Theranostic Approaches to Radioiodine-Refractory Differentiated Thyroid Cancer: A Narrative Review

Simple Summary

Differentiated thyroid cancer is typically managed with surgery and radioiodine, a treatment that exploits the ability of thyroid cells to take up iodine. In a subset of patients, however, the disease loses this capacity and progresses despite repeated therapy, a condition known as radioiodine-refractory disease. Current options for these patients include local and systemic treatments, which can delay progression but are not curative and carry cumulative toxicity. This review examines the role of theranostics, an approach that combines molecular imaging with matched radionuclide therapy to individualize treatment. Particular attention is given to redifferentiation strategies, in which targeted drugs are used to restore iodine uptake and allow renewed radioiodine therapy. Emerging non-iodine theranostic approaches are also discussed. The aim is to clarify patient selection and inform the integration of these strategies into clinical practice.

Abstract

Background: Radioiodine (Na[¹³¹I]I) therapy is the cornerstone of systemic treatment for differentiated thyroid cancer (DTC), exploiting sodium–iodide symporter (NIS) expression for durable control. Up to 30–40% of advanced cases develop radioiodine-refractory disease (RAI-R DTC), marked by impaired iodine uptake, aggressive behavior, and poor response to Na[¹³¹I]I. Locoregional treatments, multikinase inhibitors (MKIs), and selective targeted agents improve progression-free survival but are not curative and carry cumulative toxicity, motivating precision-based alternatives. The primary objective of this review is to clarify the evolving theranostic paradigm in RAI-R DTC; the secondary objectives are to appraise redifferentiation and iodine-based theranostics for restoring or exploiting iodine avidity and to evaluate non-iodine theranostic strategies for cases where iodine biology is absent, impaired, or unreliable. Methods: This narrative review synthesizes contemporary evidence on theranostic strategies in RAI-R DTC, drawn from available studies, clinical trials, and current guidelines, with an emphasis on redifferentiation and non-iodine approaches; a systematic search protocol was not applied. Results: Theranostics couples target-specific molecular imaging with matched radionuclide therapy and response-adapted sequencing. Its most transformative application is redifferentiation, in which pharmacologic modulation of oncogenic signaling can restore iodine avidity and enable renewed, dosimetry-guided Na[¹³¹I]I treatment. Beyond iodine, somatostatin receptor (SSTR) imaging and peptide receptor radionuclide therapy (PRRT) have re-emerged in very selected cases, whereas alpha emitters remain investigational. Refractoriness is increasingly viewed as a reversible continuum rather than a fixed state. Conclusions: Theranostics can individualize RAI-R DTC treatment, restoring or exploiting iodine biology where possible and shifting to non-iodine targets where it is unreliable. Patient selection, timing, and integration with systemic therapy are central, and prospective validation is needed.

1. Introduction

Radioiodine therapy represents the earliest truly successful theranostic story, with diagnostic iodine imaging identifying sodium–iodide symporter (NIS)-expressing disease and compounded therapeutic Na[¹³¹I] delivering cytotoxic beta radiation [1,2]. Although differentiated thyroid cancer (DTC) overall carries an excellent prognosis, with 10-year disease-specific survival exceeding 90%, the subset that becomes RAI-R behaves very differently. Distant metastases develop in roughly 10% of DTC patients, and approximately two-thirds of these ultimately become radioactive iodine (RAI)-refractory (RAI-R). Once refractoriness is established, 10-year overall survival falls to approximately 10%, placing RAI-R DTC among the more lethal endocrine malignancies and justifying a distinct clinical and research focus [3,4]. For these patients, survival is driven less by thyroid lineage biology and more by oncogenic signaling, dedifferentiation, and the emergence of alternative vulnerabilities. Accordingly, a multimodal treatment strategy should be tailored to disease burden, growth kinetics, and symptomatology [5]. Locoregional therapies—including surgery, external beam radiotherapy, thermal ablation, and embolization techniques—remain central for controlling oligometastatic or symptomatic lesions and for preventing local complications. Systemic treatment is indicated for patients with progressive, unresectable, or symptomatic disease [6]. Multikinase inhibitors targeting angiogenic and oncogenic pathways represent the standard first-line systemic option (e.g., Sorafenib, Lenvatinib), providing meaningful disease control but often at the cost of chronic toxicity. More recently, selective targeted therapies directed at actionable genomic alterations (e.g., BRAF, RET, or NTRK fusions) have demonstrated higher response rates and improved tolerability, reshaping treatment algorithms in molecularly selected patients [7,8,9,10].

Despite these advances, systemic therapies are largely non-curative and require continuous administration, highlighting the need for careful patient selection and timing of treatment initiation [11,12,13]. As a result, active surveillance remains appropriate for indolent, asymptomatic disease, while emerging strategies, including redifferentiation and non-iodine theranostic approaches, are increasingly being explored to complement or defer long-term systemic therapy. Accordingly, the main aims of this review are to discuss the evolving theranostic paradigm in RAI-R DTC and to examine its two complementary objectives: (a) restoring or exploiting iodine avidity through redifferentiation strategies and iodine-based theranostics and (b) shifting toward non-iodine theranostic approaches when iodine biology is absent, impaired, or clinically unreliable.

2. Methodology

This article is a narrative review. We searched PubMed/MEDLINE for English-language publications up to April 2026, combining terms for radioiodine-refractory differentiated thyroid cancer, theranostics, redifferentiation, sodium–iodide symporter, peptide receptor radionuclide therapy, and radioligand strategies. Reference lists of key articles and recent society guidelines were screened for additional sources. Priority was given to prospective and randomized clinical trials, meta-analyses, and guideline statements, supplemented by influential preclinical and translational work; case reports and small series were cited where they illustrated emerging or investigational approaches. Study selection and emphasis reflected author consensus rather than a predefined, registered protocol, and no formal risk-of-bias or quality appraisal was undertaken.

3. Definition of Radioiodine-Refractory Differentiated Thyroid Cancer

Radioiodine therapy has been the mainstay of treatment for metastatic DTC since the 1940s. About half of patients respond well, but the remainder develop RAI-R DTC—a condition that has caused ongoing debate regarding its precise definition [14,15,16]. In 2014, two attempts were made to define RAI-R DTC. Schlumberger et al. proposed a straightforward rule: discontinue radioiodine when at least one metastasis is iodine-negative and progressing [17]. Sacks and Braunstein went further, suggesting that patients could also be labeled refractory on the basis of a negative diagnostic scan with known structural disease, [¹⁸F]FDG-positive lesions, or cumulative Na[¹³¹I]I exceeding 22 GBq [18]. The first and third criteria proved controversial, as diagnostic scans can miss lesions that show uptake only after application of therapeutic activities [19,20], and arbitrary cumulative limits do not capture the heterogeneity of real clinical practice, but instead oversimplify it [21,22]. The 2015 American Thyroid Association (ATA) guidelines offered a more nuanced view, defining RAI-R disease as lesions not accumulating iodine on the first post-therapy scan; lesions that lose previously documented uptake; mixed uptake patterns; or progression despite adequate uptake [23]. A joint statement from the ATA, the European Association of Nuclear Medicine (EANM), the Society of Nuclear Medicine and Molecular Imaging (SNMMI), and the European Thyroid Association (ETA) later acknowledged that current criteria remain inadequate and should not serve as absolute determinants for stopping radiodiodine therapy [16]. The 2025 ATA guidelines provide further clarification [24]. Critically, they emphasize that RAI-R characteristics should risk-stratify patients according to the likelihood of response; they should not necessarily mandate whether radioiodine is recommended. Strong criteria suggesting RAI-R disease include (i) absence of Na[¹³¹I]I uptake on a post-therapy scan with confirmed structural disease and/or (ii) progression less than 6 months after a positive post-therapy radioiodine scan. Supplemental criteria suggesting reduced radioiodine sensitivity include no uptake on diagnostic whole-body scan despite confirmed disease on morphological or [¹⁸F]FDG positron emission tomography (PET) imaging and uptake in some but not all tumor foci, indicating that radioiodine therapy alone is insufficient (Figure 1).

The guidelines also acknowledge that RAI-R criteria will continue to evolve as imaging techniques improve and redifferentiation therapies enhance the effectiveness of radioiodine. The evolution of RAI-R DTC criteria is summarized in

Table 1. Evolution of radioiodine-refractory differentiated thyroid cancer criteria.

Year	Source	Key Criteria
2014	Schlumberger et al. [17]	At least one lesion is radioiodine-negative and progressive.
2014	Sacks & Braunstein [18]	Negative diagnostic scan with structural disease; [18F]FDG-positive lesions; cumulative activity > 22 GBq.
2015	ATA Guidelines [23]	No uptake on first post-therapy whole-body scan; loss of prior uptake; mixed uptake; progression despite uptake.
2019	ATA/EANM/SNMMI/ETA Joint Statement [16]	Current criteria are inadequate, and clinicians should not treat them as hard rules for stopping radioiodine.
2025	ATA Guidelines [24]	Strong criteria: no uptake on post-therapy scan despite structural disease; progression < 6 months after therapy despite positive scan. Supplemental: negative diagnostic scan; mixed uptake. Criteria should risk-stratify, not mandate therapy decisions.

Legend: FDG, Fluorodeoxyglucose; ATA, American Thyroid Association; EANM, European Association of Nuclear Medicine; SNMMI, Society of Nuclear Medicine and Molecular Imaging; ETA, European Thyroid Association.

.

The 6-month interval, according to the ATA 2025 guideline, refers to the period of progression after a positive post-therapy scan. RAI refractoriness frequently manifests late, sometimes years after the first manifestation and after several treatment courses; in this situation, it is best assessed against the last qualifying administration with documented uptake, since cumulative-activity thresholds are no longer considered appropriate for this purpose. Progression within 6–12 months of such administration, despite adequate uptake, supports the presence of clinically significant refractory disease; periodic reassessment is recommended, particularly after attempts at redifferentiation.

The loss of iodine-concentrating ability stems from reduced NIS expression, driven largely by activation of the MAPK pathway [25,26,27,28,29,30]. BRAF mutations, present in roughly half of papillary thyroid cancers, strongly inhibit NIS expression [31,32]. The effect is variable, however, and is amplified when a TERT promoter mutation co-occurs; this genetic duet robustly predicts loss of radioiodine avidity [33]. RAS and RET alterations also activate this pathway, though with less pronounced effects [34]. Not all driver mutations are equal: BRAF V600E dedifferentiates tumors more aggressively than RAS or RET alterations [35]. However, the real paradigm shift comes from redifferentiation studies: if you can restore iodine uptake with redifferentiation therapies, then ‘refractory’ was never the appropriate term in the first place [36,37,38,39,40]. Refractoriness is better understood as a continuum [14]. A revised definition should incorporate absence of clinically relevant uptake on post-therapy imaging using optimal techniques, i.e., single-photon-emission computed tomography/computed tomography (SPECT/CT) after high administered radioiodine activity (>7400 MBq), provided that adequate thyroid-stimulating hormone (TSH) stimulation (TSH > 30 mU/L, achieved through thyroid hormone withdrawal or recombinant human TSH) is ensured and significant iodine contamination is excluded (low-iodine diet for at least 1–2 weeks; iodinated contrast media avoided for ≥8 weeks or until urinary iodine normalizes)—with ‘relevant’ defined quantitatively: a target-to-background ratio above 4 and uptake at least twice the liver mean uptake [14,41], or an absorbed dose reaching 20 Gy [14,42]; while accepting that mixed uptake reflects tumor heterogeneity, not total resistance [43]; progression within 6–12 months despite adequate uptake; and absence of response to redifferentiation attempts when clinically appropriate based on the tumor’s molecular profile. This status should be reassessed periodically, particularly after attempts at redifferentiation. Dosimetry can help characterize tumor heterogeneity. Taprogge et al. found absorbed doses ranging from <1 Gy to 1170 Gy across lesions [44]. The evidence no longer supports treating radioiodine refractoriness as irreversible, and our definitions must evolve accordingly. These proposed criteria are supported by moderate-quality evidence derived from small prospective phase II trials and expert consensus; prospective validation in larger, multicenter cohorts is required before incorporation into routine guidelines.

4. Molecularly Targeted Therapies in Radioiodine-Refractory DTC

Multikinase inhibitors, including lenvatinib and sorafenib, remain standard systemic therapies for progressive, unresectable disease, improving progression-free survival but with substantial toxicity that often limits long-term use [45,46] (

Table 2. Clinical maturity of therapeutic approaches in radioiodine-refractory differentiated thyroid cancer.

Example of Agents	Clinical Maturity	Current Clinical Positioning	Main Limitations
Multikinase inhibitors Lenvatinib (preferred), sorafenib	Standard of care Guideline-supported	First-line systemic therapy in progressive, symptomatic, or RECIST-progressive RAI-R DTC. Recommended by 2025 ATA.	Non-curative; continuous administration required; substantial chronic toxicity (hypertension, proteinuria, diarrhea, HFSR, fatigue); frequent dose reductions and interruptions
Genotype-matched selective TKIs e.g., BRAF + MEK: dabrafenib–trametinib; RET: selpercatinib, pralsetinib; NTRK: larotrectinib, entrectinib	Standard of care Guideline-supported in molecularly selected disease	Preferred first-line when an actionable driver is identified on NGS. Endorsed by the 2025 ATA in genotype-matched settings.	Requires comprehensive NGS access; acquired resistance, high drug cost, and reimbursement variability
Redifferentiation therapy Short-course targeted therapy followed by Na[131I]I selumetinib; dabrafenib ± trametinib; trametinib; selpercatinib; larotrectinib	Emerging Limited, mostly phase II evidence	Individualized option in selected patients with an actionable driver, and preferably predicted lesional absorbed dose ≥ 20 Gy at activity ≤ MTA	Heterogeneous response across drivers (BRAF V600E 33–60%; RAS 60–100%); preferably with Na[131I]I PET/CT dosimetry access; optimal duration and sequencing have not been established; response low with TERT co-mutation
PRRT (SSTR-targeted) [177Lu]Lu-DOTATATE, [90Y]Y-DOTATOC	Investigational Retrospective series; no DTC-specific RCT	Off-label use in highly SSTR-expressing RAI-R DTC after failure of standard systemic therapy. No regulatory approval for DTC.	Heterogeneous SSTR expression in DTC; not all lesions are targetable; overlapping toxicity with prior Na[131I]I; no randomized prospective data in DTC
PSMA-directed radioligand therapy [177Lu]Lu-PSMA	Investigational Case reports; no prospective data	Not routinely recommended. May be considered in individual cases with strong lesional PSMA uptake after exhaustion of standard options. Ideally, within prospective clinical trials.	PSMA expression heterogeneous and predominantly neovascular; no validated dosimetric predictor of response; potential salivary toxicity additive to prior Na[131I]I
FAPI-based theranostics [177Lu]Lu-FAPI	Investigational Pilot series only	Not recommended outside clinical trials. May be considered in FAPI-avid lesions when standard options have failed.	Limited reports and short follow-up; response durability unknown
Targeted alpha therapy Na[211At]At (NIS-mediated uptake; not retained via TPO)	Investigational First-in-human phase I	Clinical trials only. Potential future role in patients with preserved NIS expression who have progressed on Na[131I]I and systemic therapies.	Limited geographic availability; dose-limiting hematological toxicity at higher activities; Na[211At]At, not organized by TPO, limited intracellular retention; optimal multi-cycle regimen and combinations not yet defined

Legend: ATA, American Thyroid Association; RAI-R DTC, radioiodine-refractory differentiated thyroid carcinoma; FAPI fibroblast activation protein inhibitor; HFSR, hand–foot skin reaction; MTA, maximum tolerated activity; NGS, next-generation sequencing; NIS, sodium–iodide symporter; PRRT, peptide receptor radionuclide therapy; PSMA, prostate-specific membrane antigen; RCT, randomized controlled trial; SSTR, somatostatin receptor; TERT, telomerase reverse transcriptase; TKI, tyrosine-kinase inhibitor; TPO, thyroid peroxidase.

).

Among MKIs, lenvatinib is generally preferred over sorafenib based on higher objective response rates (≈60–70%) and longer progression-free survival, although dose reductions are common due to hypertension, proteinuria, diarrhea, fatigue, and weight loss [45,47,48,49]. Sorafenib achieves more modest response rates (≈10–15%) and is frequently limited by hand–foot skin reaction, rash, and diarrhea [50]. Highly selective agents demonstrate superior efficacy and improved tolerability in molecularly defined subgroups. BRAF inhibitors combined with MEK inhibitors produce response rates of 50–70% in BRAF V600E–mutant DTC, with common toxicities including pyrexia, rash, and fatigue [51]. RET inhibitors (e.g., selpercatinib) and NTRK inhibitors (e.g., larotrectinib) achieve response rates exceeding 70%, often with durable disease control; adverse events are typically low-grade and include transaminase elevations, hypertension, and dizziness [52,53]. These favorable profiles support prioritization of molecular profiling and genotype-matched therapy whenever possible [46,54]. Next-generation sequencing (NGS) is central to therapeutic selection, enabling identification of actionable alterations, stratification for redifferentiation strategies, and prioritization of selective kinase inhibitors over MKIs when available [46,54]. Accordingly, early integration of NGS is recommended in patients with RAI-R DTC requiring systemic therapy [55].

For patients with progressive RAI-R DTC and no actionable driver who progress on first-line MKI therapy, cabozantinib is the evidence-based second-line option, having significantly prolonged progression-free survival versus placebo in the phase III COSMIC-311 trial after progression on prior VEGFR-targeted therapy [56]. A theranostic approach is a less well-supported alternative in this context, since redifferentiation is generally not applicable to driver-negative disease and receptor-directed radioligand therapy remains investigational. Selection between second-line cabozantinib and a theranostic strategy should weigh (i) the presence and intensity of a targetable imaging phenotype (SSTR, PSMA, or FAPI uptake); (ii) disease kinetics and burden, with rapidly progressive or symptomatic disease favoring a proven systemic agent over an investigational one; (iii) cumulative prior toxicity, particularly hematological and salivary, which overlaps with several radioligand therapies; (iv) regulatory status and the availability of a prospective trial; and (v) multidisciplinary tumor-board review. In practice, cabozantinib remains the standard second-line choice, with theranostic approaches reserved for selected patients with a strong target phenotype, preferably within prospective trials.

Donafenib, a deuterated derivative of sorafenib developed in China, was evaluated in the Multicenter Randomized Double-Blind Phase III Trial: 191 patients with progressive RAI-R DTC randomized 2:1 to donafenib 300 mg twice daily or placebo [57]. Median progression-free survival was 12.9 versus 6.4 months (HR, 0.39; 95% CI, 0.25–0.61; p < 0.0001), with objective response rates of 23.3% versus 1.7% and disease control rates of 93.3% versus 79.3%. The most common grade ≥ 3 treatment-related adverse events were hypertension (13.3%) and hand–foot syndrome (12.5%); 42.2% required dose reduction or interruption, and 6.3% were discontinued. The toxicity profile is typical of the VEGFR-directed MKI class. Generalizability is limited, however: the trial enrolled only Chinese patients, donafenib is not approved for thyroid cancer outside China, and there is no head-to-head comparison with lenvatinib, sorafenib, or cabozantinib.

A comprehensive discussion of molecularly targeted pharmacological therapies is beyond the scope of the present work; therefore, we refer readers to the excellent studies cited above. With regard to the theranostic approach, however, it is important to emphasize that certain targeted agents have demonstrated the ability to restore radioiodine uptake and retention in RAI-R DTC cells, thereby enabling so-called redifferentiation therapy, as described below (Figure 2).

5. Redifferentiation Strategies and Radioiodine Therapy

The mitogen-activated protein kinase (MAPK) pathway is a central intracellular signaling cascade that transmits signals from cell-surface receptors to the nucleus, regulating cell proliferation, differentiation, survival, and metabolism [58]. In DTC, constitutive MAPK activation—most commonly driven by BRAF V600E mutations or RAS alterations—promotes tumor growth and progressive dedifferentiation [28,59,60,61,62]. MAPK hyperactivation suppresses thyroid-specific genes involved in iodine handling, including the NIS, thyroglobulin (Tg), and thyroid peroxidase, thereby contributing to RAI refractoriness. Pharmacologic inhibition of MAPK signaling can partially reverse this dedifferentiated phenotype. Targeted agents, particularly BRAF and MEK inhibitors, may restore expression of iodine-handling machinery, re-enable iodine uptake, and allow renewed use of RAI (“redifferentiation therapy”). However, redifferentiation is not guaranteed—even in molecularly favorable tumors—and should be conceptualized as a testable intervention rather than a definitive therapeutic pathway [63,64,65]. Consequently, contemporary practice emphasizes an imaging-gated theranostic strategy: a defined course of targeted therapy is followed by reassessment of iodine uptake and dosimetry using quantitative molecular imaging Na[¹²⁴I]I positron emission tomography/computed tomography (PET/CT) or Na[¹³¹I]I single-photon emission computed tomography/computed tomography (SPECT/CT) [66]. Only patients predicted to achieve a therapeutically meaningful absorbed dose proceed to high-activity Na[¹³¹I]I. Treatment sequencing must remain individualized, particularly in patients with bulky or rapidly progressive disease, where delaying effective systemic therapy to attempt resensitization may carry clinical risk.

The initial proof of concept for redifferentiation was provided by MEK inhibition. In a landmark study, selumetinib increased iodine uptake in a subset of RAI-R DTC patients, enabling dosimetrically guided Na[¹³¹I]I therapy. Tumors harboring RAS mutations appeared particularly responsive, consistent with their dependence on downstream MAPK signaling. This study established the paradigm of imaging-gated redifferentiation and underscored the importance of dosimetry-driven patient selection [42]. In BRAF V600E-mutated RAI-refractory DTC, combined BRAF (i.e., vemurafenib) and MEK inhibition has generally demonstrated greater efficacy in restoring iodine avidity than single-agent approaches. Prospective phase II data using dabrafenib plus trametinib followed by Na[¹³¹I]I have shown clinically meaningful results, providing a structured framework for resensitization that integrates targeted therapy, repeat imaging, and therapeutic RAI [40]. Beyond MAPK-directed therapies, highly selective inhibitors targeting oncogenic fusions have shown potential for redifferentiation in molecularly defined subgroups. RET inhibitors such as selpercatinib have been associated with restored iodine uptake in RET-rearranged thyroid cancers, including pediatric cases, enabling successful RAI therapy [67,68]. Similarly, the TRK inhibitor larotrectinib has demonstrated redifferentiating effects in NTRK-fusion-positive thyroid cancers in both adult and pediatric populations [69,70,71].

Overall, clinical outcomes following redifferentiation therapy are heterogeneous. Restoration of iodine uptake, durability of redifferentiation, and response to subsequent RAI vary widely and are influenced by tumor genotype, disease burden, duration of targeted therapy, and dosimetric strategy. While some patients achieve partial responses or prolonged disease stabilization, others experience only transient uptake with early relapse [37,72,73,74,75]. Recent syntheses report wide ranges in iodine uptake restoration (approximately 33–95%) and tumor response rates (11–80%), highlighting the need for continued refinement of patient selection, standardized imaging protocols, and integration of real-world evidence into future trial design [21]. A recent comprehensive synthesis highlights wide variability in reported iodine uptake restoration (33–95%) and tumor response rates (11–80%) across studies, reflecting differences in inclusion criteria, therapy duration, dosimetry, and trial design [62] (

Table 3. Clinical redifferentiation studies in radioiodine-refractory differentiated thyroid cancer: effect sizes by driver mutation.

Study (First Author, Year)	Driver/Agent(s)	Design	Preparation	N	Meaningful RAI Reuptake	Treated with [131I] → Response	Key Limitations
BRAF V600E-mutated (and mixed) RAI-R DTC
Rothenberg, 2015 [76]	BRAF V600E Dabrafenib 150 mg bid	Prospective, phase I/pilot	rhTSH	10	6/10 (60%) new sites of uptake on diagnostic Na[131I]I WBS	6/10 received empiric 5.5 GBq Na[131I]I; at 3 months PR 2/6, SD 4/6	Small cohort; diagnostic rather than quantitative imaging; short follow-up; no lesional dosimetry; one grade 3 cutaneous squamous cell carcinoma
Dunn, 2019 [55]	BRAF V600E Vemurafenib 960 mg bid × 4 weeks	Prospective, phase II	rhTSH; Na[124I]I PET/CT dosimetry	12 enrolled; 10 completed	6/10 (60%) new or increased uptake on Na[124I]I PET/CT	4/6 met dosimetric threshold (≥20 Gy, AA ≤ 11.1 GBq) and received Na[131I]I; at 6 months PR 2/4, SD 2/4	Small cohort; supportive molecular data (increased thyroid-differentiation score and NIS mRNA) in 3 biopsied cases
Jaber, 2018 [77]	BRAF V600E (9), NRAS (3), WT (1) Single-agent or combination BRAF ± MEK inhibitors (dabrafenib; vemurafenib; dabrafenib + trametinib; trametinib)	Retrospective, single-center	Variable; mostly empiric	13	8/13 (62%) restored Na[131I]I uptake on diagnostic WBS	8 received median 7.4 GBq Na[131I]I; 1 treated empirically without clear restoration. At median 14.3-month follow-up: PR 3/9, SD 6/9 (including the empirically treated patient)	Retrospective design; median targeted therapy duration 14.4 months (range 0.9–76.4); only 3/9 underwent dosimetric pre-therapy evaluation; heterogeneous prior exposure
Leboulleux, 2019 [38]	BRAFK601E (poorly diff. DTC) Dabrafenib + trametinib	Case report	rhTSH	1	Restored Na[124I]I uptake	1/1 received Na[131I]I; biochemical and structural PR	Single patient; not generalizable
Weber, 2022 [41]	BRAF-mutated (6)/BRAF wild-type (14) Dabrafenib 75 mg bid + trametinib 2 mg (BRAF-mut); trametinib 2 mg (BRAF-WT) × 21 days	Prospective, single-center, two-arm	rhTSH; Na[123I]I SPECT/CT-gated; dosimetry-guided Na[131I]I	20 (6 BRAF-mut, 14 BRAF-WT)	7/20 (35%) met redifferentiation criteria (T/B > 4 and >2× liver uptake): 2/6 BRAF-mutated, 5/14 BRAF wild-type	All 7 responders received dosimetry-guided Na[131I]I (mean 300 mCi, range 273–422). Thyroglobulin decline in 4/7 (57%). At 1 year: PR 1/7, SD 5/7, PD 1/7 SUVpeak < 10 on [18F]FDG PET associated with successful redifferentiation (p = 0.01).	Small per-arm numbers; no comparator; short Na[131I]I preparation window (21 days)
Leboulleux, 2023 [40]	BRAF V600E Dabrafenib 150 mg bid + trametinib 2 mg × 6 weeks	Prospective, phase II, multicenter *	rhTSH; empiric 5.5 GBq Na[131I]I on day 35 regardless of diagnostic WBS	24 enrolled; 21 evaluable	Post-therapy WBS uptake in 20/21 patients	All evaluable patients received Na[131I]I; at 6 months PR 8/21 (38%), SD 11/21 (52%), PD 2/21 (10%). Second course in 11 eligible patients: 1 CR, 6 PR, 2 SD, 1 PD, 1 NE PFS 82% at 1 year, 68% at 2 years.	No randomized comparator; separating targeted-therapy effect from Na[131I]I contribution is difficult; requires longer follow-up for durability
RAS-mutated (and mixed) RAI-R DTC
Ho, 2013 [42]	Mixed (NRAS 5, BRAF 9, other 6) Selumetinib 75 mg bid × 4 weeks	Prospective, phase I (landmark proof-of-concept)	rhTSH; Na[124I]I PET/CT dosimetry	24 enrolled; 20 evaluable	12/20 (60%) restored uptake overall NRAS: 5/5 (100%) BRAF: 4/9 (44%)	8/12 met dosimetric threshold (≥20 Gy, AA ≤ 11.1 GBq) and received Na[131I]I; at 6 months PR 5/8, SD 3/8. All NRAS patients with restored uptake met the threshold and received RIT Only 1/4 of BRAF patients met the threshold.	Landmark proof-of-concept; small BRAF subgroup; MEK single-agent likely suboptimal in BRAF-mutated disease
Leboulleux, 2023 [39]	RAS-mutated Trametinib 2 mg daily × 6 weeks	Prospective, phase II, multicenter *	rhTSH; fixed empiric 5.5 GBq Na[131I]I on day 35	11 enrolled; 10 evaluable	Increased Na[131I]I uptake in ~2/3 of patients	All evaluable patients received Na[131I]I; at 6 months PR 2/10 (20%), SD 7/10 (70%), PD 1/10. Second course in 3 patients: 1 maintained PR 18 months; 2 PD. Median PFS 1 year.	Small cohort; single-agent MEK inhibition suboptimal in some patients; 2/11 treatment discontinued (grade 3 colitis; grade 2 LVEF decrease)
Burman, 2022 [78]	RAS-mutant and RAS wild-type Trametinib + Na[131I]I	Phase II, prospective	rhTSH; Na[124I]I PET/CT dosimetry (≥20 Gy, AA ≤11.1 GBq)	34 (25 RAS-mut; 9 RAS-WT)	RAS-mut: 15/25 (60%) met dosimetric threshold RAS-WT: 3/4 BRAF and 1/4 RET met threshold	RAS-mut: 14/15 received Na[131I]I; at 6 months PR 8/14 (57%), SD 3/14 (21%), PD 3/14 (21%); PFS 44% at 6 months RAS-WT (BRAF/RET/STK11): 3 SD, 1 PR (PR in a BRAF-mutant patient)	Heterogeneous comparator groups
RET-altered RAI-R DTC
Chan, 2023 [79]	RET fusion Pralsetinib	Case report	rhTSH	1	Reversed “flip-flop” of Na[131I]I and [18F]FDG avidity; restored RAI uptake in metastatic lesions	Received Na[131I]I; structural and biochemical response	Single patient
Werner, 2023 [67]	RET-altered Selpercatinib 160 mg bid × 3 weeks (after 15.5 months of prior selpercatinib therapy)	Case report with pre-therapy dosimetry	rhTSH; individualized dosimetric planning	1	Intense uptake on diagnostic Na[131I]IWBS; lesional absorbed dose up to 197 Gy	Received 9.4 GBq Na[131I]I; previously iodine-negative lung nodules showed intense uptake on post-therapy scan; Tg decline; nodule shrinkage on CT	Single patient; proof-of-principle with unusually high achievable lesion doses
NTRK fusion–positive (and mixed) RAI-R DTC
Groussin, 2020 [69]	EML4–NTRK3 fusion Larotrectinib 100 mg bid	Case report	rhTSH	1	Marked restoration of Na[124I]I uptake after 6 months of larotrectinib	Received Na[131I]I; structural and biochemical response	Single patient; durability limited by TERT co-mutation concern
Lee, 2021 [80]	TPR–NTRK1 and CCDC6–RET fusions (pediatric) Larotrectinib 100 mg bid (NTRK); selpercatinib 80 mg bid (RET)	Two pediatric cases with comprehensive genomic profiling	Variable	2	Both cases: decreased tumor size and reinduced RAI uptake following fusion-directed therapy	Both received Na[131I]I; structural and biochemical response	Very small case series; pediatric cohort; generalizability to adult disease unproven
Mixed drivers
Iravani, 2019 [81]	BRAF V600E (3) and NRAS (3) BRAF: dabrafenib + trametinib or vemurafenib + cobimetinib × 4 weeks NRAS: trametinib 2 mg daily × 4 weeks	Prospective, single-center	Variable	6 (3 BRAF V600E PTC; 2 FTC + 1 PDTC with NRAS)	4/6 (67%) restored uptake: 3/3 BRAF V600E; 1/3 NRAS	4/6 proceeded to Na[131I]I. At median 16.6-month follow-up: partial imaging response beyond 3 months in 3/4 (2 BRAF, 1 NRAS)	Small cohort; heterogeneous histologies and drivers. Protocol evolved during enrolment.
Toro-Tobon, 2024 [37]	Mixed (BRAF, RAS, RET, ALK) Trametinib, selpercatinib, pralsetinib, alectinib, or dabrafenib + trametinib × 4 weeks	Retrospective, single-center	rhTSH-stimulated Na[123I]I WBS at week 3	33	19/33 (57.6%) restored uptake overall By driver: RAS 100%; invasive EFVPTC and FTC 100%; classical PTC 42.1%; BRAF-mutant 38.9%	Redifferentiated patients proceeded to high-activity Na[131I]I. Additional ~20% tumor reduction at 6 months beyond the ~12% shrinkage at 3 weeks. No significant PFS or time-to-next-therapy differences vs. the non-redifferentiated group. Anaplastic transformation in 2/33 (6.1%); 5/33 (15.1%) died during follow-up (all post-RIT).	Retrospective; selection bias. Safety signal on anaplastic transformation warrants longer follow-up
von Hinten, 2025 [63]	Mixed drivers Short-course driver-matched targeted therapy	Feasibility, prospective	rhTSH	Small cohort (feasibility)	Clinically meaningful RAI reuptake in a subset	Patients meeting imaging criteria proceeded to Na[131I]I; biochemical responses reported with short follow-up.	Feasibility focus; emphasizes short-course approach; follow-up <12 months

Legend: AA, administered activity; CR, complete response; DTC, differentiated thyroid carcinoma; EFVPTC, encapsulated follicular variant of papillary thyroid carcinoma; FTC, follicular thyroid carcinoma; LVEF, left ventricular ejection fraction; NE, not evaluable; NIS, sodium–iodide symporter; PD, progressive disease; PDTC, poorly differentiated thyroid carcinoma; PFS, progression-free survival; PR, partial response; PTC, papillary thyroid carcinoma; RAI, radioactive iodine; RAI-R, radioiodine-refractory; rhTSH, recombinant human thyroid-stimulating hormone; SD, stable disease; T/B, target-to-background ratio; TERT, telomerase reverse transcriptase; Tg, thyroglobulin; WBS, whole-body scintigraphy; WT, wild-type. * Inclusion required documented RECIST progression within 18 months before enrolment (and no target lesion > 30 mm).

).

Pre-treatment quantification of lesional iodine avidity (preferably Na[¹²⁴I]I PET/CT, alternatively Na[¹³¹I]I SPECT/CT) determines whether a clinically meaningful absorbed dose is achievable; only patients with predicted lesional absorbed dose ≥ 20 Gy at an administered activity at or below the maximum tolerated activity usually proceed to Na[¹³¹I]I therapy. The absorbed dose delivered to a lesion reflects the cumulative number of radioactive decays occurring within it, i.e., the area under the lesion time–activity curve, normalized to lesion mass. Dose therefore depends jointly on uptake (initial trapping) and retention (biological half-life and organification), rather than on a single static imaging measurement. In practice, the 20 Gy lesion threshold is best conceptualized not as an absolute requirement for every lesion but as a target for the majority of lesions, achievable with an administered activity ≤ the maximum tolerated activity (MTA). MTA in DTC is typically derived from healthy-tissue dosimetry (red marrow blood-based dose ≤ 2 Gy; 48 h retention limit of ≤4.44 GBq, in general, and ≤2.96 GBq for diffuse pulmonary metastases). If the estimated lesion dose per administered activity (LDpA, Gy/GBq) is such that the MTA delivers < 20 Gy to most lesions, redifferentiation is unlikely to translate into a meaningful therapeutic outcome, even when iodine uptake is visually restored. Pretreatment dosimetry must be performed under the same biological conditions as the planned therapy (recombinant human stimulation if recombinant human TSH is planned, or thyroid hormone withdrawal if thyroid hormone withdrawal is planned), since iodine handling and lesion uptake differ between the two preparations. In practical terms, redifferentiation is best suited to patients with progressive RAI-R DTC and an actionable driver, predominantly BRAF V600E, whose disease burden allows a short induction course of targeted therapy before radioiodine. Response should be confirmed objectively with Na[¹²⁴I]I PET/CT, alternatively Na[¹³¹I]I SPECT/CT, supported by Tg and structural imaging. Failure to restore meaningful iodine avidity should prompt a timely return to standard systemic therapy.

6. Non-Radioiodine Theranostics in Thyroid Cancer

In recent years, theranostic research has also focused on the repurposing of already available radiopharmaceuticals used in non-thyroid malignancies, as well as on the development of thyroid-targeting agents with mechanisms distinct from radioiodine. In particular, although still not definitive, emerging data are available for radiolabeled somatostatin analogs, prostate-specific membrane antigen (PSMA)-targeted agents, fibroblast activation protein inhibitor (FAPI)-targeting radiopharmaceuticals, and Astatine-211.

6.1. Somatostatin-Receptor-Targeting Radiopharmaceuticals

Radiopharmaceuticals targeting SSTRs enable imaging of tumors with elevated receptor expression [82]. [⁶⁸Ga]Ga-DOTA-TATE, [⁶⁸Ga]Ga-DOTA-TOC and [⁶⁸Ga]Ga-DOTA-NOC are the most widely applied in clinical practice, and they differ in their binding profiles across SSTR subtypes. [⁶⁸Ga]Ga-DOTA-TATE binds with high selectivity to SSTR2, whereas [⁶⁸Ga]Ga-DOTA-NOC has affinity for both SSTR2 and SSTR3, and [⁶⁸Ga]Ga-DOTA-TOC binds SSTR2 and SSTR5 [83]. DTC cells have been shown to express SSTR2, SSTR3, and SSTR5. On this basis, radiolabeled SSTR analogs may assist in identifying recurrence or metastases, especially in the RAI-R setting [84,85]. The efficacy of radiotracers was tested in several studies. In a cohort of 13 patients with RAI-R DTC, [⁶⁸Ga]Ga-DOTA-TATE and [⁶⁸Ga]Ga-DOTA-NOC were compared for lesion detection. Somatostatin-avid disease was identified in eight patients (62%), with 45 lesions detected by [⁶⁸Ga]Ga-DOTA-TATE versus 42 by [⁶⁸Ga]Ga-DOTA-NOC. Lesion uptake was markedly higher with [⁶⁸Ga]Ga-DOTA-TATE (SUVmax 12.9 ± 9.1 vs. 6.3 ± 4.1), suggesting a potential advantage in this setting [86]. Factors linked to SSTR expression in recurrent or metastatic DTC have been examined in a study by Watanabe et al., which identified higher receptor expression in follicular thyroid carcinoma and reported that elevated thyroglobulin levels, larger tumor size, and preserved Na[¹³¹I]I uptake were more frequently associated with SSTR-positive lesions [87].

SSTR antagonists have recently emerged as an alternative to somatostatin analogs, offering a greater number of binding sites and longer cell-surface residence time, with applications in both diagnosis and therapy across several SSTR-positive tumors, including DTC, medullary thyroid carcinoma, and small-cell lung cancer. SSTR-avid lesions on imaging open the possibility of PRRTs within a theranostic approach [82,88].

[⁶⁸Ga]Ga-DOTA-TOC PET/CT has been applied to select RAI-R DTC patients for PRRT and to monitor treatment response and toxicity. Of 41 patients evaluated, 24 were [⁶⁸Ga]Ga-DOTA-TOC-positive, and 13 showed high SSTR expression; 11 ultimately received [⁹⁰Y]Y-DOTATOC. Over a follow-up of 3.5–11.5 months, disease control was achieved in 7/11 (64%), comprising partial response in two and stable disease in five. Change in functional volume on PET/CT was the only parameter that discriminated responding from non-responding lesions. Adverse events were predominantly transient and included hematologic toxicity, nausea, and asthenia, with permanent renal toxicity in one patient [89].

In a systematic review by Maghsoomi et al. of PRRT in advanced RAI-R DTC and metastatic MTC, 41 of 2284 screened papers were included, covering 157 RAI-R DTC patients treated with PRRT. Biochemical and objective (partial or complete) responses were achieved in 25.3% and 10.5%, respectively [90]. The objective response rate with PRRT is low, in contrast to the substantially higher disease-control rate. Disease control was defined by RECIST as the sum of complete response, partial response, and stable disease, thereby incorporating stable disease, which, in the typically indolent course of RAI-R DTC, may reflect the natural history rather than a treatment effect. Without a randomized comparison, the divergence cannot be attributed to PRRT and is likely inflated by selection bias inherent in non-randomized, largely retrospective designs that preferentially enroll patients with high SSTR expression and often slowly progressive disease, a limitation acknowledged by Maghsoomi and colleagues. These observations support positioning PRRT as investigational pending prospective controlled trials. Side effects were mostly minor (nausea, asthenia, raised liver enzymes), with rare and largely transient major events such as nephrotoxicity and hematologic toxicity. A recent meta-analysis by Lee and colleagues similarly assessed PRRT in metastatic RAI-R DTC and MTC, pooling 11 studies and 165 patients (98 with MTC and 67 with DTC) [91]. Pooled objective response and disease-control rates were 8.5–15.6% and 54–60.0%, respectively, with major adverse effects in 2.79–2.82% of cases; ¹⁷⁷Lu-based somatostatin analogs proved more effective than ⁹⁰Y-based analogs, with an objective response rate (ORR) of 11.48–24.52% and a disease control rate (DCR) of 61.47–67.26%, while ⁹⁰Y-based analogs had an ORR of 6.98–13.82% and a DCR of 50.86–57.29%.

A proper selection is key to avoiding inappropriate treatments with attached side effects and costs. Not all lesions detected on PET/CT are necessarily suitable for PRRT, and one of the most important parameters remains the uptake intensity: by definition, lesions with activity lower than that of the hepatic parenchyma have a low probability of response and therefore represent a contraindication to PRRT (Figure 3).

In conclusion, positive RAI-R DTC lesions on SSTR imaging provide an opportunity to treat selected patients with peptide receptor radionuclide therapy (PRRT) within a theranostic framework. In this setting, SSTR PET/CT enables in vivo assessment of receptor expression, which directly determines treatment eligibility. Importantly, only lesions demonstrating uptake equal to or higher than hepatic activity are considered suitable for PRRT, as lower uptake is associated with a low probability of therapeutic response. Thus, SSTR imaging functions not only as a diagnostic modality but also as a gatekeeper for therapy selection. PRRT using ¹⁷⁷Lu- or ⁹⁰Y-labeled somatostatin analogs may therefore be considered in patients with sufficient receptor expression. However, RAI-refractory tumors often exhibit a slow progression kinetics, while currently available studies evaluated results in a relatively short period. Accordingly, given the heterogeneous response rates, short follow-up studies, and limited prospective evidence, careful selection of patients and the use of reliable imaging-based methods remain essential to optimize outcomes and avoid ineffective treatment.

6.2. PSMA-Targeting Radiopharmaceuticals

Prostate-specific membrane antigen, encoded by the folate hydrolase 1 gene, is a type II transmembrane glycoprotein [92]. Beyond prostate cancer, it is expressed on tumor neovascular endothelium in many solid cancers, including thyroid [82,93,94,95,96]. High neovascular PSMA expression in thyroid cancer correlates with a more aggressive course, and moderate-to-strong expression predicts higher RAI-R risk and disease-specific mortality in DTC [94,97], supporting PSMA-targeting tracers for both diagnosis and therapy. These are usually labeled with ⁶⁸Ga or ¹⁸F, the latter giving better image quality through its shorter positron range, lower maximum β⁺ energy, and higher positron yield [98,99]. However, ¹⁸F-labeled PSMA compounds exhibit higher bone uptake [100,101], resulting in more nonspecific bone uptake than ⁶⁸Ga-labeled PSMA [102]. No data are available in RAI-R DTC patients, but it is relevant to note that no differences in clinical classification are observed in patients with prostate cancer when evaluated by either ¹⁸F- or ⁶⁸Ga-labeled PSMA PET/CT, evaluated by expert nuclear medicine physicians.

The correlation between PSMA expression on histology and PSMA-positive lesions on PET/CT was examined in a large immunohistochemical analysis of 44 DTC patients with persistent or recurrent neck disease. PSMA expression was quantified using the immunoreactive score (IRS), with at least one positive lesion (IRS ≥ 2) detected in 68% of cases—66% in RAI-negative and 75% in RAI-positive patients [93].

The diagnostic performance of PSMA-targeted radiopharmaceuticals in RAI-R DTC has been assessed in a systematic review of six studies (n = 49) [103]. Detection rates varied widely (25–100%) and were generally lower than those of [¹⁸F]FDG PET/CT in comparative analyses, ranging from 25 to 83% per patient and around 65% per lesion, with only two studies reporting 100% patient-based detection. SUVmax values showed marked heterogeneity (1.0–39.7). Overall, the current evidence does not support PSMA-targeted imaging as an alternative for restaging RAI-R DTC, although these tracers have the potential for selecting candidates for PSMA-directed radioligand therapy in selected cases, i.e., those with significant uptake in tumor lesions (Figure 4).

A recent phase II study evaluated [¹⁸F]AlF-PSMA-11 PET/CT alongside [¹⁸F]FDG PET/CT in eight patients with RAI-R thyroid carcinoma, identifying 39 lesions in total [104]. Detection rates were 76.9% for [¹⁸F]AlF-PSMA-11 and 84.6% for [¹⁸F]FDG; importantly, the two modalities were complementary, with nine lesions visualized only on [¹⁸F]FDG and six only on [¹⁸F]AlF-PSMA-11. PSMA expression was confirmed immunohistochemically in the primary tumor of five patients, whereas serum-soluble PSMA showed no correlation with disease parameters. These findings support a potential complementary diagnostic role for PSMA-targeted imaging in RAI-R disease.

Receptor expression has been evaluated in metastatic DTC patients with rising Tg and negative iodine scintigraphy using [⁶⁸Ga]Ga-DOTA-TATE (Krenning’s score) and [⁶⁸Ga]Ga-PSMA-11 PET/CT (miPSMA score), with eligibility for [¹⁷⁷Lu]Lu-DOTA-TATE or [¹⁷⁷Lu]Lu-PSMA-617 therapy defined as a Krenning’s score ≥ 3 and miPSMA score ≥ 2. Patient-based positivity rates were 40.5% for [⁶⁸Ga]Ga-DOTA-TATE, 41.9% for [⁶⁸Ga]Ga-PSMA-11, and 75.7% for [¹⁸F]FDG. Of 74 patients, 14 (18.9%) met criteria for radioligand therapy: seven (9.5%) for [¹⁷⁷Lu]Lu-DOTA-TATE, nine (12.2%) for [¹⁷⁷Lu]Lu-PSMA-targeted therapy, and four for both [105].

Published experience with [¹⁷⁷Lu]Lu-PSMA-targeted therapy in RAI-R DTC remains limited, with reported outcomes including temporary partial responses and early disease progression [106,107]. The eligibility thresholds in this study were originally established for neuroendocrine and prostate cancer, respectively. However, they were also applied as minimum requirements for patient selection. Notably, pretherapy dosimetry using PSMA-ligands is also not straightforward, as theranostics couples with long enough half-lives are currently not at hand to determine if one could reach the minimum effective dose in the lesions. Further studies are needed to establish specific thresholds for different diseases and optimize patient selection. At the moment, however, it is worthwhile to use PET data as stratifier, taking into account that uptake does not necessarily mean retention, and PSMA-1007 (imaging) is not the same as PSMA-617 (therapy). Reported toxicity has been minimal, limited to transient nausea after the second therapy course in one patient [107]. Although experience with [¹⁷⁷Lu]Lu-PSMA radioligand therapy in DTC remains scarce, its safety profile is expected to parallel that established in prostate cancer [108,109]. In thyroid cancer patients, the risk of xerostomia may be further increased due to prior salivary gland damage from Na[¹³¹I] therapy (synergistic effect). Overall, the literature on [¹⁷⁷Lu]Lu-PSMA-targeting radioligands in RAI-R DTC is limited. Prospective multicentric studies are needed to properly evaluate its potential role in RAI-R DTC patients.

6.3. Fibroblast Activation Protein-Targeting Radiopharmaceuticals

Fibroblast activation protein (FAP) is minimally expressed in normal fibroblasts but markedly upregulated in cancer-associated fibroblasts in over 90% of epithelial carcinomas [82,110,111,112,113]. Thus, radiolabeled fibroblast activation protein inhibitors (FAPIs) are suitable for tumor imaging.

[⁶⁸Ga]Ga-DOTA-FAPI-04 PET/CT has been evaluated for the detection of RAI-R DTC lesions in 24 patients, with FAPI-avid disease identified in 21 (87.5%) and a mean SUVmax of 4.25; SUVmax correlated positively with lesion growth rate [114]. In comparative studies, [⁶⁸Ga]Ga-DOTA-labeled FAPI radiopharmaceuticals have outperformed [¹⁸F]FDG in selected metastatic RAI-R DTC cases, attributable to a higher target-to-background ratio [115].

In 23 DTC patients with biochemical recurrence (suppressed Tg ≥ 1 ng/mL or Tg-Ab > 115 IU/mL), [¹⁸F]AlF-NOTA-FAPI-04 PET/CT demonstrated significantly higher lesion-level sensitivity (75% vs. 13%; p = 0.002) and accuracy (80% vs. 27%; p < 0.001) than [¹⁸F]FDG, with participant-level sensitivity reaching 100% versus 14% (p = 0.003). Seven small cervical lymph node metastases (mean diameter 5.9 ± 1.4 mm) were detected exclusively on FAPI imaging [116]. Strong, detectable expression of FAP presents an opportunity for new therapeutic options in RAI-R DTC patients.

In a recent head-to-head comparison of [⁶⁸Ga]Ga-FAPI-46 and [¹⁸F]FDG PET/CT in 14 patients with RAI-R DTC and elevated thyroglobulin, both tracers detected active lesions in 10 patients with comparable lesion number and intensity. However, [⁶⁸Ga]Ga-FAPI-46 imaging showed significantly lower background uptake in the blood pool and liver (p = 0.001), yielding superior target-to-background ratios—lesion-to-liver SUVmax exceeded 3 in all [⁶⁸Ga]Ga-FAPI-46 scans versus half of [¹⁸F]FDG scans. In one of five patients with pulmonary metastases, lung nodules were [¹⁸F]FDG-negative but [⁶⁸Ga]Ga-FAPI-46-positive, resulting in upstaging. The authors concluded that [⁶⁸Ga]Ga-FAPI-46 accumulates at least as effectively as [¹⁸F]FDG in recurrent and metastatic RAI-R DTC lesions, with the added advantages of lower injected activity and improved contrast [117].

Ballal et al. conducted a study to evaluate the efficiency and safety of [¹⁷⁷Lu]Lu-DOTAGA.(SA.FAPi)₂ in patients with metastatic RAI-R DTC who progressed on MKI therapy (sorafenib/lenvatinib) [118]. They prospectively enrolled 15 patients with metastatic disease who had moderate-to-strong uptake of [⁶⁸Ga]Ga-DOTA.SA.FAPi. Treatment with [¹⁷⁷Lu]Lu-DOTAGA.(SA.FAPi)₂ in RAI-R DTC produced a marked decline in serum Tg, with partial response in four patients and stable disease in three, although no complete responses were observed. Symptom burden also improved, with the maximal visual analog score decreasing from 9 to 6 and Eastern Cooperative Oncology Group (ECOG) performance status from 3 to 2. No grade III–IV hematological, renal, or hepatic toxicity was recorded, supporting FAPI-based theranostics as a potential additional option in metastatic RAI-R DTC after failure of standard therapies. However, several limitations must be acknowledged. The studies were retrospective in part, lesion validation was incomplete, and response criteria were not standardized (no RECIST or defined Tg cut-offs). Follow-up was limited, and key outcomes such as duration of response were not reported. Well-designed prospective studies are therefore required to establish the clinical efficacy of these approaches.

These studies, however, had several limitations. They were retrospective, and not every lesion detected by FAPI PET was biopsied; some lesions were considered positive based on follow-up images. The criteria for therapy response were not clearly defined. A partial response (PR) was characterized by a continuous decrease in levels of Tg, accompanied by a concurrent decrease in uptake values on PET/CT or structural imaging. An increase in structural size or the appearance of new lesions, along with a rising trend in Tg values, was classified as disease progression. Stable disease was defined as the absence of these changes. No RECIST criteria or Tg cut-off was established. Additionally, the studies had limited follow-up, and clinically important factors, such as the duration of response, were not clearly stated. Furthermore, currently available data remain investigational, and a properly designed prospective study is warranted for evaluating the efficacy of these theranostic approaches.

6.4. Astatine-211

Astatine-211 (Na[²¹¹At]At) is an alpha-emitting radioisotope produced by a cyclotron, with a half-life of 7.2 h. It shares similar properties with iodine [119,120,121]. Both Na[¹³¹I]I and Na[²¹¹At]At are monovalent halogen anions and are transported into thyrocytes. NIS is not substrate-selective—it also translocates Na[²¹¹At]At. Intracellular retention, however, depends on thyroid peroxidase (TPO)-mediated organification into thyroglobulin-bound organic compounds. Because Na[²¹¹At]At is not an efficient TPO substrate, it is taken up but poorly organified and therefore poorly retained, behaving more like [^99mTc]pertechnetate than Na[¹³¹I]I. However, its alpha particles impart significantly higher linear energy transfer compared to the beta emissions of Na[¹³¹I]I, leading to more efficient DNA double-strand breaks, enhanced colony formation inhibition, and improved tumor control, along with extended survival in preclinical thyroid cancer models [120,122].

Given the limitations of systemic therapies due to adverse effects, Na[²¹¹At]At may offer a valuable option for patients with disease progression following multiple Na[¹³¹I]I treatments. Watabe et al. recently conducted the first in-human Alpha-T1 trial, where they administered a single course of Na[²¹¹At]At to 11 patients with RAI-R DTC. This was done using a modified 3 + 3 activity-escalation design at doses of 1.25, 2.5, and 3.5 MBq/kg following rhTSH stimulation and a two-week iodine-restricted diet [123]. SPECT/CT confirmed that Na[²¹¹At]At was taken up by the expected organs, including the stomach and salivary glands (2.75 ± 1.46 and 1.96 ± 0.80 mGy/MBq, respectively). Red marrow exposure was lower, at 0.57 ± 0.15 mGy/MBq. Hematologic toxicity emerged as the main dose-limiting factor; at 3.5 MBq/kg, half of the patients experienced grade 3 lymphopenia or leukopenia lasting more than one week, though recovery was typically complete by eight weeks. Common side effects included salivary gland swelling, xerostomia, nausea, vomiting, and decreased appetite, which were mostly manageable. The use of ice packs to cool the salivary glands helped alleviate swelling, although its effect on xerostomia remains uncertain. Preliminary efficacy was promising. About a third of patients who received 2.5 or 3.5 MBq/kg showed a reduction of more than 50% in thyroglobulin, and Na[¹³¹I] SPECT demonstrated corresponding reductions in tracer uptake. Computed tomography-based Response Evaluation Criteria in Solid Tumors (RECIST) assessment at six months revealed stable disease in nine out of ten evaluable patients, most of whom had bone metastases or small lung metastases that were difficult to assess using conventional size-based criteria. In Na[¹³¹I] SPECT, one of three patients who received 2.5 MBq/kg achieved a partial response, while the other two had stable disease. Among the five evaluable patients who received 3.5 MBq/kg, one achieved a complete response with the disappearance of uptake in a skull metastasis, one showed a partial response, one had progressive disease, and two had stable disease. Based on the balance of toxicity and early efficacy signals, the investigators recommended 2.5 MBq/kg as the preferred dose for future phase 2 trials, while acknowledging that multiple cycles, similar to the multi-cycle regimens used in PSMA-targeted alpha therapy, may be required to optimize therapeutic outcomes [124]. The key practical advantage of Na[²¹¹At]At is radiation-protection simplicity, driven primarily by the decay scheme: alpha decay is accompanied by low-abundance characteristic X-rays (the 77–92 keV polonium X-rays can, incidentally, be used for post-therapy SPECT imaging). This reduces the external dose to contacts and healthcare staff to levels that are manageable in an outpatient setting, so admission to dedicated isolation rooms—often a logistical bottleneck that delays Na[¹³¹I]I therapy—is not required in most jurisdictions. The short 7.2 h physical half-life is a complementary advantage that limits residual patient activity rapidly, but is not, on its own, the primary driver of outpatient feasibility [125]. There is growing interest in combining Na[²¹¹At]At with redifferentiation strategies, such as BRAF or MEK inhibitors, although this approach will require careful management of overlapping hematologic toxicity [123].

7. Discussion

Despite substantial therapeutic advances, outcomes in RAI-R DTC remain suboptimal. Disease stabilization is often transient; a clear survival benefit has yet to be consistently demonstrated; and—of particular clinical relevance—MKI systemic therapies are associated with a significant burden on patients’ quality of life due to adverse events [126,127]. In this context, current research is increasingly focused on developing more selective, targeted agents to improve efficacy while minimizing off-target toxicity. Importantly, patients with advanced disease who are not receiving systemic therapy—apart from TSH suppression—may maintain a good quality of life for prolonged periods. This underscores the need to balance therapeutic aggressiveness with clinical prudence, integrating evolving scientific evidence with individualized management and shared decision-making. Within this framework, Na[¹³¹I] remains, to date, the only potentially curative option in advanced thyroid carcinoma [128]. Accurate identification of RAI-R disease is therefore crucial in order to avoid both premature discontinuation of a potentially effective treatment and the continuation of ineffective Na[¹³¹I]I therapy. The integration of multimodal imaging, histopathology, and molecular profiling is expected to improve phenotypic characterization and guide more appropriate therapeutic strategies. Accordingly, the management of RAI-R DTC should be conceptualized as a dynamic, stepwise process integrating clinical, molecular, and theranostic parameters [126,127] (Figure 5).

The sequence proposed in Figure 5—with selective inhibitors considered ahead of MKIs when an actionable driver is identified—departs from some guidelines (e.g., ESMO), which position lenvatinib and sorafenib as the established first-line systemic options in RAI-R DTC [129]. We present this ordering as a pragmatic framework that reflects the growing body of data on BRAF- and NTRK-directed agents and the possibility of subsequent redifferentiation. We acknowledge that, e.g., BRAF inhibition with dabrafenib (with or without trametinib) is not synonymous with redifferentiation; restoration of RAI avidity is a potential downstream consequence. Furthermore, irrespective of systemic therapy selection, locoregional treatments remain central to the management of locally progressive disease and should be considered throughout the pathway to control oligometastatic or symptomatic lesions and to prevent local complications, in line with the 2025 ATA guidelines [24].

Figure 5 does not include a formal economic dimension, which we accept as a limitation. The cost impact of each step is bidirectional: shorter kinase inhibitor exposure lowers both drug- and toxicity-related expenditure—the latter often underestimated, covering dose adjustments, supportive care, and admissions—while pretherapeutic dosimetry adds upfront cost. Prospective evaluation of the algorithm should therefore include cost-effectiveness and budget-impact endpoints alongside clinical outcomes.

After establishing RAI-R status, patients should be stratified by disease kinetics and symptom burden to guide the choice between active surveillance and systemic intervention. Central to this approach is early molecular profiling, which enables a personalized therapeutic algorithm: patients harboring actionable alterations (e.g., BRAF, RAS, RET, NTRK) may be candidates for redifferentiation strategies to restore iodine avidity, whereas those without targetable drivers are generally candidates for MKI therapy. In metastatic disease that has lost iodine avidity, redifferentiation uses a short course of therapy, e.g., dabrafenib plus trametinib in BRAF V600E tumors and trametinib in RAS-mutant disease, to restore NIS expression and re-enable Na[¹³¹I]I therapy. In phase II trials, uptake was restored in a proportion of patients, and those who reached the dosimetric threshold mainly achieved partial responses or disease stabilization, although durable complete responses were rare. The aim in incurable disease is palliative: reducing burden, controlling symptoms, and deferring continuous MKI therapy by delivering treatment in discrete cycles. Given the small, non-randomized evidence base, this approach suits selected patients and, where possible, prospective studies. The question is whether uptake can deliver a tumoricidal dose within normal-tissue limits. Blood- and marrow-based dosimetry defines the safe ceiling, classically under 2 Gy to the blood, with tighter limits in diffuse pulmonary metastases. Lesion dosimetry from Na[¹²⁴I]I PET/CT assesses whether each target reaches a therapeutic dose, typically defined as 20 Gy. Taprogge et al. found lesion-level doses ranging from below 1 Gy to over 1000 Gy within a single patient, which supports lesion-level decision-making [44]. The decision to treat should rest on post-redifferentiation dosimetry under the stimulation conditions intended for therapy. Avidity can be regained and lost again, so refractory status should be reassessed over time. Na[¹²⁴I]I PET dosimetry is resource-intensive and not yet standardized, and the 20 Gy threshold rests on limited data. Subsequent Na[¹³¹I]I is therefore best reserved for selected lesions with verified dosimetry rather than given as routine retreatment. The theranostic principle applies most directly to patients with distant metastases, where it can guide both the choice and the intensity of treatment. In metastatic disease that retains iodine avidity, lesional and whole-body dosimetry supports personalized Na[¹³¹I]I activity rather than fixed empiric dosing, maximizing absorbed tumor dose while respecting marrow and pulmonary constraints—particularly relevant in diffuse pulmonary or extensive osseous metastases.

Redifferentiation should not delay active systemic therapy in patients with high, rapidly growing tumor burden requiring immediate disease control, or in those with TERT promoter co-mutation and high [¹⁸F]FDG avidity, where dedifferentiation is advanced and response to MAPK inhibition is unlikely to restore clinically useful iodine handling [130,131]. In parallel, emerging non-radioiodine theranostic approaches are gaining increasing attention. These include PRRT, PSMA, and FAPI-targeted radioligand therapy, as well as alpha-emitter–based strategies such as Na[²¹¹At]At. In each scenario, imaging functions as a companion diagnostic for patient selection and response monitoring, favoring a stratified, lesion-level strategy over uniform systemic treatment. While these approaches are biologically appealing and may offer a more favorable toxicity profile than conventional systemic therapies, current evidence remains limited and largely derived from small, retrospective, and monocentric studies. Among these, Na[²¹¹At]At appears particularly promising, although larger, more robust prospective data are still lacking. Response assessment should rely on an integrated evaluation that combines imaging (RECIST, [¹⁸F]FDG PET/CT, post-therapy whole-body scan), biochemical markers, and clinical parameters, enabling continuous adaptation of management within a precision oncology paradigm [22,132,133,134].

A distinct problem arises in progressive RAI-R DTC when MKIs are not an option, whether because of comorbidity, cumulative toxicity, progression through the available agents, or patient preference. The conventional algorithm then narrows to locoregional measures, and a target-selected theranostic option becomes attractive precisely because it can be delivered in discrete cycles rather than as continuous therapy. The premise is that imaging should define the target before any radioligand is administered. Multi-tracer PET makes this phenotyping possible. Na[¹²⁴I]I PET/CT quantifies residual iodine avidity and supports lesional dosimetry, identifying patients in whom high-activity Na[¹³¹I]I, with or without preceding redifferentiation, remains feasible. [¹⁸F]FDG PET/CT characterizes glycolytic, dedifferentiated disease; intense and diffuse [¹⁸F]FDG uptake indicates aggressive biology and predicts a low probability of successful redifferentiation, thereby steering selection away from iodine-based strategies. Where iodine biology is absent, receptor- and stroma-directed tracers map alternative targets, e.g., [⁶⁸Ga]-labeled somatostatin analogs for SSTR-expressing lesions amenable to PRRT, and, more experimentally, e.g., [⁶⁸Ga]-PSMA and [⁶⁸Ga]-FAPI for neovascular and stromal targets that may be addressed with matched therapies. These tracers expose the intra- and inter-lesional heterogeneity characteristic of RAI-R DTC, and a discordant target map may favor a lesion-directed or sequential approach over a single uniform treatment. The therapeutic radionuclide then mirrors the diagnostic tracer, preserving the theranostic pair.

The evidence reviewed here supports a clear hierarchy of readiness for clinical use. Multikinase inhibitors—lenvatinib and, to a lesser extent, sorafenib—remain the standard first-line systemic option for progressive, symptomatic RAI-R DTC, supported by phase III randomized data. In molecularly selected patients, genotype-matched selective inhibitors deliver higher response rates and better tolerability than MKIs and should be prioritized whenever a targetable alteration is identified on next-generation sequencing, e.g., BRAF plus MEK inhibition for BRAF V600E disease, selpercatinib or pralsetinib for RET-altered disease, and larotrectinib or entrectinib for NTRK fusions. For clinical practice, it would also be important to clarify whether mutation status can be reliably assessed on archival thyroidectomy specimens or if repeat tumor sampling closer to the initiation of systemic therapy is recommended; current pragmatic consensus suggests that archival tissue is generally adequate, although repeat biopsy-based profiling may be considered in advanced disease or prior to targeted therapy to account for potential molecular evolution. Every other theranostic strategy discussed in this review—redifferentiation outside clinical trials, PRRT in SSTR-expressing cases, PSMA-directed radioligand therapy, FAPI-based theranostics, and alpha-emitter-based approaches, including Na[²¹¹At]At—remains investigational. Early signals are encouraging, and the biological rationale is sound, but the current evidence base is dominated mostly by small, retrospective, often monocentric series with heterogeneous response criteria and short follow-up. Translating these approaches to practice will require adequately powered prospective multicenter trials with standardized patient selection, harmonized imaging and dosimetry, and clinically meaningful endpoints. Until such data are available, these strategies should be offered to patients preferentially within clinical trials, rather than as routine alternatives to established therapies.

8. Conclusions

Despite ongoing research efforts, clinical outcomes in advanced RAI-R DTC remain unsatisfactory; therefore, management should be approached with caution. First, patients may maintain an excellent quality of life for years with TSH suppression alone, supporting a conservative approach in selected cases. Second, currently available MKI systemic therapies provide modest benefit and are associated with significant toxicity, requiring careful patient selection and timing of initiation. Third, alternative therapeutic strategies—including redifferentiation approaches and non-radioiodine theranostics—remain investigational and are not yet supported by sufficient evidence for routine use. Nevertheless, in this complex and heterogeneous clinical scenario, such strategies may be considered in selected patients, with particular regard to redifferentiation approaches, while awaiting results from ongoing prospective studies.