Artificial Intelligence for Thyroid Nodule Characterization: Where Are We Standing?

Meaning of ACR-TIRADS recommendation in favor of follow-up rather than FNAC in thyroid nodules

Thyroid Imaging Reporting and Data Systems (TIRADSs) have been largely diffused for their high accuracy in risk stratification of thyroid nodules (TNs) and their selection for fine-needle aspiration cytology (FNAC). The most popular TIRADSs are ACR-, EU-, and K-TIRADS, with some discrepancies each other. One major difference is that ACR-TIRADS includes a recommendation in favor of follow-up in TNs having a major diameter insufficient to indicate FNAC. The present study aimed to explore prevalence and significance of this recommendation. EU- and K-TIRADS were used as comparator. A retrospective series of thyroidectomies was searched according to a pre-defined protocol. The study period was 2019-2023. Preoperative ultrasound images were reviewed by radiologists blinded of clinical data. Matching of TIRADS and histology was performed later. Histology was the gold standard. The study series included 39 TNs classified as category 3, 4, or 5 and assessed for follow-up according to ACR-TIRADS. The overall cancer frequency was 25.6%, being 13% in category 3, 20% in category 4, and 83.3% in category 5. The category assessment according to ACR-, EU-, and K-TIRADS was not significantly different. EU-TIRADS indicated FNAC in 10 TNs of which two cancers and eight benign lesions. K-TIRADS recommended FNAC in 32 TNs of which seven cancers and 25 benign lesions. TNs assessed for follow-up according to ACR-TIRADS are cancer in one-fourth of cases. EU- and, especially, K-TIRADS allow us to select for FNAC cancers, with the burden of non-negligible frequency of unnecessary FNACs.

The Utility of Contrast-Enhanced Ultrasound (CEUS) in Assessing the Risk of Malignancy in Thyroid Nodules

BACKGROUND

Ultrasonography is a primary method used in the evaluation of thyroid nodules, but no single feature of this method predicts malignancy with high accuracy. Therefore, this paper aims to assess the utility of contrast-enhanced ultrasound (CEUS) in the differential diagnosis of thyroid nodules.

METHODS

The study group comprised 188 adult patients (155 women and 33 men) who preoperatively underwent CEUS of a thyroid nodule classified as Bethesda categories II-VI after fine-needle aspiration biopsy. During the CEUS examination, 1.5 mL of SonoVue contrast was injected intravenously, after which 15 qualitative CEUS enhancement patterns were analysed.

RESULTS

The histopathologic results comprised 65 benign thyroid nodules and 123 thyroid carcinomas. The dominant malignant CEUS features, such as hypo- and heterogeneous enhancement and slow wash-in phase, were evaluated, whereas high enhancement, ring enhancement, and a slow wash-out phase were assessed as predictors of benign lesions. Two significant combinations of B-mode and CEUS patterns were noted, namely, hypoechogenicity with heterogeneous enhancement and non-smooth margins with hypo- or iso-enhancement.

CONCLUSIONS

The preliminary results indicate that CEUS is a useful tool in assessing the risk of malignancy of thyroid lesions. The combination of the qualitative enhancement parameters and B-mode sonographic features significantly increases the method's usefulness.

Head-to-head comparison of American, European, and Asian TIRADSs in thyroid nodule assessment: systematic review and meta-analysis

Context

Ultrasound-based risk stratification systems (Thyroid Imaging Reporting and Data Systems (TIRADSs)) of thyroid nodules (TNs) have been implemented in clinical practice worldwide based on their high performance. However, it remains unexplored whether different TIRADSs perform uniformly across a range of TNs in routine practice. This issue is highly relevant today, given the ongoing international effort to establish a unified TIRADS (i.e. I-TIRADS), supported by the leading societies specializing in TNs. The study aimed to conduct a direct comparison among ACR-, EU-, and K-TIRADS in the distribution of TNs: (1) across the TIRADS categories, and (2) based on their estimated cancer risk.

Methods

A search was conducted on PubMed and Embase until June 2023. Original studies that sequentially assessed TNs using TIRADSs, regardless of FNAC indication, were selected. General study characteristics and data on the distribution of TNs across TIRADSs were extracted.

Results

Seven studies, reporting a total of 41,332 TNs, were included in the analysis. The prevalence of ACR-TIRADS 1-2 was significantly higher than that of EU-TIRADS 2 and K-TIRADS 2, with no significant difference observed among intermediate- and high-risk categories of TIRADSs. According to malignancy risk estimation, K-TIRADS often classified TNs as having more severe risk, ACR-TIRADS as having moderate risk, and EU-TIRADS classified TNs as having lower risk.

Conclusion

ACR-, EU-, and K-TIRADS assess TNs similarly across their categories, with slight differences in low-risk classifications. Despite this, focusing on cancer risk estimation, the three TIRADSs assess TNs differently. These findings should be considered as a prerequisite for developing the I-TIRADS.

MV-MS-FETE: Multi-view multi-scale feature extractor and transformer encoder for stenosis recognition in echocardiograms

BACKGROUND

aortic stenosis is a common heart valve disease that mainly affects older people in developed countries. Its early detection is crucial to prevent the irreversible disease progression and, eventually, death. A typical screening technique to detect stenosis uses echocardiograms; however, variations introduced by other tissues, camera movements, and uneven lighting can hamper the visual inspection, leading to misdiagnosis. To address these issues, effective solutions involve employing deep learning algorithms to assist clinicians in detecting and classifying stenosis by developing models that can predict this pathology from single heart views. Although promising, the visual information conveyed by a single image may not be sufficient for an accurate diagnosis, especially when using an automatic system; thus, this indicates that different solutions should be explored.

METHODOLOGY

following this rationale, this paper proposes a novel deep learning architecture, composed of a multi-view, multi-scale feature extractor, and a transformer encoder (MV-MS-FETE) to predict stenosis from parasternal long and short-axis views. In particular, starting from the latter, the designed model extracts relevant features at multiple scales along its feature extractor component and takes advantage of a transformer encoder to perform the final classification.

RESULTS

experiments were performed on the recently released Tufts medical echocardiogram public dataset, which comprises 27,788 images split into training, validation, and test sets. Due to the recent release of this collection, tests were also conducted on several state-of-the-art models to create multi-view and single-view benchmarks. For all models, standard classification metrics were computed (e.g., precision, F1-score). The obtained results show that the proposed approach outperforms other multi-view methods in terms of accuracy and F1-score and has more stable performance throughout the training procedure. Furthermore, the experiments also highlight that multi-view methods generally perform better than their single-view counterparts.

CONCLUSION

this paper introduces a novel multi-view and multi-scale model for aortic stenosis recognition, as well as three benchmarks to evaluate it, effectively providing multi-view and single-view comparisons that fully highlight the model's effectiveness in aiding clinicians in performing diagnoses while also producing several baselines for the aortic stenosis recognition task.

Quantitative ultrasound parameters from scattering and propagation may reduce the biopsy rate for breast tumor

Breast cancer has become the most common cancer worldwide, and early screening improves the patient's survival rate significantly. Although pathology with needle-based biopsy is the gold standard for breast cancer diagnosis, it is invasive, painful, and expensive. Meanwhile it makes patients suffer from misplacement of the needle, resulting in misdiagnosis and further assessment. Ultrasound imaging is non-invasive and real-time, however, benign and malignant tumors are hard to differentiate in grayscale B-mode images. We hypothesis that breast tumors exhibit characteristic properties, which generates distinctive spectral patterns not only in scattering, but also during propagation. In this paper, we propose a breast tumor classification method that evaluates the spectral pattern of the tissues both inside the tumor and beneath it. First, quantitative ultrasonic parameters of these spectral patterns were calculated as the representation of the corresponding tissues. Second, parameters were classified by the K-Nearest Neighbor machine learning model. This method was verified with an open access dataset as a reference, and applied to our own dataset to evaluate the potential for tumors assessment. With both datasets, the proposed method demonstrates accurate classification of the tumors, which potentially makes it unnecessary for certain patients to take the biopsy, reducing the rate of the painful and expensive procedure.

Interobserver variability in thyroid ultrasound

PURPOSE

Ultrasound evaluation of thyroid nodules is the preferred technique, but it is dependent on operator interpretation, leading to inter-observer variability. The current study aimed to determine the inter-physician consensus on nodular characteristics, risk categorization in the classification systems, and the need for fine needle aspiration puncture.

METHODS

Four endocrinologists from the same center blindly evaluated 100 ultrasound images of thyroid nodules from 100 different patients. The following ultrasound features were evaluated: composition, echogenicity, margins, calcifications, and microcalcifications. Nodules were also classified according to ATA, EU-TIRADS, K-TIRADS, and ACR-TIRADS classifications. Krippendorff's alpha test was used to assess interobserver agreement.

RESULTS

The interobserver agreement for ultrasound features was: Krippendorff's coefficient 0.80 (0.71-0.89) for composition, 0.59 (0.47-0.72) for echogenicity, 0.73 (0.57-0.88) for margins, 0.55 (0.40-0.69) for calcifications, and 0.50 (0.34-0.67) for microcalcifications. The concordance for the classification systems was 0.7 (0.61-0.80) for ATA, 0.63 (0.54-0.73) for EU-TIRADS, 0.64 (0.55-0.73) for K-TIRADS, and 0.68 (0.60-0.77) for K-TIRADS. The concordance in the indication of fine needle aspiration puncture (FNA) was 0.86 (0.71-1), 0.80 (0.71-0.88), 0.77 0.67-0.87), and 0.73 (0.64-0.83) for systems previously described respectively.

CONCLUSIONS

Interobserver agreement was acceptable for the identification of nodules requiring cytologic study using various classification systems. However, limited concordance was observed in risk stratification and many ultrasonographic characteristics of the nodules.

Comparative Cyto-Histological Genetic Profile in a Series of Differentiated Thyroid Carcinomas

INTRODUCTION

Molecular tests can contribute to improve the preoperative diagnosis of thyroid nodules. Tests available are expensive and not adapted to different populations.

AIM

This study aimed to compare the cyto-histological genetic profile and to evaluate the reliability of molecular tests using ultrasound-guided fine needle aspiration cytology (US-FNAC) in accurately diagnosing differentiated thyroid carcinomas (DTCs) and predicting biologic behavior of papillary thyroid carcinomas (PTCs).

MATERIALS AND METHODS

The series included 259 patients with paired cyto-histological samples totaling 518 samples. The genetic alterations were analyzed via PCR/Sanger sequencing. The association with clinicopathologic features was evaluated in PTCs.

RESULTS/DISCUSSION

From the 259 patients included, histologies were 50 (19.3%) benign controls and 209 (80.7%) DTC cases, from which 182 were PTCs; cytologies were 5.8% non-diagnostic, 18.2% benign, 39% indeterminate, and 37.1% malignant. In histology, indeterminate nodules (n = 101) were 22.8% benign and 77.2% malignant. Mutation frequencies in cytology and histology specimens were, respectively, TERTp: 3.7% vs. 7.9%; BRAF: 19.5% vs. 25.1%; and RAS: 11% vs. 17.5%. The overall cyto-histological agreement of the genetic mutations was 94.9%, with Cohen's k = 0.67, and in indeterminate nodules agreement was 95.7%, k = 0.64. The identified mutations exhibited a discriminative ability in diagnosing DTC with a specificity of 100% for TERTp and BRAF, and of 94% for RAS, albeit with low sensitivity. TERTp and BRAF mutations were associated with aggressive clinicopathological features and tumor progression in PTCs (p < 0.001). The obtained good cyto-histological agreement suggests that molecular analysis via US-FNAC may anticipate the genetic profile and the behavior of thyroid tumors, confirming malignancy and contributing to referring patients to surgery.

Exploring non-invasive precision treatment in non-small cell lung cancer patients through deep learning radiomics across imaging features and molecular phenotypes

BACKGROUND

Accurate prediction of tumor molecular alterations is vital for optimizing cancer treatment. Traditional tissue-based approaches encounter limitations due to invasiveness, heterogeneity, and molecular dynamic changes. We aim to develop and validate a deep learning radiomics framework to obtain imaging features that reflect various molecular changes, aiding first-line treatment decisions for cancer patients.

METHODS

We conducted a retrospective study involving 508 NSCLC patients from three institutions, incorporating CT images and clinicopathologic data. Two radiomic scores and a deep network feature were constructed on three data sources in the 3D tumor region. Using these features, we developed and validated the 'Deep-RadScore,' a deep learning radiomics model to predict prognostic factors, gene mutations, and immune molecule expression levels.

FINDINGS

The Deep-RadScore exhibits strong discrimination for tumor molecular features. In the independent test cohort, it achieved impressive AUCs: 0.889 for lymphovascular invasion, 0.903 for pleural invasion, 0.894 for T staging; 0.884 for EGFR and ALK, 0.896 for KRAS and PIK3CA, 0.889 for TP53, 0.895 for ROS1; and 0.893 for PD-1/PD-L1. Fusing features yielded optimal predictive power, surpassing any single imaging feature. Correlation and interpretability analyses confirmed the effectiveness of customized deep network features in capturing additional imaging phenotypes beyond known radiomic features.

INTERPRETATION

This proof-of-concept framework demonstrates that new biomarkers across imaging features and molecular phenotypes can be provided by fusing radiomic features and deep network features from multiple data sources. This holds the potential to offer valuable insights for radiological phenotyping in characterizing diverse tumor molecular alterations, thereby advancing the pursuit of non-invasive personalized treatment for NSCLC patients.

Artificial Intelligence Model Assisting Thyroid Nodule Diagnosis and Management: A Multicenter Diagnostic Study

CONTEXT

It is not clear how to integrate artificial intelligence (AI)-based models into diagnostic workflows.

OBJECTIVE

To develop and validate a deep-learning-based AI model (AI-Thyroid) for thyroid cancer diagnosis, and to explore how this improves diagnostic performance.

METHODS

The system was trained using 19 711 images of 6163 patients in a tertiary hospital (Ajou University Medical Center; AUMC). It was validated using 11 185 images of 4820 patients in 24 hospitals (test set 1) and 4490 images of 2367 patients in AUMC (test set 2). The clinical implications were determined by comparing the findings of six physicians with different levels of experience (group 1: 4 trainees, and group 2: 2 faculty radiologists) before and after AI-Thyroid assistance.

RESULTS

The area under the receiver operating characteristic (AUROC) curve of AI-Thyroid was 0.939. The AUROC, sensitivity, and specificity were 0.922, 87.0%, and 81.5% for test set 1 and 0.938, 89.9%, and 81.6% for test set 2. The AUROCs of AI-Thyroid did not differ significantly according to the prevalence of malignancies (>15.0% vs ≤15.0%, P = .226). In the simulated scenario, AI-Thyroid assistance changed the AUROC, sensitivity, and specificity from 0.854 to 0.945, from 84.2% to 92.7%, and from 72.9% to 86.6% (all P < .001) in group 1, and from 0.914 to 0.939 (P = .022), from 78.6% to 85.5% (P = .053) and from 91.9% to 92.5% (P = .683) in group 2. The interobserver agreement improved from moderate to substantial in both groups.

CONCLUSION

AI-Thyroid can improve diagnostic performance and interobserver agreement in thyroid cancer diagnosis, especially in less-experienced physicians.

Artificial intelligence technology enhances the performance of shear wave elastography in thyroid nodule diagnosis

OBJECTIVE

To investigate the diagnostic value of Artificial Intelligence (AI) in thyroid nodules diseases.

METHODS

This study included 100 patients (100 nodules - 23 benign; 77 malignant) who underwent shear wave elastography (SWE) and AI imaging of nodules prior to biopsy and/or surgery in Zhangjiakou First Hospital from January 2021 to December 2021. The image diagnostic value of AI was analyzed.

RESULTS

Among the 100 patients, there were 77 malignant nodules (77%) and 23 benign nodules (23%). Papillary thyroid carcinoma accounted for 94.8% (74/77) of the malignant nodules, and nodular goiter accounted for 100% of the benign nodules. The overall detection rate of AI+SWE was higher than that of SWE alone (P < 0.05). The accuracy, sensitivity, specificity, negative predictive value, and positive predictive value of AI+SWE were all higher than those of SWE only (P < 0.05). The ROC curve results showed that the area under the curve of AI+SWE in the diagnosis of thyroid nodules was 0.903. This was higher than that of SWE (P < 0.05).

CONCLUSION

SWE+AI is effective in the diagnosis of thyroid nodules, and its sensitivity and specificity are better than those of SWE only.

Knowledge mapping of immunotherapy for thyroid cancer from 1980 to 2022: A review

With the gradual development of immunotherapy for thyroid cancer, relevant research has increased. To better understand the current situation, development trend, evolution process, and research hotspots of this field, we conducted this comprehensive bibliometrics visual analysis. We retrieved papers published from 1980 to 2022 from Web of Science Core Collection on January 31, 2023. CiteSpace, Pajek, VOSviewer, R-Bibliometrix, and Scimago Graphics are the tools to perform the analysis. Analysis methods mainly include co-occurrence analysis and cluster analysis. Analysis objects are countries or regions, institutions, authors, journals, and keywords, etc. In terms of publication number, the recent decade has witnessed rapid growth. USA was the most prolific country and has the most influence in the cooperation team. Sweden took the lead in focus on this research field and lasted for 21 years. Garden State Cancer Center was released most papers (28). INSERM played a major role in institutional cooperation. Goldenberg DM published the most papers (48), with H-Index 25 and G-Index 43. Journal of Nuclear Medicine has the greatest papers published (41). The average impactor factor of the top 10 journals is 7.2058. The top keywords with high burst strength are: radioimmunotherapy (14.85), monoclonal antibody (13.78), non hodgkins lymphoma (12.54). The research field of immunotherapy for thyroid cancer will be further developed. This study provides a valuable reference for future research in the field.

AI-Assisted Ultrasound for the Early Diagnosis of Antibody-Negative Autoimmune Thyroiditis

The prevalence of antibody-negative chronic autoimmune thyroiditis (SN-CAT) is increasing. The early diagnosis of SN-CAT can effectively prevent its further development. Thyroid ultrasound can diagnose autoimmune thyroiditis and predict hypothyroidism. Primary hypothyroidism with a hypoechoic pattern suggested by thyroid ultrasound and negative thyroid serum antibodies is the main basis for the diagnosis of SN-CAT. However, for early SN-CAT, only hypoechoic thyroid changes and serological antibodies are currently available. This study explored how to achieve an accurate and early diagnosis of SN-CAT and prevent the development of SN-CAT combined with hypothyroidism. The diagnosis of a hypoechoic thyroid by artificial intelligence is expected to be a breakthrough in the accurate diagnosis of SN-CAT.

Updates in Thyroid Cancer Surgery

This Special Issue of Cancers entitled "Updates in thyroid surgery" is a collection of nine articles that covers a wide range of topics, providing a comprehensive picture of the latest developments in thyroid surgery [...].

A systematic review on artificial intelligence techniques for detecting thyroid diseases

The use of artificial intelligence approaches in health-care systems has grown rapidly over the last few years. In this context, early detection of diseases is the most common area of application. In this scenario, thyroid diseases are an example of illnesses that can be effectively faced if discovered quite early. Detecting thyroid diseases is crucial in order to treat patients effectively and promptly, by saving lives and reducing healthcare costs. This work aims at systematically reviewing and analyzing the literature on various artificial intelligence-related techniques applied to the detection and identification of various diseases related to the thyroid gland. The contributions we reviewed are classified according to different viewpoints and taxonomies in order to highlight pros and cons of the most recent research in the field. After a careful selection process, we selected and reviewed 72 papers, analyzing them according to three main research questions, i.e., which diseases of the thyroid gland are detected by different artificial intelligence techniques, which datasets are used to perform the aforementioned detection, and what types of data are used to perform the detection. The review demonstrates that the majority of the considered papers deal with supervised methods to detect hypo- and hyperthyroidism. The average accuracy of detection is high (96.84%), but the usage of private and outdated datasets with a majority of clinical data is very common. Finally, we discuss the outcomes of the systematic review, pointing out advantages, disadvantages, and future developments in the application of artificial intelligence for thyroid diseases detection.

Predicting Extrathyroidal Extension in Papillary Thyroid Carcinoma Using a Clinical-Radiomics Nomogram Based on B-Mode and Contrast-Enhanced Ultrasound

Papillary thyroid carcinoma (PTC) is the most common pathological type of thyroid cancer. PTC patients with extrathyroidal extension (ETE) are associated with poor prognoses. The preoperative accurate prediction of ETE is crucial for helping the surgeon decide on the surgical plan. This study aimed to establish a novel clinical-radiomics nomogram based on B-mode ultrasound (BMUS) and contrast-enhanced ultrasound (CEUS) for the prediction of ETE in PTC. A total of 216 patients with PTC between January 2018 and June 2020 were collected and divided into the training set (n = 152) and the validation set (n = 64). The least absolute shrinkage and selection operator (LASSO) algorithm was applied for radiomics feature selection. Univariate analysis was performed to find clinical risk factors for predicting ETE. The BMUS Radscore, CEUS Radscore, clinical model, and clinical-radiomics model were established using multivariate backward stepwise logistic regression (LR) based on BMUS radiomics features, CEUS radiomics features, clinical risk factors, and the combination of those features, respectively. The diagnostic efficacy of the models was assessed using receiver operating characteristic (ROC) curves and the DeLong test. The model with the best performance was then selected to develop a nomogram. The results show that the clinical-radiomics model, which is constructed by age, CEUS-reported ETE, BMUS Radscore, and CEUS Radscore, showed the best diagnostic efficiency in both the training set (AUC = 0.843) and validation set (AUC = 0.792). Moreover, a clinical-radiomics nomogram was established for easier clinical practices. The Hosmer-Lemeshow test and the calibration curves demonstrated satisfactory calibration. The decision curve analysis (DCA) showed that the clinical-radiomics nomogram had substantial clinical benefits. The clinical-radiomics nomogram constructed from the dual-modal ultrasound can be exploited as a promising tool for the pre-operative prediction of ETE in PTC.

US Evaluation of Topical Hemostatic Agents in Post-Thyroidectomy

BACKGROUND

the aim of this study was to describe the ultrasound appearance of topical hemostatics after thyroidectomy.

METHODS

we enrolled 84 patients who were undergoing thyroid surgery and were treated with two types of topical hemostats, 49 with an absorbable hemostat of oxidized regenerated cellulose (Oxitamp^®) and 35 with a fibrin glue-based hemostat (Tisseel^®). All patients were examined using B-mode ultrasound.

RESULTS

In 39 patients of the first group (approximately 80%), a hemostatic residue was detected and in some cases confused with a native gland residue, or with cancer recurrence in oncological patients. No residue was detected in patients in the second group. The main ultrasound characteristics of the tampon were analyzed and arranged according to predefined patterns, and suggestions to recognize it and avoid wrong diagnoses were provided. A part of the group of patients with tampon residue was re-evaluated after 6-12 months, ensuring that the swab remained for months after the maximum resorption time declared by the manufacturer.

CONCLUSIONS

with equal hemostatic effectiveness, the fibrin glue pad is more favorable in the ultrasound follow-up because it creates reduced surgical outcomes. It is also important to know and recognize the ultrasound characteristics of oxidized cellulose-based hemostats in order to reduce the number of diagnostic errors and inappropriate diagnostic investigations.

Effect of the categorization method on the diagnostic performance of ultrasound risk stratification systems for thyroid nodules

Objective

To evaluate whether the categorization methods of risk stratification systems (RSSs) is a decisive factor that influenced the diagnostic performances and unnecessary FNA rates in order to choose optimal RSS for the management of thyroid nodules.

Methods

From July 2013 to January 2019, 2667 patients with 3944 thyroid nodules had undergone pathological diagnosis after thyroidectomy and/or US-guided FNA. US categories were assigned according to the six RSSs. The diagnostic performances and unnecessary FNA rates were calculated and compared according to the US-based final assessment categories and the unified size thresholds for biopsy proposed by ACR-TIRADS, respectively.

Results

A total of 1781 (45.2%) thyroid nodules were diagnosed as malignant after thyroidectomy or biopsy. Significantly lowest specificity and accuracy, along with the highest unnecessary FNA rates were seen in EU-TIRADS for both US categories (47.9%, 70.2%, and 39.4%, respectively, all P < 0.05) and indications for FNA (54.2%, 50.0%, and 55.4%, respectively, all P < 0.05). Diagnostic performances for US-based final assessment categories exhibited similar accuracy for AI-TIRADS, Kwak-TIRADS, C-TIRADS, and ATA guidelines (78.0%, 77.8%, 77.9%, and 76.3%, respectively, all P > 0.05), while the lowest unnecessary FNA rate was seen in C-TIRADS (30.9%) and without significant differences to that of AI-TIRADS, Kwak-TIRADS, and ATA guideline (31.5%, 31.7%, and 33.6%, respectively, all P > 0.05). Diagnostic performance for US-FNA indications showed similar accuracy for ACR-TIRADS, Kwak-TIRADS, C-TIRADS and ATA guidelines (58.0%, 59.7%, 58.7%, and 57.1%, respectively, all P > 0.05). The highest accuracy and lowest unnecessary FNA rate were seen in AI-TIRADS (61.9%, 38.6%) and without significant differences to that of Kwak-TIRADS(59.7%, 42.9%) and C-TIRADS 58.7%, 43.9%, all P > 0.05).

Conclusion

The different US categorization methods used by each RSS were not determinant influential factors in diagnostic performance and unnecessary FNA rate. For daily clinical practice, the score-based counting RSS was an optimal choice.

The Complex Cyto-Molecular Landscape of Thyroid Nodules in Pediatrics

Thyroid fine-needle aspiration (FNA) is a commonly used diagnostic cytological procedure in pediatric patients for the evaluation of thyroid nodules, triaging them for the detection of thyroid cancer. In recent years, greater attention has been paid to thyroid FNA in this setting, including the use of updated ultrasound score algorithms to improve accuracy and yield, especially considering the theoretically higher risk of malignancy of these lesions compared with the adult population, as well as to minimize patient discomfort. Moreover, molecular genetic testing for thyroid disease is an expanding field of research that could aid in distinguishing benign from cancerous nodules and assist in determining their clinical management. Finally, artificial intelligence tools can help in this task by performing a comprehensive analysis of all the obtained data. These advancements have led to greater reliance on FNA as a first-line diagnostic tool for pediatric thyroid disease. This review article provides an overview of these recent developments and their impact on the diagnosis and management of thyroid nodules in children.

Explore the diagnostic performance of 2020 Chinese Thyroid Imaging Reporting and Data Systems by comparing with the 2017 ACR-TIRADS guidelines: a single-center study

OBJECTIVE

To compare the diagnostic efficacy of the Chinese Thyroid Imaging Reporting and Data Systems (C-TIRADS) with the well-accepted ACR-TIRADS guidelines in identifying benign from malignant thyroid nodules.

METHODS

A total of 2064 nodules were collected from 1627 patients undergoing thyroid ultrasonography in our center between October 2019 and November 2021. Nodules were divided into two groups: "≥1 cm" and "<1 cm". Ultrasound features of each nodule were observed and recorded by two physicians with more than 15 years of experience and classified according to the ACR-TIRADS and C-TIRADS guidelines, respectively.

RESULTS

The area under the curve of the ACR-TIRADS guideline was higher than that of the C-TIRADS guideline (0.922, P = 0.017), the specificity and positive predictive value of the C-TIRADS guideline were higher (81.64%, 88.72%, all P < 0.05), which was more significant in the subgroup of nodules <1 cm (P = 0.001). In addition, there was no statistical difference between the two guidelines in the diagnostic efficacy indicators for nodules ≥1 cm. The ACR-TIRADS effectively reduced unnecessary biopsies compared with the C-TIRADS (P < 0.05).

CONCLUSIONS

There was high agreement between the two guidelines for the diagnosis of thyroid nodules, C-TIRADS guidelines had a higher specificity and simplicity while were inferior to the ACR-TIRADS guidelines in terms of reducing the number of biopsies.

Nomogram incorporating preoperative clinical and ultrasound indicators to predict aggressiveness of solitary papillary thyroid carcinoma

Objective

To construct a nomogram based on preoperative clinical and ultrasound indicators to predict aggressiveness of solitary papillary thyroid carcinoma (PTC).

Methods

Preoperative clinical and ultrasound data from 709 patients diagnosed with solitary PTC between January 2017 and December 2020 were analyzed retrospectively. Univariate and multivariate logistic regression analyses were performed to identify the factors associated with PTC aggressiveness, and these factors were used to construct a predictive nomogram. The nomogram's performance was evaluated in the primary and validation cohorts.

Results

The 709 patients were separated into a primary cohort (n = 424) and a validation cohort (n = 285). Univariate analysis in the primary cohort showed 13 variables to be associated with aggressive PTC. In multivariate logistic regression analysis, the independent predictors of aggressive behavior were age (OR, 2.08; 95% CI, 1.30-3.35), tumor size (OR, 4.0; 95% CI, 2.17-7.37), capsule abutment (OR, 2.53; 95% CI, 1.50-4.26), and suspected cervical lymph nodes metastasis (OR, 2.50; 95% CI, 1.20-5.21). The nomogram incorporating these four predictors showed good discrimination and calibration in both the primary cohort (area under the curve, 0.77; 95% CI, 0.72-0.81; Hosmer-Lemeshow test, P = 0.967 and the validation cohort (area under the curve, 0.72; 95% CI, 0.66-0.78; Hosmer-Lemeshow test, P = 0.251).

Conclusion

The proposed nomogram shows good ability to predict PTC aggressiveness and could be useful during treatment decision making.

Advances in knowledge

Our nomogram-based on four indicators-provides comprehensive assessment of aggressive behavior of PTC and could be a useful tool in the clinic.

The Use of Artificial Intelligence in the Diagnosis and Classification of Thyroid Nodules: An Update

The incidence of thyroid nodules diagnosed is increasing every year, leading to a greater risk of unnecessary procedures being performed or wrong diagnoses being made. In our paper, we present the latest knowledge on the use of artificial intelligence in diagnosing and classifying thyroid nodules. We particularly focus on the usefulness of artificial intelligence in ultrasonography for the diagnosis and characterization of pathology, as these are the two most developed fields. In our search of the latest innovations, we reviewed only the latest publications of specific types published from 2018 to 2022. We analyzed 930 papers in total, from which we selected 33 that were the most relevant to the topic of our work. In conclusion, there is great scope for the use of artificial intelligence in future thyroid nodule classification and diagnosis. In addition to the most typical uses of artificial intelligence in cancer differentiation, we identified several other novel applications of artificial intelligence during our review.

Diagnostic efficiency of existing guidelines and the AI-SONIC™ artificial intelligence for ultrasound-based risk assessment of thyroid nodules

INTRODUCTION

The thyroid ultrasound guidelines include the American College of Radiology Thyroid Imaging Reporting and Data System, Chinese-Thyroid Imaging Reporting and Data System, Korean Society of Thyroid Radiology, European-Thyroid Imaging Reporting and Data System, American Thyroid Association, and American Association of Clinical Endocrinologists/American College of Endocrinology/Associazione Medici Endocrinologi guidelines. This study aimed to compare the efficiency of the six ultrasound guidelines vs. an artificial intelligence system (AI-SONICTM) in differentiating thyroid nodules, especially medullary thyroid carcinoma.

METHODS

This retrospective study included patients with medullary thyroid carcinoma, papillary thyroid carcinoma, or benign nodules who underwent nodule resection between May 2010 and April 2020 at one hospital. The diagnostic efficacy of the seven diagnostic tools was evaluated using the receiver operator characteristic curves.

RESULTS

Finally, 432 patients with 450 nodules were included for analysis. The American Association of Clinical Endocrinologists/American College of Endocrinology/Associazione Medici Endocrinologi guidelines had the best sensitivity (88.1%) and negative predictive value (78.6%) for differentiating papillary thyroid carcinoma or medullary thyroid carcinoma vs. benign nodules, while the Korean Society of Thyroid Radiology guidelines had the best specificity (85.6%) and positive predictive value (89.6%), and the American Thyroid Association guidelines had the best accuracy (83.7%). When assessing medullary thyroid carcinoma, the American Thyroid Association guidelines had the highest area under the curve (0.78), the American College of Radiology Thyroid Imaging Reporting and Data System guidelines had the best sensitivity (90.2%), and negative predictive value (91.8%), and AI-SONICTM had the best specificity (85.6%) and positive predictive value (67.5%). The Chinese-Thyroid Imaging Reporting and Data System guidelines had the best under the curve (0.86) in diagnosing malignant tumors vs. benign tumors, followed by the American Thyroid Association and Korean Society of Thyroid Radiology guidelines. The best positive likelihood ratios were achieved by the Korean Society of Thyroid Radiology guidelines and AI-SONICTM (both 5.37). The best negative likelihood ratio was achieved by the American Association of Clinical Endocrinologists/American College of Endocrinology/Associazione Medici Endocrinologi guidelines (0.17). The highest diagnostic odds ratio was achieved by the American Thyroid Association guidelines (24.78).

DISCUSSION

All six guidelines and the AI-SONICTM system had satisfactory value in differentiating benign vs. malignant thyroid nodules.

The progress of radiomics in thyroid nodules

Due to the development of Artificial Intelligence (AI), Machine Learning (ML), and the improvement of medical imaging equipment, radiomics has become a popular research in recent years. Radiomics can obtain various quantitative features from medical images, highlighting the invisible image traits and significantly enhancing the ability of medical imaging identification and prediction. The literature indicates that radiomics has a high potential in identifying and predicting thyroid nodules. So in this article, we explain the development, definition, and workflow of radiomics. And then, we summarize the applications of various imaging techniques in identifying benign and malignant thyroid nodules, predicting invasiveness and metastasis of thyroid lymph nodes, forecasting the prognosis of thyroid malignancies, and some new advances in molecular level and deep learning. The shortcomings of this technique are also summarized, and future development prospects are provided.

Editorial: Recent Advances in Thyroid Surgery

Thyroid surgery has been, since its earliest application, one of the most notable fields in medicine, illustrated by the fact that the Nobel Prize in Medicine was won, for the first time, for thyroid surgery by Emil Theodor Kocher (1841-1917) in 1909, for his contributions to thyroid physiology, pathology, and surgery [...].

Deep learning for the diagnosis of suspicious thyroid nodules based on multimodal ultrasound images

OBJECTIVES

This study aimed to differentially diagnose thyroid nodules (TNs) of Thyroid Imaging Reporting and Data System (TI-RADS) 3-5 categories using a deep learning (DL) model based on multimodal ultrasound (US) images and explore its auxiliary role for radiologists with varying degrees of experience.

METHODS

Preoperative multimodal US images of 1,138 TNs of TI-RADS 3-5 categories were randomly divided into a training set (n = 728), a validation set (n = 182), and a test set (n = 228) in a 4:1:1.25 ratio. Grayscale US (GSU), color Doppler flow imaging (CDFI), strain elastography (SE), and region of interest mask (Mask) images were acquired in both transverse and longitudinal sections, all of which were confirmed by pathology. In this study, fivefold cross-validation was used to evaluate the performance of the proposed DL model. The diagnostic performance of the mature DL model and radiologists in the test set was compared, and whether DL could assist radiologists in improving diagnostic performance was verified. Specificity, sensitivity, accuracy, positive predictive value, negative predictive value, and area under the receiver operating characteristics curves (AUC) were obtained.

RESULTS

The AUCs of DL in the differentiation of TNs were 0.858 based on (GSU + SE), 0.909 based on (GSU + CDFI), 0.906 based on (GSU + CDFI + SE), and 0.881 based (GSU + Mask), which were superior to that of 0.825-based single GSU (p = 0.014, p< 0.001, p< 0.001, and p = 0.002, respectively). The highest AUC of 0.928 was achieved by DL based on (G + C + E + M)US, the highest specificity of 89.5% was achieved by (G + C + E)US, and the highest accuracy of 86.2% and sensitivity of 86.9% were achieved by DL based on (G + C + M)US. With DL assistance, the AUC of junior radiologists increased from 0.720 to 0.796 (p< 0.001), which was slightly higher than that of senior radiologists without DL assistance (0.796 vs. 0.794, p > 0.05). Senior radiologists with DL assistance exhibited higher accuracy and comparable AUC than that of DL based on GSU (83.4% vs. 78.9%, p = 0.041; 0.822 vs. 0.825, p = 0.512). However, the AUC of DL based on multimodal US images was significantly higher than that based on visual diagnosis by radiologists (p< 0.05).

CONCLUSION

The DL models based on multimodal US images showed exceptional performance in the differential diagnosis of suspicious TNs, effectively increased the diagnostic efficacy of TN evaluations by junior radiologists, and provided an objective assessment for the clinical and surgical management phases that follow.

Ultrasound super-resolution imaging for the differential diagnosis of thyroid nodules: A pilot study

OBJECTIVE

Ultrasound imaging provides a fast and safe examination of thyroid nodules. Recently, the introduction of super-resolution imaging technique shows the capability of breaking the Ultrasound diffraction limit in imaging the micro-vessels. The aim of this study was to evaluate its feasibility and value for the differentiation of thyroid nodules.

METHODS

In this study, B-mode, contrast-enhanced ultrasound, and color Doppler flow imaging examinations were performed on thyroid nodules in 24 patients. Super-resolution imaging was performed to visualize the microvasculature with finer details. Microvascular flow rate (MFR) and micro-vessel density (MVD) within thyroid nodules were computed. The MFR and MVD were used to differentiate the benign and malignant thyroid nodules with pathological results as a gold standard.

RESULTS

Super-resolution imaging (SRI) technique can be successfully applied on human thyroid nodules to visualize the microvasculature with finer details and obtain the useful clinical information MVD and MFR to help differential diagnosis. The results suggested that the mean value of the MFR within benign thyroid nodule was 16.76 ± 6.82 mm/s whereas that within malignant thyroid was 9.86 ± 4.54 mm/s. The mean value of the MVD within benign thyroid was 0.78 while the value for malignant thyroid region was 0.59. MFR and MVD within the benign thyroid nodules were significantly higher than those within the malignant thyroid nodules respectively (p < 0.01).

CONCLUSIONS

This study demonstrates the feasibility of ultrasound super-resolution imaging to show micro-vessels of human thyroid nodules via a clinical ultrasound platform. The important imaging markers, such as MVD and MFR, can be derived from SRI to provide more useful clinical information. It has the potential to be a new tool for aiding differential diagnosis of thyroid nodules.

Diagnostic performance of six ultrasound-based risk stratification systems in thyroid follicular neoplasm: A retrospective multi-center study

This study aimed to compare the diagnostic performances of six commonly used ultrasound-based risk stratification systems for distinguishing follicular thyroid adenoma (FTA) from follicular thyroid carcinoma (FTC), including the American Thyroid Association Sonographic Pattern System (ATASPS), ultrasound classification systems proposed by American Association of Clinical Endocrinologists, American College of Endocrinology, and Associazione Medici Endocrinology (AACE/ACE/AME), Korean thyroid imaging reporting and data system (K-TIRADS), European Thyroid Association for the imaging reporting and data system (EU-TIRADS), American College of Radiology for the imaging reporting and data system (ACR-TIRADS), and 2020 Chinese Guidelines for Ultrasound Malignancy Risk Stratification of Thyroid Nodules (C-TIRADS). A total of 225 FTA or FTC patients were retrospectively analyzed, involving 251 thyroid nodules diagnosed by postoperative pathological examinations in three centers from January 2013 to October 2021. The diagnostic performances of six ultrasound-based risk stratification systems for distinguishing FTA from FTC were assessed by plotting the receiver operating characteristic (ROC) curves and compared at different cut-off values. A total of 205 (81.67%) cases of FTA and 46 (18.33%) cases of FTC were involved in the present study. Compared with those of FTA, FTC presented more typical ultrasound features of solid component, hypoechoic, irregular margin and sonographic halo (all P<0.001). There were no significant differences in ultrasound features of calcification, shape and comet-tail artifacts between cases of FTA and FTC. There was a significant difference in the category of thyroid nodules assessed by the six ultrasound-based risk stratification systems (P<0.001). The areas under the curve (AUCs) of ATASPS, AACE/ACE/AME, K-TIRADS, EU-TIRADS, ACR-TIRADS and C-TIRADS in distinguishing FTA from FTC were 0.645, 0.729, 0.766, 0.635, 0.783 and 0.798, respectively. Our study demonstrated that all the six ultrasound-based risk stratification systems present potential in the differential diagnosis of FTA and FTC. Specifically, C-TIRADS exerts the best diagnostic performance among the Chinese patients. ATASPS possesses a high sensitivity, while K-TIRADS possesses a high specificity in distinguishing FTA from FTC.

Thyroid and renal cancers: A bidirectional association

There is a deep interrelation between the thyroid gland and the kidney parenchyma, with dysfunction of the first leading to significant changes in renal metabolism and vice versa. Given the recognition of cancer as a systemic disease, the raise of thyroid tumors and the common association of several malignancies, such as breast cancer, prostate cancer, colorectal cancer, and other, with an increased risk of kidney disease, public health alert for these conditions is warranted. A systematic review of the current evidence on the bidirectional relationship between thyroid and renal cancers was conducted including 18 studies, highlighting patient’s characteristics, histology, time for secondary malignancy to develop from the first diagnosis, treatment, and follow-up. A total of 776 patients were identified; median age was 64 years (range: 7–76 years). Obesity and family history were identified as the most common risk factors, and genetic susceptibility was suggested with a potential strong association with Cowden syndrome. Controversy on chemo and radiotherapy effects was found, as not all patients were previously exposed to these treatments. Men were more likely to develop kidney cancer after a primary thyroid malignancy, with 423/776 (54%) experiencing renal disease secondarily. Median time after the first malignancy was 5.2 years (range: 0–20 years). With the advancement of current oncological therapy, the prognosis for thyroid cancer patients has improved, although there has been a corresponding rise in the incidence of multiple secondary malignancy within the same population, particularly concerning the kidney. Surgery can achieve disease-free survival, if surveillance follow-up allows for an early localized form, where radical treatment is recommended.

The TNAPP web-based algorithm improves thyroid nodule management in clinical practice: A retrospective validation study

BACKGROUND

The detection of thyroid nodules has been increasing over time, resulting in an extensive use of fine-needle aspiration (FNA) and cytology. Tailored methods are required to improve the management of thyroid nodules, including algorithms and web-based tools.

STUDY AIMS

To assess the performance of the Thyroid Nodule App (TNAPP), a web-based, readily modifiable, interactive algorithmic tool, in improving the management of thyroid nodules.

METHODS

One hundred twelve consecutive patients with 188 thyroid nodules who underwent FNA from January to December 2016 and thyroid surgery were retrospectively evaluated. Neck ultrasound images were collected from a thyroid nodule registry and re-examined to extract data to run TNAPP. Each nodule was evaluated for ultrasonographic risk and suitability for FNA. The sensitivity, specificity, positive and negative predictive values, and overall accuracy of TNAPP were calculated and compared to the diagnostic performance of the other two algorithms by the American Association of Clinical Endocrinology/American College of Endocrinology/Associazione Medici Endocrinologi (AACE/ACE/AME), which it was derived from the American College of Radiology Thyroid Imaging Reporting and Data System (ACR TI-RADS).

RESULTS

TNAPP performed better in terms of sensitivity (>80%) and negative predictive value (68%) with an overall accuracy of 50.5%, which was similar to that found with the AACE/ACE/AME algorithm. TNAPP displayed a slightly better performance than AACE/ACE/AME and ACR TI-RADS algorithms in selectively discriminating unnecessary FNA for nodules with benign cytology (TIR 2 - Bethesda class II: TNAPP 32% vs. AACE/ACE/AME 31% vs. ACR TI-RADS 29%). The TNAPP reduced the number of missed diagnoses of thyroid nodules with suspicious and highly suspicious cytology (TIR 4 + TIR 5 - Bethesda classes V + VI: TNAPP 18% vs. AACE/ACE/AME 26% vs. ACR TI-RADS 20.5%). A total of 14 nodules that would not have been aspirated were malignant, 13 of which were microcarcinomas (92.8%).

DISCUSSION

The TNAPP algorithm is a reliable, easy-to-learn tool that can be readily employed to improve the selection of thyroid nodules requiring cytological characterization. The rate of malignant nodules missed because of inaccurate characterization at baseline by TNAPP was lower compared to the other two algorithms and, in almost all the cases, the tumors were microcarcinomas. TNAPP's use of size >20 mm as an independent determinant for considering or recommending FNA reduced its specificity.

CONCLUSION

TNAPP performs well compared to AACE/ACE/AME and ACR-TIRADS algorithms. Additional retrospective and, ultimately, prospective studies are needed to confirm and guide the development of future iterations that incorporate different risk stratification systems and targets for diagnosing malignancy while reducing unnecessary FNA procedures.

An innovative synthetic support for immunocytochemical assessment of cytologically indeterminate (Bethesda III) thyroid nodules

BACKGROUND

Fine needle aspiration (FNA) is the procedure of choice in the evaluation of thyroid nodules. Nodules with indeterminate cytological categories, Bethesda III and IV, pose challenges in clinical practice and are frequently submitted to diagnostic surgery. CytoFoam Core (CFCS) uses an absorbent foam device inserted into the needle hub to collect the cytological sample aspirated during FNA. Specimen is formalin-fixed and paraffin-embedded.

AIM OF THE STUDY

Assessing diagnostic efficacy of CFCS, compared to traditional cytology, in re-evaluating thyroid nodules classified as Bethesda III, using post-surgical histology as reference standard.

METHOD

Retrospective study on 89 patients with a first indeterminate cytological report who were referred to the Department of Endocrinology of Regina Apostolorum Hospital (Albano L. Rome, Italy) for a second FNA. FNA was performed after at least one month under ultrasound guidance with a 23G needle according to the established procedure. During the second procedure, both traditional cytological (TC) smears and a single-pass CFCS specimen were obtained for each patient. On CFCS samples immunocytochemical staining for Galectin-3, HBME-1, and CK-19 was also performed. 51 patients eventually underwent surgery, and their histological diagnoses were compared to the TC and CFCS reports. Four parameters were evaluated: inadequacy rate, rate of persistent indeterminate (Bethesda III and IV) reports, rate of malignancy in persistently indeterminate nodules, and rate of cancer in lesions cytologically classified as malignant.

RESULTS

Non-diagnostic samples were 6 (11.8%) in TC vs 3 (5.9%) in CFCS (p=0.4). Persistent indeterminate samples were 31 (60.8%) in TC vs 19 (37.2%) in CFCS (p=0.01). Rate of malignancy in persistently indeterminate nodules was 8/19 (42.1%) in CFCS vs 9/31 (29%) in TC group (p=0.3). Nine/51 (17.6%) samples were classified as benign by TC vs 21/51 (41.2%) samples by CFCS (p<0.01). All nodules resulted benign at post-surgical evaluation. Five/51 (9.8%) samples were classified as suspicious for malignancy/malignant in TC group against 8/51 (15.7%) samples in CFCS (p=0.5). Post-surgical evaluation confirmed malignancy in all these cases.

CONCLUSION

CFCS demonstrated greater diagnostic accuracy than TC in repeat FNA assessment of cytologically indeterminate nodules. CFCS increased the conclusive diagnosis rate and decreased the number of cytologically indeterminate cases.

Risk prediction for central lymph node metastasis in isolated isthmic papillary thyroid carcinoma by nomogram: A retrospective study from 2010 to 2021

BACKGROUND

Isthmic papillary thyroid carcinoma (IPTC) is an aggressive thyroid cancer associated with a poor prognosis. Guidelines elaborating on the extent of surgery for IPTC are yet to be developed. This study aims to construct and validate a model to predict central lymph node metastasis (CLNM) in patients with IPTC, which could be used as a risk stratification tool to determine the best surgical approach for patients.

METHODS

Electronic medical records for patients diagnosed with isolated papillary thyroid carcinoma who underwent surgery at Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, from January 2010 to December 2021 were reviewed. All patients who underwent thyroidectomy with central neck dissection (CND) for isolated IPTC were included. We conducted univariate and multivariate logistic regression analyses to assess risk factors for ipsilateral and contralateral CLNM and the number of CLNM in IPTC patients. Based on the analysis, the nomogram construction and internal validations were performed.

RESULTS

A total of 147 patients with isolated IPTC were included. The occurrence of CLNM was 53.7% in the patients. We identified three predictors of ipsilateral CLNM, including age, gender, and size. For contralateral CLNM, three identified predictors were age, gender, and capsular invasion. Predictors for the number of CLNM included age, gender, capsular invasion, tumor size, and chronic lymphocytic thyroiditis (CLT). The concordance index(C-index) of the models predicting ipsilateral CLNM, contralateral CLNM, 1-4 CLNM, and ≥5 CLNM was 0.779 (95%CI, 0.704, to 0.854), 0.779 (95%CI, 0.703 to 0.855), 0.724 (95%CI, 0.629 to 0.818), and 0.932 (95%CI, 0.884 to 0.980), respectively. The corresponding indices for the internal validation were 0.756 (95%CI, 0.753 to 0.758), 0.753 (95%CI, 0.750 to 0.756), 0.706 (95%CI, 0.702 to 0.708), and 0.920 (95%CI, 0.918 to 0.922). Receiver operating characteristic (ROC) curves, calibration, and decision curve analysis (DCA) results confirmed that the three nomograms could precisely predict CLNM in patients with isolated IPTC.

CONCLUSION

We constructed predictive nomograms for CLNM in IPTC patients. A risk stratification scheme and corresponding surgical treatment recommendations were provided accordingly. Our predictive models can be used as a risk stratification tool to help clinicians make individualized surgical plans for their patients.

Identification of benign and malignant thyroid nodules based on dynamic AI ultrasound intelligent auxiliary diagnosis system

BACKGROUND

Dynamic artificial intelligence (AI) ultrasound intelligent auxiliary diagnosis system (Dynamic AI) is a joint application of AI technology and medical imaging data, which can perform a real-time synchronous dynamic analysis of nodules. The aim of this study is to investigate the value of dynamic AI in differentiating benign and malignant thyroid nodules and its guiding significance for treatment strategies.

METHODS

The data of 607 patients with 1007 thyroid nodules who underwent surgical treatment were reviewed and analyzed, retrospectively. Dynamic AI was used to differentiate benign and malignant nodules. The diagnostic efficacy of dynamic AI was evaluated by comparing the results of dynamic AI examination, preoperative fine needle aspiration cytology (FNAC) and postoperative pathology of nodules with different sizes and properties in patients of different sexes and ages.

RESULTS

The sensitivity, specificity and accuracy of dynamic AI in the diagnosis of thyroid nodules were 92.21%, 83.20% and 89.97%, respectively, which were highly consistent with the postoperative pathological results (kappa = 0.737, p < 0.001). There is no statistical difference in accuracy between people with different ages and sexes and nodules of different sizes, which showed the good stability. The accuracy of dynamic AI in malignant nodules (92.21%) was significantly higher than that in benign nodules (83.20%) (p < 0.001). The specificity and positive predictive value were significantly higher, and the misdiagnosis rate was significantly lower in dynamic AI than that of preoperative ultrasound ACR TI-RADS (p < 0.001). The accuracy of dynamic AI in nodules with diameter ≤ 0.50 cm was significantly higher than that of preoperative ultrasound (p = 0.044). Compared with FNAC, the sensitivity (96.58%) and accuracy (94.06%) of dynamic AI were similar.

CONCLUSIONS

The dynamic AI examination has high diagnostic value for benign and malignant thyroid nodules, which can effectively assist surgeons in formulating scientific and reasonable individualized diagnosis and treatment strategies for patients.

Machine learning-based dynamic prediction of lateral lymph node metastasis in patients with papillary thyroid cancer

OBJECTIVE

To develop a web-based machine learning server to predict lateral lymph node metastasis (LLNM) in papillary thyroid cancer (PTC) patients.

METHODS

Clinical data for PTC patients who underwent primary thyroidectomy at our hospital between January 2015 and December 2020, with pathologically confirmed presence or absence of any LLNM finding, were retrospectively reviewed. We built all models from a training set (80%) and assessed them in a test set (20%), using algorithms including decision tree, XGBoost, random forest, support vector machine, neural network, and K-nearest neighbor algorithm. Their performance was measured against a previously established nomogram using area under the receiver operating characteristic curve (AUC), decision curve analysis (DCA), precision, recall, accuracy, F1 score, specificity, and sensitivity. Interpretable machine learning was used for identifying potential relationships between variables and LLNM, and a web-based tool was created for use by clinicians.

RESULTS

A total of 1135 (62.53%) out of 1815 PTC patients enrolled in this study experienced LLNM episodes. In predicting LLNM, the best algorithm was random forest. In determining feature importance, the AUC reached 0.80, with an accuracy of 0.74, sensitivity of 0.89, and F1 score of 0.81. In addition, DCA showed that random forest held a higher clinical net benefit. Random forest identified tumor size, lymph node microcalcification, age, lymph node size, and tumor location as the most influentials in predicting LLNM. And the website tool is freely accessible at http://43.138.62.202/.

CONCLUSION

The results showed that machine learning can be used to enable accurate prediction for LLNM in PTC patients, and that the web tool allowed for LLNM risk assessment at the individual level.

Influence of Hashimoto thyroiditis on diagnosis and treatment of thyroid nodules

BACKGROUND

As the prevalence of Hashimoto's thyroiditis (HT) and thyroid cancer (TC) has been increasing dramatically in recent years, the association between the two diseases has been widely debated and studied. However, no consistent findings are available and the relationship remains controversial. In this study, we analyzed the influence of HT on the diagnosis and treatment of thyroid nodules and investigated the relationship between HT and TC.

METHODS

From Jan 2017 to Apr 2021, 4678 patients underwent thyroid surgery were collected. Of these patients, 440 were diagnosed with HT (242 nodular goiter (NG) with HT, 198 TC with HT). These patients were grouped when appropriate and the data from these patients were statistically analyzed by using SPSS and GraphPad Prism 6.

RESULTS

HT occurred in 198 of 1089 (18.2%) TC patients and 242 of 3589 (6.74%) patients without TC (p=0.000). In terms of the ultrasonography features, in the NG with HT group, 33.1% (80/242) of patients had fine calcification and 45.9% (111/242) of patients whose TI-RADS classification were > Level 3. In the isolated PTC group, 32.3% (2343/7260) LN were metastasis-positive while in the NG with HT group, only 26.0% (504/1939) LN were metastasis-positive (P=0.000). The proportion of PTMC was significantly higher (P=0.000), while the proportion of multifocal carcinoma was significantly lower when coexisting with HT (P=0.029). When comparing the data from the two groups diagnosed as PTMC coexisting with HT or not, there was no significant difference in the composition ratio of tumor number, LN metastasis, LN dissection area, regional LN metastasis and number of negative/positive LN (P=0.614, P=0.051, P=0.139, P=0.350, P=1.000 and P=0.333 respectively). In the MPTC group, 42.2% (872/2065) LN were metastasis-positive while in the MPTC with HT group, only 23.6% (50/212) LN were metastasis-positive (P=0.000).

CONCLUSIONS

Our data suggested that HT is associated with an increased risk of developing TC but may be a protective factor against PTC progression and metastasis. The coexistence of HT affects the judgment of thyroid nodules by ultrasonography.

Significance of radiofrequency ablation in large solid benign thyroid nodules

OBJECTIVE

The aim of this study is to explore efficacy and safety for radiofrequency ablation (RFA) among cases attacked by large benign solid thyroid nodules, mainly focusing on volume reduction, complication rate, and thyroid function.

METHODS AND MATERIALS

From June 2015 to November 2019, 24 patients with 25 large benign solid thyroid nodules (more than 25 ml) underwent single or sequential RFA in our institution. Eleven nodules achieved complete ablation after single RFA, whereas the other 14 nodules received sequential RFA. Volume reduction in large nodules was evaluated. Following single or sequential RFA, all patients received clinical and ultrasound evaluations, and the median follow-up duration among them was 23.5 months. Technical success, complication rate, and recurrence rate were assessed as well.

RESULTS

In single RFA group, volume reduction ranged from 62.6% to 99.4% (mean ± SD, 93.6 ± 9.9%) 6 months after RFA. In sequential RFA group, volume reduction ranged from 30.6% to 92.9% (mean ± SD, 67.4 ± 17.8%) after the first RFA and was between 83.4% and 98.4% (mean ± SD, 94.8± 3.8%) 6 months after the second RFA. The concentrations of FT3 and FT4 increased slightly 1 day after RFA and returned to normal level 1 month after.

CONCLUSIONS

Single or sequential RFA is safe and effective in treating large benign solid thyroid nodules (more than 25 ml) that cause obvious compressive symptoms. Hence, compression symptoms and cosmetic conditions could be effectively improved through single or sequential RFA without marginal recurrence.

Reconceptualize tall-cell variant papillary thyroid microcarcinoma: From a "sonographic histology" perspective

OBJECTIVE

This study aimed to examine the relationship between sonographic features and histological manifestations in the tall-cell variant of papillary thyroid microcarcinoma (TCV-PTMC), thus proposing the concept of "sonographic histology" and examine its value in the clinical management of the aggressive tall-cell variant.

METHODS

This study retrospectively included 104 participants who were admitted to Peking University Third Hospital from 2015 to 2022 and were histopathologically confirmed as having TCV-PTMC or classical PTMC. We mainly compared the general characteristics, sonographic characteristics, and pathological specimens between the two cohorts.

RESULTS

Hypoechoic nodules with a localized central isoechoic lesion and hypoechoic halo around nodules were most often observed in TCV-PTMC, which correlated with circumferentially distributed tumor epithelium and densely distributed tumor stroma histopathologically. Additionally, TCV-PTMC showed nodules with a more regular margin and less microcalcification than classical PTMC, which led to an underestimation of the risk of TCV-PTMC.

CONCLUSION

The good association between the ultrasound echo pattern and tissue cell arrangement was defined as sonographic histology in this study and can be applied in the preoperative identification of TCV-PTMC. This concept may provide novel insight for the identification of special subtypes of thyroid tumors and may modify pitfalls of the Thyroid Imaging Reporting and Data System in aggressive variants of microcarcinoma.

Ultrasound images-based deep learning radiomics nomogram for preoperative prediction of RET rearrangement in papillary thyroid carcinoma

PURPOSE

To create an ultrasound -based deep learning radiomics nomogram (DLRN) for preoperatively predicting the presence of RET rearrangement among patients with papillary thyroid carcinoma (PTC).

METHODS

We retrospectively enrolled 650 patients with PTC. Patients were divided into the RET/PTC rearrangement group (n = 103) and the non-RET/PTC rearrangement group (n = 547). Radiomics features were extracted based on hand-crafted features from the ultrasound images, and deep learning networks were used to extract deep transfer learning features. The least absolute shrinkage and selection operator regression was applied to select the features of nonzero coefficients from radiomics and deep transfer learning features; then, we established the deep learning radiomics signature. DLRN was constructed using a logistic regression algorithm by combining clinical and deep learning radiomics signatures. The prediction performance was evaluated using the receiver operating characteristic curve, calibration curve, and decision curve analysis.

RESULTS

Comparing the effectiveness of the models by linking the area under the receiver operating characteristic curve of each model, we found that the area under the curve of DLRN could reach 0.9545 (95% confidence interval: 0.9133-0.9558) in the test cohort and 0.9396 (95% confidence interval: 0.9185-0.9607) in the training cohort, indicating that the model has an excellent performance in predicting RET rearrangement in PTC. The decision curve analysis demonstrated that the combined model was clinically useful.

CONCLUSION

The novel ultrasonic-based DLRN has an important clinical value for predicting RET rearrangement in PTC. It can provide physicians with a preoperative non-invasive primary screening method for RET rearrangement diagnosis, thus facilitating targeted patients with purposeful molecular sequencing to avoid unnecessary medical investment and improve treatment outcomes.

A radiomics nomogram for the ultrasound-based evaluation of central cervical lymph node metastasis in papillary thyroid carcinoma

PURPOSE

To develop and validate a radiomics nomogram based on ultrasound (US) to predict central cervical lymph node (LN) metastasis in patients with papillary thyroid carcinoma (PTC).

METHODS

PTC patients with pathologically confirmed presence or absence of central cervical LN metastasis in our hospital between March 2021 and November 2021 were enrolled as the training cohort. Radiomics features were extracted from the preoperative US images, and a radiomics signature was constructed. Univariate and multivariate logistic regression analyses were used to screen out the independent risk factors, and a radiomics nomogram was established. The performance of the model was verified in the independent test cohort of PTC patients who underwent thyroidectomy and cervical LN dissection in our hospital from December 2021 to March 2022.

RESULTS

In the independent test cohort, the radiomics model based on long-axis cross-section and short-axis cross-section images outperformed the radiomics models based on either one of these sections (the area under the curve (AUC), 0.69 vs. 0.62 and 0.66). The radiomics signature consisted of 4 selected features. The US radiomics nomogram included the radiomics signature, age, gender, BRAF V600E mutation status, and extrathyroidal extension (ETE) status. In the independent test cohort, the AUC of the receiver operating curve(ROC) of this nomogram was 0.76, outperformingthe clinical model and the radiomics model (0.63 and 0.69, respectively), and also much better than preoperative US examination (AUC, 0.60). Decision curve analysis indicated that the radiomics nomogram was clinically useful.

CONCLUSIONS

This study presents an efficient and useful US radiomics nomogram that can provide comprehensive information to assist clinicians in the individualized preoperative prediction of central cervical LN metastasis in PTC patients.