Medullary thyroid carcinoma (MTC) is a rare, aggressive neuroendocrine tumor with limited treatment options and frequent recurrence. Comprehensive recurrence risk stratification remains lacking. Here, we profile 482 MTC samples from 452 patients across ten Chinese clinical centers, identifying 10,092 proteins and mutations in 87.0% of patients. Clinically, MTC grading, concurrent papillary thyroid carcinoma, and lymph node metastasis are significant recurrence risk factors, whereas at the genetic level, RET M918T and RET S891A mutations are correlated with high recurrence risk in sporadic and hereditary MTC, respectively. Ubiquitinomics show downregulated E3 ligases CUL4B and TRIM32 are associated with structural recurrence. We define three molecular subtypes with distinct outcomes and present an integrative machine learning model combining clinical, genomic, and proteomic features, validated in an independent test dataset of 105 patients and a published dataset. This multi-center, multi-omics study enhances the understanding of MTC heterogeneity and facilitates personalized patient management.

The paucity of public awareness regarding thyroid eye disease (TED) usually leads to delayed medical care. While large language models (LLMs) hold great potential for augmenting patient education, their ability in answering TED-related questions has yet to be comprehensively evaluated. This study aims to assess the capability of LLMs to address TED-related questions and explore the practicability of customizing LLMs for disease-specific domains. Considering the diverse LLM candidates, we deployed a cascade pipeline to search for the best model for TED. We first evaluated performances of several prevailing LLMs on multiple-choice questions. The best-performing models, GPT-4 and Claude 3.5, were selected and customized to create TED-GPT and TED-Claude. Chain-of-Thought (CoT) was then utilized, resulting in CoT-GPT and CoT-Claude. We also evaluated newer LLMs with native CoT capabilities (GPT-4-o1, GPT-4-o3, Gemini-2.0-Flash, Gemini-2.5-Pro, Claude 3.7). These models, along with their original versions, were then assessed and compared on multiple-choice questions. The better-performing TED-GPT and TED-Claude were evaluated on short-answer and case questions, with comparisons made to their original ones using the QUEST framework (Quality, Understanding/Reasoning, Expression, Safety/Harm, Trust). For multiple-choice questions, GPT-4 and Claude 3.5 arrived competitive accuracies (76.2% and 83.2%, respectively). The addition of CoT, along with customization into GPT-4 and Claude 3.5, led to improvement in the accuracy (CoT-GPT 86.1%, CoT-Claude 87.1%, TED-GPT 86.1%, TED-Claude 89.1%), outperforming all other newer LLMs. For case and short-answer questions, the customized TED-GPT and TED-Claude also performed better than their original versions. TED-Claude showed the best performance in accuracy, readability, comprehensiveness, likelihood of harm and reasoning. Thus, LLMs, particularly TED-Claude, achieved relatively satisfactory performances in answering TED-related questions. Besides, using LLMs’ customization modules, along with CoT, effectively enhanced model performances. This indicates that clinicians can use these simple and universal methods to construct LLMs suitable for specific medical domains.

Thyroid eye disease (TED) is a prevalent autoimmune orbital disorder that significantly impairs visual function and appearance, adversely affecting patients’ quality of life. The development of effective preventive strategies is paramount to optimizing clinical outcomes and disease management. Within the paradigm of predictive, preventive, and personalised medicine (3PM), the identification of modifiable risk factors for TED and the formulation of individualized prevention protocols are critical for targeted risk minimization. However, despite advances in understanding TED pathogenesis, the susceptible populations of TED remain inadequately defined, resulting in generalized and often ineffective preventive measures. Furthermore, while several risk factors for TED development have been identified, particularly tobacco exposure, radioiodine treatment and thyroid dysfunction, their differential contributions across distinct disease courses (onset, progression, and recurrence) remain poorly elucidated. This review comprehensively summarizes current clinical evidence on susceptible populations for TED and classifies modifiable risk factors into five categories: physical and chemical factors, endocrine and autoimmune factors, metabolic and nutritional factors, social and psychological factors, and other factors, with updated clinical evidence of their impacts on disease course. Building upon contemporary clinical evidence, we present a stratified strategy for TED prevention spanning onset, progression, and recurrence that tailored to disease phase-specific vulnerabilities. Early identification and mitigation of modifiable risk factors are pivotal to transitioning TED management from reactive medical services to a 3PM paradigm, ultimately reducing disease incidence, improving clinical outcomes and patients’ quality of life.

Accurate preoperative diagnosis of thyroid nodules via fine-needle aspiration (FNA) biopsy remains challenging, particularly in cases with indeterminate cytology. This prospective, noninterventional, blinded, multicenter study establishes ThyroProt, a diagnostic classifier that integrates targeted mass-spectrometry-based quantification of a 3-protein signature with BRAFV600E mutation status, age, and gender. Developed and validated on 837 FNA samples, the classifier is evaluated in a prospective test set of 322 samples, achieving an area under the curve (AUC) of 0.94 with an overall accuracy of 90.7%. For the critical subgroup of Bethesda III/IV nodules, ThyroProt demonstrates an accuracy of 88.0%, with 82.4% sensitivity and 100% specificity. The classifier's robust performance is further evaluated in two independent multicenter cohorts, where it maintains an AUC of 0.87-0.91 and an accuracy of 84.3%-85.7%. This study supports the clinical utility of mass-spectrometry-based targeted proteomics for improving preoperative diagnosis of thyroid nodules, particularly those with indeterminate cytology.

Autoimmune thyroid disease (AITD) is associated with adverse health outcomes, yet routine thyroid autoantibody testing may not always be practical in general screening settings. The pan-immune-inflammation value (PIV), derived from complete blood counts, represents a simple, low-cost marker of systemic inflammation. This study investigated the association between PIV and thyroid autoantibody positivity in a health examination cohort. This cross-sectional study included 5,785 Chinese adults undergoing routine health examinations at the First Hospital of China Medical University. Associations between PIV and anti-thyroglobulin (TgAb) or anti-thyroid peroxidase (TPOAb) antibody positivity were evaluated using multivariable logistic regression and restricted cubic spline (RCS) models. Subgroup analyses explored interactions with demographic and metabolic factors. RCS analyses revealed significant non-linear associations between PIV and TgAb (P for non-linearity = 0.002) as well as TPOAb (P = 0.043) positivity. In fully adjusted models, the highest PIV quartile was associated with an increased risk of TgAb positivity compared to the lowest quartile (OR 1.23, 95

BACKGROUND:Microplastics (MPs) have been identified in multiple human tissues and are increasingly implicated in systemic health risks. Their presence in the thyroid gland, however, remains unexamined. Autoimmune thyroiditis (AIT) is the most frequent autoimmune thyroid disorder and the leading cause of hypothyroidism. This study aims to detect the presence of MPs in the thyroid and their potential relevance to AIT. METHODS:In this case-control study, thyroid tissues were obtained from 29 patients with histologically confirmed AIT and 29 age- and sex-matched non-AIT controls who underwent thyroidectomy due to thyroid nodules. MP burden was quantified by pyrolysis-gas chromatography-mass spectrometry (Py-GC/MS). Particle-level polymer identity and particle characteristics, including size, shape, and color, were assessed using micro-Raman spectroscopy, whereas scanning electron microscopy (SEM) was employed for morphological observation. RESULTS:MPs were detected in thyroid tissues from both groups. Py-GC/MS revealed significantly higher total MP concentrations in the AIT group compared to controls (median: 19.9 vs. 1.9 μg/g; p=0.012). This elevation was primarily driven by polyvinyl chloride (PVC), which was significantly higher in AIT patients. Micro-Raman spectroscopy identified particles ranging from 33.9 to 1467 µm. The AIT group contained significantly increased MPs abundance compared with the non-AIT control group (172 vs. 50.2 items/g, p=0.037). Morphological profiling revealed no significant differences in the size, shape and color of MPs between groups. CONCLUSION:An increased MPs burden with the particular enrichment of PVC was observed in patients with AIT, suggesting a potential association between environmental MPs exposure and thyroid autoimmunity. Further mechanistic and epidemiological studies to clarify the impacts of chronic MPs exposure are needed.

BACKGROUND:Thyroid disease in pregnancy, preconception, and postpartum is a common and clinically relevant problem. Since the publication of the American Thyroid Association (ATA) guidelines in 2017, substantial new clinical and scientific evidence has become available. The aim of these guidelines is to provide clinicians, patients, researchers, and policymakers with evidence-based recommendations on the care of women with thyroid disease before, during, and after pregnancy. METHODS:The clinical questions addressed were informed by prior ATA guidelines, stakeholder feedback, a global needs assessment, and input from the multidisciplinary task force. Systematic literature searches were conducted with the support from a medical librarian and evaluated using the Grading of Recommendations, Assessment, Development, and Evaluation framework. Recommendations were formulated based on the quality of evidence, balance of benefits and harms, patient values, feasibility, and equity. Where data were limited, Good Practice Statements were formulated. The task force included representatives from 10 international societies as well as patient advocacy groups and a methodologist. RESULTS:The updated guidelines include recommendations on thyroid function testing, iodine supplementation, thyroid autoimmunity, hypothyroidism, hypothyroxinemia, hyperthyroidism and Graves' disease, thyroid nodules and cancer, and postpartum thyroid dysfunction for women with infertility, pregnant women, and women during postpartum and/or lactation. Recommendations are presented using recommendation tables, additional practical considerations are highlighted in boxes, and background information is provided in the text, tables, and figures per disease entity and chronological subset. CONCLUSIONS:These 2026 ATA guidelines provide updated, evidence-based recommendations for the diagnosis and management of thyroid disease in women during preconception, pregnancy, and postpartum. While acknowledging that much of the evidence remains of low-to-moderate quality, these guidelines represent current best practices and consensus among international experts from different fields, offering an optimized framework for individualized patient care.

Background and Objectives Congenital hypothyroidism (CH) is the most common neonatal endocrine disorder with largely elusive underlying mechanisms, although thyroid dysformation has been deemed as the most frequent cause. Methyltransferase like 3 (METTL3) serves as a pivotal writer for N6-methyladenosine (N6-methyladenosine, m6A) required for various organ development, but little is known about the significance of METTL3 and m6A modification in thyroid formation, in CH either. In this study, we aimed to clarify the new regulatory role of METTL3 in the occurrence and development of CH, and to provide new theoretical support and treatment ideas for the clinical treatment of CH.Methods Thyrocyte-specific Mettl3 knockout mouse model was constructed and subjected to morphological and functional analyses. Representative differentiation, polarization, and hormone synthesis factors were studied via immunohistochemistry, immunofluorescence staining, RT-qPCR, and thyroid hormone in serum were quantified. In vitro, function of Mettl3 and molecular mechanisms were further investigated through thyrocyte cells from different species via lentivirus mediated silencing and rescue experiments.Results Thyrocyte specific removal of Mettl3 caused a typical CH phenotype, with reduced thyroid hormones and body weight. Histologically, the thyroid follicle of Mettl3 deficient mice appeared as abnormally fused and enlarged structure, with significantly disturbed polarity and patterning. Mechanistically, Pax8 expression was reduced upon METTL3 loss due to damaged m6A modification, which resulted in compromised thyroid epithelial cell polarization, differentiation and hormone synthesis.Conclusions Mettl3 functions as a key player of thyroid folliculogenesis and hormone secretion by coordinating thyrocyte polarization and differentiation progression, and its deficiency may lead to congenital hypothyroidism.

Importance:Thyroid eye disease (TED), a disfiguring and potentially sight-threatening condition with racial phenotypic variations, currently has limited effective treatments. Insulin-like growth factor 1 receptor (IGF-1R) inhibitors therapy has emerged as a promising treatment option, although it remains less accessible and lacks substantial evidence in Asian patients. Objective:To assess efficacy and safety of IBI311, an IGF-1R inhibitor with an identical amino acid sequence to teprotumumab but a different dosage form, in Chinese patients with active TED. Design, Setting, and Participants:This was a randomized, double-masked, placebo-controlled, multicenter, 24-week phase 3 trial with recruitment conducted across 20 tertiary hospitals in China from May to December 2023. Chinese participants with active (clinical activity score [CAS] ≥3) moderate to severe TED were included after excluding individuals with active TED onset over 270 days; sight-threatening TED; or history of steroid pulse therapy, radiotherapy, or surgery for TED. Interventions:Eighty-two participants were randomized 2:1 to receive intravenous infusions of either IBI311 or placebo once every 3 weeks for 21 weeks with follow-up through week 24. Main Outcomes and Measures:The primary outcome was the proptosis response rate (proptosis reduction ≥2 mm) in the study eye at week 24. Results:Participants (mean [SD] age, 39.6 [10.9] years; 56 [68.3%] women) were randomized to receive IBI311 (n = 54) or placebo (n = 28). At week 24, 45 of 52 participants receiving IBI311 (85.8%) and 1 of 26 receiving placebo (3.8%) had proptosis response (difference, 81.9 percentage points; 95% CI, 69.8 to 93.9; P < .001). The secondary outcomes included overall response (proptosis reduction ≥2 mm and CAS reduction ≥2, 80.2% vs 3.6%; difference, 76.3 percentage points; 95% CI, 63.3 to 89.4), CAS of 0 or 1 (83.5% vs 16.6%; difference, 67.1 percentage points; 95% CI, 49.4 to 84.8), least-squares mean (SE) change from baseline in proptosis (-2.85 [0.18] mm vs -0.02 [0.24] mm; difference, -2.83 mm; 95% CI, -3.39 mm to -2.27 mm) in the study eye (all P < .001), and diplopia response (diplopia reduction ≥1 grade, 66.0% vs 53.3%; P = .46). All adverse events of interest (infusion reaction, hearing impairment, hyperglycemia, muscle spasm, and nausea or diarrhea) were mild or moderate in severity. No serious adverse event or death occurred in the IBI311 group. Conclusions and Relevance:In this phase 3 randomized clinical trial, IBI311 demonstrated better and clinically relevant outcomes in proptosis and CAS than placebo with no new safety issues not identified in previous clinical trials. The results suggest that IBI311 represents a viable treatment option for Chinese patients with active TED. Trial Registration:ClinicalTrials.gov Identifier: NCT05795621.

CONTEXT:While the association between maternal thyroid peroxidase antibody (TPOAb) positivity and preterm birth (PTB) risk has been established, the association between thyroglobulin antibody (TgAb) and PTB remains unclear. OBJECTIVE:This study aimed to explore the association between TgAb and PTB risk in euthyroid women. METHODS:This single-center, prospective cohort study enrolled euthyroid women in the first trimester. Data on serum concentrations of thyrotropin (TSH), free thyroxine (FT4), TgAb, and TPOAb were collected. Participants were categorized into 2 groups (TgAb-negative and TgAb-positive). PTB was subtyped into spontaneous PTB (S-PTB) and medically induced PTB (MI-PTB); and into early PTB (E-PTB) and late PTB (L-PTB). Logistic regression models examined the associations between TgAb and PTB and its subtypes, with stratification by first-trimester TSH levels (0.1-2.5 mIU/L, 2.5-4.0 mIU/L). RESULTS:This study comprised 58 247 euthyroid pregnant women. Adjusting for confounders, TgAb positivity was associated with a 16% increased risk of PTB (adjusted odds ratio (aOR) 1.16, 95% CI, 1.03-1.29; P = .01) compared to the TgAb-negative group. Specifically, TgAb positivity showed a higher risk of S-PTB and L-PTB, aOR 1.22 (95% CI, 1.06-1.39) and aOR 1.17 (95% CI, 1.04-1.32), respectively. Consistent results were observed when analyzing TgAb concentration as a continuous variable. TSH stratification analysis revealed that these associations were statistically significant only among women with TSH levels between 0.1 and 2.5 mIU/L. CONCLUSION:In euthyroid women, TgAb positivity was associated with a higher risk of PTB that mainly manifested as S-PTB and L-PTB. However, the clinical significance of these findings is limited.

Free triiodothyronine (FT3) exerts a significant influence on glucose metabolism. The relationship between gestational diabetes mellitus (GDM) and FT3 during pregnancy is complex and inconsistently reported. Our study aims to explore the bidirectional association between FT3 during pregnancy and GDM, and to assess whether this association is causal. The observational analysis included two clinical studies. Study 1 involved 6,221 pregnant women and applied multivariate logistic regression analysis to investigate the association between FT3 in early pregnancy and the subsequent risk of GDM. Study 2 comprised 387 pregnant women and employed linear regression analysis to examine the impact of GDM on FT3 in late pregnancy. Additionally, genome-wide association study (GWAS) summary statistics of FT3 and GDM were used to perform a bidirectional two-sample Mendelian randomization (MR) analysis to test for causal associations. In Study 1, after adjusting for potential confounding factors, increased FT3 levels in early pregnancy were associated with the subsequent risk of GDM [odds ratio (OR) 1.122; 95

Thyroid eye disease (TED), a disfiguring and potentially sight-threatening condition with racial phenotypic variations, currently has limited effective treatments. Insulin-like growth factor 1 receptor (IGF-1R) inhibitors therapy has emerged as a promising treatment option, although it remains less accessible and lacks substantial evidence in Asian patients. To assess efficacy and safety of IBI311, an IGF-1R inhibitor with an identical amino acid sequence to teprotumumab but a different dosage form, in Chinese patients with active TED. This was a randomized, double-masked, placebo-controlled, multicenter, 24-week phase 3 trial with recruitment conducted across 20 tertiary hospitals in China from May to December 2023. Chinese participants with active (clinical activity score [CAS] ≥3) moderate to severe TED were included after excluding individuals with active TED onset over 270 days; sight-threatening TED; or history of steroid pulse therapy, radiotherapy, or surgery for TED. Eighty-two participants were randomized 2:1 to receive intravenous infusions of either IBI311 or placebo once every 3 weeks for 21 weeks with follow-up through week 24. The primary outcome was the proptosis response rate (proptosis reduction ≥2 mm) in the study eye at week 24. Participants (mean [SD] age, 39.6 [10.9] years; 56 [68.3%] women) were randomized to receive IBI311 (n = 54) or placebo (n = 28). At week 24, 45 of 52 participants receiving IBI311 (85.8%) and 1 of 26 receiving placebo (3.8%) had proptosis response (difference, 81.9 percentage points; 95% CI, 69.8 to 93.9; P < .001). The secondary outcomes included overall response (proptosis reduction ≥2 mm and CAS reduction ≥2, 80.2% vs 3.6%; difference, 76.3 percentage points; 95% CI, 63.3 to 89.4), CAS of 0 or 1 (83.5% vs 16.6%; difference, 67.1 percentage points; 95% CI, 49.4 to 84.8), least-squares mean (SE) change from baseline in proptosis (−2.85 [0.18] mm vs −0.02 [0.24] mm; difference, −2.83 mm; 95% CI, −3.39 mm to −2.27 mm) in the study eye (all P < .001), and diplopia response (diplopia reduction ≥1 grade, 66.0% vs 53.3%; P = .46). All adverse events of interest (infusion reaction, hearing impairment, hyperglycemia, muscle spasm, and nausea or diarrhea) were mild or moderate in severity. No serious adverse event or death occurred in the IBI311 group. In this phase 3 randomized clinical trial, IBI311 demonstrated better and clinically relevant outcomes in proptosis and CAS than placebo with no new safety issues not identified in previous clinical trials. The results suggest that IBI311 represents a viable treatment option for Chinese patients with active TED. ClinicalTrials.gov Identifier: NCT05795621

Disclosure: J. Fu: None. S. Zhang: None. Q. Ye: None. H. Guan: None. Introduction: Autoimmune thyroiditis (AIT) is characterized by a loss of immune tolerance to thyroid antigens, leading to follicular cell damage and, in severe cases, the development of hypothyroidism. In recent years, dietary factors have gained attention for their potential role in triggering or exacerbating AIT. Time-restricted eating (TRE), a form of intermittent fasting, has shown promise in improving metabolic, immune, and inflammatory conditions. However, the effects of dietary factors on AIT and its underlying mechanisms remain poorly understood. Methods: We aimed to investigate the effects of TRE on thyroid inflammation and gut microbiota in AIT using C57BL/6 mice immunized with thyroglobulin (Tg) and fed with different fasting regimens. The mice were then divided into four groups: no fasting (Tg0), 12-hour fasting (Tg12), 16-hour fasting (Tg16), and alternate-day fasting (Tg24), with unimmune C57BL/6 mice serving as the control (blank) group. Thyroid autoantibody levels were measured using ELISA, thyroid and colon histology were assessed through H&E staining and AB-PAS staining, and flow cytometry was employed to analyze the distribution of T lymphocyte subsets in the spleen. The expression of colonic mucus markers MUC2 and ZO1 were evaluated by immunohistochemistry (IHC). Additionally, 16S rRNA sequencing was conducted to investigate the gut microbiota composition in fecal samples. Results: After 4 weeks of TRE, compared to the Tg0 group, the levels of TPOAb in the TRE groups were significantly decreased, with Tg12 group showed the lowest levels of TPOAb (252.7 vs 163.5 mU/L, P<0.05). The proportion of Th2 cells in the Tg12, Tg16, and Tg24 groups was significantly lower than that in the Tg0 group, while the proportion of Th17 cells was notably higher in the Tg12 and Tg16 group. AB-PAS staining showed a reduction in mucus and goblet cells in AIT mice, with a significant decrease in goblet cells in the Tg16 group compared to the Tg0 group. The levels of MUC2 and ZO1 were significantly lower in the Tg16 group compared to the Tg0 group. For gut microbiota analysis, the Tg16 group showed a composition more similar to the blank group, with an increase in Bacteroidota and a decrease in Firmicutes compared to the Tg0 group. LDA score analysis highlighted significant enrichment of Lactobacillus, Bacteroidaceae, and Bacteroides in the Tg16 group, indicating notable microbial differences across the groups. Conclusion: These results indicate TRE reduced the production of thyroid autoantibodies and Th2 cells in AIT. The 16:8 time-restricted feeding regimen induced a gut microbiota profile in AIT mice similar to that of the blank group, while further investigation should examine whether alterations in gut microbiota mediates the impact of TRE on AIT. Presentation: Monday, July 14, 2025

OBJECTIVE:To investigate the linear and nonlinear relationships between the triglyceride-glucose body mass index (TyG-BMI) and aggressiveness and risk of recurrence in papillary thyroid carcinoma (PTC). METHODS:This retrospective single-center cohort study included 11 317 patients with PTC. The associations between the TyG-BMI and PTC aggressiveness as well as moderate-to-high recurrence risk were analyzed with binary logistic regression and odds ratios (ORs). Linear and nonlinear relationships between the TyG-BMI and these outcomes were evaluated with restricted cubic spline and smoothed curve-fitted logistic risk regression models. Key factors contributing to TyG-BMI prediction outcomes were weighted with machine learning algorithms. RESULTS:After adjusting for confounding factors, higher TyG-BMI was associated with a significantly increased risk for tumors with a maximum diameter of >1 cm (OR adjust = 1.35, P < 0.001), multifocality (OR adjust = 1.42, P < 0.001), and extrathyroidal extension (OR adjust = 1.53, P < 0.001). Conversely, higher TyG-BMI was associated with a significantly decreased risk for positive lymph nodes with a maximum diameter of >0.2 cm (OR adjust = 0.36, P < 0.001) and intermediate-to-high risk of PTC recurrence (OR adjust = 0.68, P < 0.001). TyG-BMI exhibited a linear relationship with the risk of tumors with a maximum diameter of >1 cm and multifocality, but a nonlinear relationship with extrathyroidal extension and intermediate-to-high recurrence risk of PTC. TyG-BMI showed a positive linear correlation with free triiodothyronine (FT3) and thyroglobulin (Tg); a negative linear correlation with free thyroxine (FT4), thyroid peroxidase antibody (TPOAb), and thyroglobulin antibody (TgAb); and no linear relationship with thyroid-stimulating hormone (TSH). Among the components of the TyG-BMI and potential confounding factors, machine learning algorithms consistently identified triglyceride (TG) level as the primary contributor when predicting PTC aggressiveness and intermediate-to-high risk of PTC recurrence. CONCLUSION:This study reveals complex relationships, both linear and nonlinear, between the TyG-BMI, PTC aggressiveness and intermediate-to-high risk of PTC recurrence, with TG playing a pivotal role within the TyG-BMI. There were linear correlations between the TyG-BMI and thyroid function.

OBJECTIVE:In recent years, the application of deep learning (DL) technology in the thyroid field has expanded rapidly, driving substantial innovation in thyroid disease research. This review aims to provide clinicians with the latest research advances in the application of DL to the diagnosis and treatment of thyroid cancer. METHODS:A systematic review was conducted of studies published in the past five years in the PubMed database on the application of deep learning in the diagnosis, treatment, and prognosis of thyroid cancer. RESULTS:DL has made substantial advances in the diagnosis and treatment of thyroid cancer, particularly through the application of advanced models such as convolutional neural networks, long short-term memory networks, and generative adversarial networks. These models have delivered breakthrough performance in key areas, including ultrasound image analysis of thyroid nodules, automated classification of pathological images, and assessment of extrathyroidal extension. DL also shows considerable promise for individualized treatment planning and prognosis prediction. Nonetheless, its widespread clinical adoption is hindered by substantial technical, clinical, and ethical challenges. Addressing these barriers is crucial to achieving meaningful improvements in thyroid cancer care and realizing the full potential of DL in precision medicine. CONCLUSION:DL techniques are advancing the precision diagnosis and treatment of thyroid cancer and hold the potential to enhance diagnostic accuracy and improve therapeutic outcomes for patients.

CONTEXT:Soluble immune checkpoints play an important role in peripheral tolerance that has seldom been investigated in Graves' disease (GD) and thyroid eye disease (TED). OBJECTIVE:The objective of this work is to examine the alteration of soluble immune checkpoints in GD and TED. METHODS:We performed a quantitative multiplex analysis of 17 immune checkpoint proteins in serum from 50 GD patients without TED, 28 GD patients with TED, and 40 healthy controls. The association with demographic, serologic, clinical features and 27 cytokines was analyzed. A follow-up was conducted in GD patients without TED. Functional outcomes of sLAG-3 and sGITR were assessed in cell cultures using rh-LAG3, rh-GITR, an antagonistic LAG-3 antibody, and an antagonistic GITR antibody. RESULTS:GD Patients with TED had distinct sICP and cytokine profiles compared with GD patients without TED. Active patients with TED exhibited elevation in the levels of sBTLA, sLAG-3, sGITR, sCD80, sCD86, and sPD-L1. Further, GD patients without TED with high sBTLA, sCD27, and sCD40 levels at baseline showed a better improvement in thyrotropin receptor antibody titers after antithyroid drug treatment. Adding recombinant human GITR and LAG-3 to peripheral blood mononuclear cell cultures resulted in increased inflammatory cytokine secretion and decreased anti-inflammatory cytokine secretion. CONCLUSION:The present study uncovers disturbed soluble immune checkpoints and cytokines in GD patients with and without TED and may pave the way for novel immunological screening, allowing for identification of patients with TED at higher risk of developing active disease and patients with GD a better treatment response after antithyroid drug treatment.