (7)
Origin	​	​	​	​
Newly diagnosed	58% (29)	38% (19)	0% (0)	20% (10)
Recurrent disease	42% (21)	0% (0)	30% (15)	12% (6)
T staging	​	​	​	​
Unmeasurable	40% (20)	0% (0)	28% (14)	12% (6)
T1	34% (17)	22% (11)	2% (1)	10% (5)
T2–4	26% (13)	16% (8)	0% (0)	10% (5)
N staging	​	​	​	​
N0	22% (11)	22% (11)	0% (0)	0% (0)
N1a	10% (5)	6% (3)	2% (1)	2% (1)
N1b	68% (34)	10% (5)	28% (14)	30% (15)
M staging	​	​	​	​
M0	88% (44)	38% (19)	30% (15)	20% (10)
M1	12% (6)	0% (0)	0% (0)	12% (6)
Preoperative calcitonin (median and IQR)	501.5 pg/mL (IQR, 227.8–2,224.5)	690 pg/mL (IQR, 229.5–2,467)	287 pg/mL (IQR, 122–442)	1,005.5 pg/mL (IQR, 455–4403.3)

Note: Data are presented as percentage (number) or median (IQR).

Biochemical cure rate in MTC achieving R0 resection

[⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT–guided surgery achieved favorable biochemical cure rates in patients reaching R0 resection (Table 2).

Table 2.

Treatment outcomes of each cohort.

Subgroup	Biochemical cure rate	6-month EFS
Arm 1	84.2% (60.4%–96.6%)	100% (76.8%–100%)
Arm 2	46.7% (21.3%–73.4%)	88.9% (51.8%–99.7%)
Arm 3	0% (0%–20.6%)	78.6% (49.2%–95.3%)

Note: Data are presented as rates with 95% CI, as appropriate.

In arm 1 (newly diagnosed, R0 resection), 19 patients (median calcitonin, 690 pg/mL; IQR, 229.5–2,467) received [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT–guided surgery and achieved a biochemical cure rate of 84.2% (95% CI, 60.4%–96.6%) at 1 month after surgery. The post-to-pre calcitonin ratios at 1 month were 0.036 ± 0.081 for the entire arm, 0.010 ± 0.035 in biochemically cured patients, and 0.173 ± 0.150 in uncured patients. Pathologic grading was similar in both whole arm (63.2% of high grade, 12/19) and biochemically cured patients (62.5% of high grade, 10/16; ref. 30). Among 14 evaluable patients at 6 months, calcitonin levels remained comparable with 1-month values (6-month to 1-month calcitonin ratio of 0.909 ± 0.983). No biochemical relapses occurred in patients with complete remission, and one additional patient attained biochemical cure by 6 months.

In arm 2 (recurrent, R0 resection), 15 patients (median calcitonin, 287 pg/mL; IQR, 122–442) achieved a biochemical cure rate of 46.7% (95% CI, 21.3%–73.4%) at 1 month after surgery with [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT guidance. The post-to-pre calcitonin ratios were 0.206 ± 0.262 for the whole arm, 0.017 ± 0.027 in biochemically cured patients, and 0.372 ± 0.265 in uncured patients. Follow-up calcitonin levels at 6 months (among nine patients reaching these timepoints) were comparable with 1-month values (6-month to 1-month calcitonin ratio of 1.037 ± 0.547). Biochemical cure status remained stable throughout follow-up.

In arm 3 (residual disease remained), none of the 16 patients (median calcitonin, 1,005.5 pg/mL; IQR, 455–4,403.3) achieved biochemical cure. The post-to-pre calcitonin ratios were 0.314 ± 0.303 at 1 month after surgery. Calcitonin gradually increased at 6-month follow-up (6-month to 1-month calcitonin ratio of 1.439 ± 1.623).

Comparison of biochemical cure rate with historical data

Although not controlled by historical data, the biochemical cure rates in arm 1 and arm 2 were further compared with historical data to evaluate whether [⁶⁸Ga]Ga-CTR-FAPI-30–guided surgery potentially improved the biochemical cure rate.

For the newly diagnosed R0-resected MTC (arm 1 comparison), a nation-wide multicentric retrospective study (time range, 1998–2024, 648 patients) reported a biochemical cure rate of 57.9% as per investigator-chosen surgical extent (6). The current arm 1 cohort showed comparable distributions in T staging, N staging, ratio of multifocal lesion, and preoperative calcitonin levels with the historical data (P = 0.120–0.948; Supplementary Table S2) but exhibited a higher biochemical cure rate (84.2% vs. 57.9%, P = 0.007; Fig. 2A).

Figure 2.

Performance of [⁶⁸Ga]Ga-CTR-FAPI-30 in MTC surgery. [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT–guided surgery showed higher biochemical cure rates in both (A) newly diagnosed and (B) recurrent MTC attaining R0 resection thanks to the better (C) diagnostic accuracy and (D) tumor uptake. Forty-six percent of patients modified surgical plan guided by [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT (E), with each column representing a single patient, and the resection extent, diagnostic status, treatment outcome, and surgical plan modification status are displayed. Data in A, B, and C are presented as value with 95% CI. Data in D are presented as mean ± SD.

For the recurrent R0-resected MTC (arm 2 comparison), a meta-analysis of 27 studies (time range, 1974–2012, 984 patients) reported a biochemical cure rate of 16.2% (7), and a recent single-center retrospective study (time range, 1994–2019, 366 patients) demonstrated a comparable rate of 17.9% (8). Arm 2 in the current study showed no differences in N staging and preoperative calcitonin levels compared with single-center retrospective historical data (P = 0.609–0.736, Supplementary Table S3; ref. 8) and had a higher biochemical cure rate (46.7% vs. 16.2%–17.9%, P = 0.038–0.049; Fig. 2B).

Sufficient EFS for R0-resected patients

Among 37 evaluable patients at 6-month follow-up, the cohort-wide EFS was 89.1% (95% CI, 74.6%–97%) under [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT–guided surgery. Patients achieving R0 resection (arm 1 and arm 2) exhibited 6-month EFS of 100% (95% CI, 76.8%–100%) and 88.9% (95% CI, 51.8%–99.7%), respectively (Table 2). In contrast, three patients with residual disease (arm 3) had structural recurrence/progression, yielding a 6-month EFS of 78.6% (95% CI, 49.2%–95.3%). No recurrent/progressive lesions met RECIST size criteria, and all patients continued disease monitoring.

Calcitonin served as a sensitive biomarker for disease progression, with the median ratio of 6-month to 1-month values of 2.803 ± 2.831 in progressing patients versus 0.943 ± 0.682 in nonprogressing patients. No biochemically cured patients developed structural disease during follow-up.

Higher diagnostic accuracy of [⁶⁸Ga]Ga-CTR-FAPI-30 than traditional imaging

A total of 231 lesions were considered malignant by one or more imaging modalities and were definitive for lesion property. Specifically, [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT identified 199 lesions as suspected malignancies, whereas conventional imaging (ultrasound + CT) suggested malignancy in 188 lesions and other imaging modalities identified 14 malignant lesions. Reference standard confirmation for the lesions was derived from pathologic diagnosis of MTC in 197 cases, pathologic diagnosis of non-MTC in 21 cases, and clinical follow-up confirmation of non-MTC in 13 cases (unresected lesions exhibiting biochemical cure, supporting benign classification).

[⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT demonstrated superior diagnostic accuracy versus conventional imaging modalities for MTC lesion identification (96.5% vs. 72.7%, P < 0.0001), with improvements in both sensitivity (98.5% vs. 81.7%, P < 0.0001) and specificity (85.3% vs. 20.6%, P < 0.0001; Fig. 2C; Supplementary Table S4). This advantage persisted across all subgroups when stratified by disease status (newly diagnosed vs. recurrent), anatomic region, and calcitonin level (Supplementary Tables S5–S10).

Differences in lesion uptake further supported the accurate classification of MTC and non-MTC lesions, as quantitative post hoc imaging analysis confirmed higher [⁶⁸Ga]Ga-CTR-FAPI-30 uptake in MTC lesions compared with non-MTC lesions (SUV_max, 11.19 ± 7.94 vs. 2.07 ± 2.50, P < 0.0001; Fig. 2D; Supplementary Table S11). Elevated tumor uptake remained consistent across disease status, anatomic region, and calcitonin level (Supplementary Table S12–S17).

Precision surgical planning guided by [⁶⁸Ga]Ga-CTR-FAPI-30

Thanks to the sufficient lesion localization and diagnostic accuracy, [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT modified surgical plans in 46% (23/50) of patients compared with investigator-proposed approaches, with surgical extent reduced in 34% (17/50) and extended in 12% (6/50) patients (Fig. 2E). When stratified by treatment arms, patients in arm 1 (newly diagnosed, R0 resection) exhibited the highest modification rate (15/19), whereas arm 2 (recurrent, R0 resection) showed moderate plan alterations (6/15). Conversely, minimal modifications occurred in arm 3 (residual disease remain, 2/16).

Postoperative histopathologic correlation and treatment outcome validation confirmed appropriate modification in 21 of 23 cases (91%), in which resected areas contained malignancy whereas spared regions showed no residual disease, and the treatment outcomes were improved. Discordant findings occurred in two of these patients: One underwent unwarranted surgical extension due to PET false positives (dissection of the contralateral neck which was traditional imaging–negative), and one received potentially insufficient resection (spared contralateral neck dissection but remained biochemically uncured and revealed no structurally evident disease during follow-up). In addition, one patient (no surgical extent modification) received inappropriate expanded surgery due to false-positive findings in both [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT and traditional imaging. Examples of surgical modification according to [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT are shown in Figs. 3 and 4; Supplementary Figs. S1 and S2.

Figure 3.

[⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT (A) enlarged the surgical extent of MTC compared with (B) investigator’s choice. A 36-year-old male, previously received total thyroidectomy and level VI lymph node (LN) dissection, currently had a serum calcitonin level of 132 pg/mL. C, [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT detected bilateral neck metastasis (right level II–V, left level IV); however, (D) traditional imaging (ultrasound and contrast-enhanced CT) missed metastasis at right level V and left level IV. Therefore, the surgical extent was enlarged from right neck dissection (levels II–IV, determined by traditional imaging extent and calcitonin level) to (E) bilateral neck dissection (right level II–V, left level II–IV). Postoperative pathology confirmed bilateral metastases, including left level IV and right level V lesions undetected by traditional imaging, and the patient reached biochemical cure at 1-month follow-up (serum calcitonin, 6.97 pg/mL).

Figure 4.

[68Ga]Ga-CTR-FAPI-30 PET-CT (A) reduced the surgical extent of MTC compared with (B) investigator’s choice. A 59-year-old female with primary thyroid tumor currently had a serum calcitonin level of 4,242 pg/mL. Both (C) [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT and (D) traditional imaging (ultrasound and contrast-enhanced CT) identified a right thyroid lesion without suspected lateral neck metastasis (LNM). The investigator originally chose total thyroidectomy, bilateral level VI dissection, and right level II–IV dissection considering the extremely high level of serum calcitonin; however, the surgical plan was reduced to (E) total thyroidectomy and bilateral level VI dissection as [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT confirmed the absence of lateral neck involvement. Biochemical cure was achieved at 1-month follow-up (serum calcitonin, 3.29 pg/mL), indicating that the reduced extent of surgery was feasible. Note: This case was presented as a surgical livestream at the 2024 National Cancer Center Head and Neck Oncology Symposium, Beijing, China.

To address the inherent subjectivity of investigator-driven planning, [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT–guided surgeries were further benchmarked against conventional imaging–based approaches, which extended surgical extent in 22% (11/50) and reduced in 10% (5/10) patients. When compared with the calcitonin-guided prophylactic dissection, [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT–guided surgeries demonstrated surgery extensions in 8% (4/50) and reductions in 46% (23/50) patients (Fig. 2E).

Explorative analysis

The study cohort was representative compared with historical data (Supplementary Table S18). Diagnostic performance of [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT for detecting distant metastasis was evaluated against other imaging modalities. Among the 50 enrolled patients, 8 patients received liver CT or MRI, 7 patients received [¹⁸F]FDG PET-CT, and 1 patient underwent [⁶⁸Ga]Ga-SSA PET-CT. Using treatment outcome and follow-up as the reference standard, liver CT/MRI yielded three true-positive, three true-negative, one false-positive, and one inconclusive results; [¹⁸F]FDG PET-CT reported one true-positive, five true-negative, and one false-negative results; [⁶⁸Ga]Ga-SSA PET-CT indicated one true-negative result. [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT accurately classified all of these patients in distant metastasis.

The role of preoperative calcitonin in MTC surgical planning was further elucidated. As anticipated, serum calcitonin levels demonstrated strong positive correlation with [⁶⁸Ga]Ga-CTR-FAPI-30–delineated tumor volume (Pearson correlation coefficient of 0.86); however, it showed no significant association with lymph node metastasis or distant metastasis status (Spearman correlation coefficients of 0.01–0.22). Therefore, calcitonin quantified disease burden but lack anatomic localization.

The potential of CEA as an alternative biomarker for MTC was also investigated. The median CEA level at enrollment was 13.3 ng/mL (IQR, 4.2–61.1) for the whole population, 19 ng/mL (IQR, 6.3–72.3) for patients with newly diagnosed MTC, and 7.9 ng/mL (IQR, 3.4–18.1) for those with recurrent disease. However, 14 patients had a CEA value within the normal range (<5 ng/mL) at initial diagnosis, making CEA-based cure rates impossible to calculate. CEA showed a moderate correlation with tumor volume delineated by [⁶⁸Ga]Ga-CTR-FAPI-30–delineated tumor volume (Pearson correlation coefficient of 0.62) and was not correlated with lymph node metastasis or distant metastasis status (Spearman correlation coefficients of −0.13 to 0.24).

In addition, the 29 newly diagnosed MTC can be divided into three distribution categories: (i) localized disease (primary tumor only, no lymph node or distant metastasis, n = 11); (ii) locoregional disease (primary tumor and lymph node metastasis, no distant metastasis, n = 13); and (iii) distant metastasis at initial diagnosis (n = 5). Although preoperative calcitonin levels failed to correlate with these categories (Spearman correlation coefficient of 0.30), [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT accurately stratified the patients into these groups.

Safety

No grade ≥2 adverse events occurred during or following [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT. Surgical complication among enrolled patients included 2% (1/50) recurrent laryngeal nerve dysfunction (tumor infiltration), 2% (1/50) phrenic nerve dysfunction (tumor infiltration), and 2% (1/50) lymphatic leakage (improved after conservative treatment); none of these correlated with the modified surgical plan according to [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT. Notably, [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT precluded the need for 2% (1/50) recurrent laryngeal nerve resection (tumor infiltration) and 2% (1/50) trachea window resection (tumor infiltration) by identifying distant metastases (rendering the disease ultimately incurable). This avoidance of unnecessary procedures contributed to the preservation of patient quality of life.

Discussion

In this prospective imaging-guided surgical trial, [⁶⁸Ga]Ga-CTR-FAPI-30–guided surgery exhibited biochemical cure rates of 84.2% (95% CI, 60.4%–96.6%) in newly diagnosed MTC and 46.7% (95% CI, 21.3%–73.4%) in recurrent MTC reaching R0 resection. This improvement is mechanistically supported by the tracer's superior lesion uptake and diagnostic accuracy, enabling comprehensive preoperative lesion mapping for surgical resection. Critically, [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT redirected surgical planning in 46% of cases compared with investigator-determined strategies, yielding better oncological outcomes through precise surgical resection.

The management of MTC has historically been constrained by limitations in lesion localization at both initial diagnosis and follow-up. Conventional imaging modalities exhibit insufficient sensitivity, failing to identify pathologically confirmed lymph node metastases in up to 36% of cases during surgical planning and missing distant metastases in 18% to 60% of patients (9, 31–34). Although guidelines from the NCCN and European Association of Nuclear Medicine recommend PET-CT with radiopharmaceuticals such as [¹⁸F]FDG, [¹⁸F]FDOPA, or [⁶⁸Ga]Ga-SSA for metastatic lesion detection (16, 29), metastatic tumors remain undetectable in 29% to 41% of patients (35–37). Consequently, their limited detection rates preclude reliable surgical extent planning.

As an alternative approach, calcitonin-directed prophylactic dissection has been employed when conventional and molecular imaging fail to delineate disease burden (11, 12). However, the threshold of calcitonin was significantly differed between studies (38–40), and retrospective studies suggested no survival benefit from prophylactic neck dissection (13–15). In our study, serum calcitonin levels correlated positively with overall disease volume (Pearson correlation coefficient of 0.86) but did not associate with regional or distant metastasis (Spearman correlation coefficient of 0.01–0.22). Thus, reliance on calcitonin for surgical guidance is inherently limited: It quantifies disease burden but lacks anatomic specificity, precluding precise lesion localization.

Functional imaging–guided surgery provided an alternative solution (41, 42). Given the excellent co-localization of MTC and FAP, targeting overexpressed FAP in tumor stroma represents a promising diagnostic alternative (43–45). Early-phase imaging studies demonstrate that [⁶⁸Ga]Ga-FAPI PET-CT achieves patient-based detection rates of 81% to 96% for metastatic MTC (23, 24). Nevertheless, conventional FAPIs exhibit transient target engagement because of reversible ligand binding and short in vivo half-lives and might fail to visualize small malignant lesions because of suboptimal tumor-to-background ratios. Our CTR strategy further increased the diagnostic performance by enabling irreversible covalent bonds between radioligand and FAP-expressing cells, which enhances the detection of occult lesions undetected by the established imaging modalities through increased tumor uptake (27, 28). Consequently, [⁶⁸Ga]Ga-CTR-FAPI-30 not only improves distant metastasis identification but also facilitates precise surgical planning. This functional imaging–guided surgery diverges from traditional approaches that rely on structural imaging or calcitonin-based determination of surgical extent. By reliably revealing latent metastatic foci through sustained tumor-specific uptake, CTR-FAPI-30 guided the resection of metastatic tumors while avoiding prophylactic dissection in tumor-free areas, thereby improving surgical precision and therapeutic outcomes.

The disease distribution delineated by [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT provides novel insights into MTC progression. Our findings support stratifying newly diagnosed MTC into three distinct categories: (i) localized disease confined to the primary tumor without lymph node or distant metastasis. These tumors exhibit indolent growth patterns and are associated with better OS (46). (ii) Locoregional disease involving the primary site and lymph node metastasis but no distant metastasis. Curative-intent resection remains achievable through comprehensive lesion identification. (iii) Distant metastasis at initial diagnosis, in which surgical intervention focuses on palliative objectives to prolong survival or control disease-related symptoms (47). Although preoperative calcitonin levels showed no correlation with these classifications (Spearman correlation coefficient of 0.30), [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT accurately stratified patients into these classifications and guided surgical resection.

The integration of [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT into surgical planning promises to redefine management of MTC: (i) Identification of occult malignant lesions undetected by conventional imaging, enabling complete tumor resection. In our cohort, 22% of patients exhibited additional metastases revealed exclusively by [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT, and excision of these lesions correlated with improved biochemical cure rates. (ii) Avoidance of calcitonin-directed prophylactic neck dissection in disease-free regions, thereby reducing operative morbidity. According to [⁶⁸Ga]Ga-CTR-FAPI-30–defined disease extent, 46% of patients in our study avoided prophylactic dissection without compromising therapeutic outcomes. It should be noted that concurrent level VI dissection was performed for all newly diagnosed patients during primary tumor resection as recommended by guidelines (11, 16) because the re-operation of level VI (if not dissected initially) increases the possibility of recurrent laryngeal nerve dysfunction and permeant hypoparathyroidism. (iii) Precise mapping of distant metastases to tailor surgical scope, maximizing resection of amenable tumor while preserving function (e.g., preserving tumor-infiltrated recurrent laryngeal nerves or trachea in cases with incurable distant metastasis). To validate these advantages, a multicenter randomized phase III trial has been initiated to assess whether [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT–guided surgery improves outcomes in newly diagnosed MTC (NCT07383246).

Beyond MTC, the covalent targeting strategy of CTR-FAPI-30 holds significant promise as a pancancer radiopharmaceutical. Its mechanism of action, forming a stable and irreversible ligation after selective FAP binding, minimizes off-target retention and maximizes tumor-specific uptake. This superior targeting profile not only enhances diagnostic sensitivity but is also ideally suited for targeted radionuclide therapy, as prolonged tumor retention permits delivery of a potent and localized therapeutic radiation dose. Consequently, we have initiated a phase I dose-escalation study of [¹⁷⁷Lu]Lu-CTR-FAPI-30 for the treatment of metastatic thyroid cancers (NCT07438847), seeking to provide a novel therapeutic option in which current targeted therapies have failed to improve OS (48–51).

Several limitations of the current study should be noted. First, the initial postoperative serum calcitonin was measured at 1 month, earlier than the guideline recommendations. Although levels remained stable when compared with 6-month values in our cohort, this earlier timepoint may introduce bias when comparing biochemical cure rates with other studies. Additionally, long-term follow-up is needed to determine whether biochemical cure translates into sustained EFS. Second, although comparisons were made with historical data, the study was not controlled because of differences in timeframe, surgical principles, techniques, and calcitonin measurement. The future randomized phase III trial is necessary to validate the study findings. Third, patients with very low or very high calcitonin levels were not enrolled, as the former population may not benefit from the PET-CT imaging and the latter population might have widespread distant metastasis that precludes surgical resection. Finally, the accuracy analysis excluded lesions classified as benign across all imaging modalities, which may underestimate the specificity of imaging modalities.

In conclusion, [⁶⁸Ga]Ga-CTR-FAPI-30–guided surgery achieved favorable biochemical cure rates in both newly diagnosed MTC and recurrent MTC, enabling precision surgical resection through accurate lesion localization. Coupled with established diagnostic advantages over conventional imaging modalities in metastatic MTC, [⁶⁸Ga]Ga-CTR-FAPI-30 PET-CT represents a promising candidate for standard-of-care imaging at both initial diagnosis and follow-up.

Supplementary Material

Supplementary Data

This file contains Supplementary Tables 1-18, Supplementary Figure 1-2, clinical protocol and notification to attending physicians.

Data Availability

The data generated in this trial are not publicly available at present because of restrictions related to patient privacy; however, these data are available from the corresponding authors upon reasonable request.

Authors’ Disclosures

X.-Y. Cui reports a patent for PCT/CN2023/096106 pending, licensed, and with royalties paid from BoomRay Pharmaceuticals. Z. Liu reports personal fees from BoomRay Pharmaceuticals outside the submitted work, as well as a patent for PCT/CN2023/096106 pending, licensed, and with royalties paid from BoomRay Pharmaceuticals. No disclosures were reported by the other authors.

Authors’ Contributions

Z. Kong: Resources, data curation, formal analysis, funding acquisition, writing–original draft, writing–review and editing. Y. Liu: Resources, data curation, formal analysis, writing–original draft, writing–review and editing. Y.-S. Lin: Conceptualization, resources, data curation, funding acquisition, investigation, writing–original draft, writing–review and editing. Z. Li: Resources, funding acquisition, validation, investigation, writing–original draft, writing–review and editing. X.-Y. Cui: Resources, formal analysis, investigation, writing–original draft, writing–review and editing. Shengyan Liu: Data curation, investigation, writing–review and editing. X. Zhang: Data curation, validation, writing–review and editing. R. Li: Data curation, investigation, writing–review and editing. Y. Yang: Data curation, formal analysis, writing–review and editing. Y. Sun: Data curation, formal analysis, writing–review and editing. Y. Nie: Data curation, formal analysis, writing–review and editing. Z. Zhang: Data curation, writing–review and editing. C. An: Data curation, writing–review and editing. S. Ni: Data curation, writing–review and editing. Y. Zhu: Data curation, formal analysis, writing–review and editing. Z. Liu: Conceptualization, formal analysis, funding acquisition, writing–original draft, writing–review and editing. J. Wang: Conceptualization, data curation, formal analysis, investigation, writing–original draft, writing–review and editing. Shaoyan Liu: Conceptualization, resources, data curation, funding acquisition, validation, writing–original draft, writing–review and editing.