A fully autonomous robotic ultrasound system for thyroid scanning.

The current thyroid ultrasound relies heavily on the experience and skills of the sonographer and the expertise of the radiologist, and the process is physically and cognitively exhausting. In this paper, we report a fully autonomous robotic ultrasound system, which is able to scan thyroid regions without human assistance and identify malignant nod- ules. In this system, human skeleton point recognition, reinforcement learning, and force feedback are used to deal with the difficulties in locating thyroid targets. The orientation of the ultrasound probe is adjusted dynamically via Bayesian optimization. Experimental results on human participants demonstrated that this system can perform high-quality ultrasound scans, close to manual scans obtained by clinicians. Additionally, it has the potential to detect thyroid nodules and provide data on nodule characteristics for American College of Radiology Thyroid Imaging Reporting and Data System (ACR TI-RADS) calculation.

Differential diagnosis of thyroid nodules using heterogeneity quantification software on ultrasound images: correlation with the Bethesda system and surgical pathology.

Ultrasonography (US)-guided fine-needle aspiration cytology (FNAC) is the primary modality for evaluating thyroid nodules. However, in cases of atypia of undetermined significance (AUS) or follicular lesion of undetermined significance (FLUS), supplemental tests are necessary for a definitive diagnosis. Accordingly, we aimed to develop a non-invasive quantification software using the heterogeneity scores of thyroid nodules. This cross-sectional study retrospectively enrolled 188 patients who were categorized into four groups according to their diagnostic classification in the Bethesda system and surgical pathology [II-benign (B) (n = 24); III-B (n = 52); III-malignant (M) (n = 54); V/VI-M (n = 58)]. Heterogeneity scores were derived using an image pixel-based heterogeneity index, utilized as a coefficient of variation (CV) value, and analyzed across all US images. Differences in heterogeneity scores were compared using one-way analysis of variance with Tukey's test. Diagnostic accuracy was determined by calculating the area under the receiver operating characteristic (AUROC) curve. The results of this study indicated significant differences in mean heterogeneity scores between benign and malignant thyroid nodules, except in the comparison between III-M and V/VI-M nodules. Among malignant nodules, the Bethesda classification was not observed to be associated with mean heterogeneity scores. Moreover, there was a positive correlation between heterogeneity scores and the combined diagnostic category, which was based on the Bethesda system and surgical cytology grades (R = 0.639, p < 0.001). AUROC for heterogeneity scores showed the highest diagnostic performance (0.818; cut-off: 30.22% CV value) for differentiating the benign group (normal/II-B/III-B) from the malignant group (III-M/V&VI-M), with a diagnostic accuracy of 72.5% (161/122). Quantitative heterogeneity measurement of US images is a valuable non-invasive diagnostic tool for predicting the likelihood of malignancy in thyroid nodules, including AUS or FLUS.

Assessing the relative effectiveness of various ultrasound-guided ablation techniques for treating benign thyroid nodules: A systematic review and network meta-analysis.

BACKGROUND: Benign thyroid nodules (BTNs) represent a prevalent clinical challenge globally, with various ultrasound-guided ablation techniques developed for their management. Despite the availability of these methods, a comprehensive evaluation to identify the most effective technique remains absent. This study endeavors to bridge this knowledge gap through a network meta-analysis (NMA), aiming to enhance the understanding of the comparative effectiveness of different ultrasound-guided ablation methods in treating BTNs. METHODS: We comprehensively searched PubMed, Embase, Cochrane, Web of Science, Ovid, SCOPUS, and ProQuest for studies involving 16 ablation methods, control groups, and head-to-head trials. NMA was utilized to evaluate methods based on the percentage change in nodule volume, symptom score, and cosmetic score. This study is registered in INPLASY (registration number 202260061). RESULTS: Among 35 eligible studies involving 5655 patients, NMA indicated that RFA2 (radiofrequency ablation, 2 sessions) exhibited the best outcomes at 6 months for percentage change in BTN volume (SUCRA value 74.6), closely followed by RFA (SUCRA value 73.7). At 12 months, RFA was identified as the most effective (SUCRA value 81.3). Subgroup analysis showed RFA2 as the most effective for solid nodule volume reduction at 6 months (SUCRA value 75.6), and polidocanol ablation for cystic nodules (SUCRA value 66.5). CONCLUSION: Various ablation methods are effective in treating BTNs, with RFA showing notable advantages. RFA with 2 sessions is particularly optimal for solid BTNs, while polidocanol ablation stands out for cystic nodules.

Thyroid abnormalities identified on CT screening for lung cancer.

PURPOSE: Because incidental thyroid nodules (ITNs) are common extrapulmonary findings in low-dose computed tomography (LDCT) scans for lung cancer screening, we aimed to investigate the frequency of ITNs on LDCT scans separately on baseline and annual repeat scans, the frequency of malignancy among the ITNs, and any association with demographic, clinical, CT characteristics. METHODS: Retrospective case series of all 2309 participants having baseline and annual repeat screening in an Early Lung and Cardiac Action Program (MS-ELCAP) LDCT lung screening program from January 2010 to December 2016 was performed. Frequency of ITNs in baseline and annual repeat rounds were determined. Multivariable regression analysis was performed to identify significant predictors. RESULTS: Dominant ITNs were seen in 2.5 % of 2309 participants on baseline and in 0.15 % of participants among 4792 annual repeat LDCTs. The low incidence of new ITNs suggests slow growth as it would take approximately an average of 16.8 years for a new ITN to be detected on annual rounds of screening. Newly detected ITNs on annual repeat LDCT were all smaller than 15 mm. Regression analysis showed that the increasing of age, coronary artery calcifications score and breast density grade were significant predictors for females having an ITN. No significant predictors were found for ITNs in males. CONCLUSION: ITNs are detected at LDCT however, no malignancy was found. Certain predictors for ITNs in females have been identified including breast density, which may point towards a common causal pathway.

Thyroid US-guided FNA techniques: A prospective,randomized controlled study.

BACKGROUND: Thyroid nodules are common in the general population. Ultrasonography is the most efficient diagnostic approach to evaluate thyroid nodules. The US FNAC procedure can be performed using either the short axis (perpendicular), or a long axis (parallel) approach to visualize the needle as it is advanced toward the desired nodule. The main aim of this study was to compare the percentage of non-diagnostic results between the long and short axis approach. METHODS: A prospective study that included a randomized controlled trial and was divided into two arms-the short axis and the long axis-was conducted. A total of 245 thyroid nodules were collected through the fine needle aspiration cytology, performed with ultrasound, from march 2021 to march 2022. The patient's demographic information were collected and also nodules characteristics. RESULTS: Of 245 nodules sampled, 122 were sampled with the long axis method, while 123 with the short axis method. There is not significantly less non diagnostic approach with either method compared to the other (11.5 % vs 16.3 % respectively). DISCUSSION: Previous studies came to the conclusion that the long axis method yields fewer non diagnostic samples. This study evaluated the two FNA approaches which were proceeded by the same physician who is expert in both techniques. CONCLUSION: The US FNAC performed in the long axis approach will not produce more conclusive results and less non diagnostic results (Bethesda category 1) than the short axis approach one.

The application value of deep learning-based nomograms in benign-malignant discrimination of TI-RADS category 4 thyroid nodules.

Thyroid nodules are a common occurrence, and although most are non-cancerous, some can be malignant. The American College of Radiology has developed the Thyroid Imaging Reporting and Data System (TI-RADS) to standardize the interpretation and reporting of thyroid ultrasound results. Within TI-RADS, a category 4 designation signifies a thyroid nodule with an intermediate level of suspicion for malignancy. Accurate classification of these nodules is crucial for proper management, as it can potentially reduce unnecessary surgeries and improve patient outcomes. This study utilized deep learning techniques to effectively classify TI-RADS category 4 thyroid nodules as either benign or malignant. A total of 500 patients were included in the study and randomly divided into a training group (350 patients) and a test group (150 patients). The YOLOv3 model was constructed and evaluated using various metrics, achieving an 84% accuracy in the classification of TI-RADS category 4 thyroid nodules. Based on the predictions of the model, along with clinical and ultrasound data, a nomogram was developed. The performance of the nomogram was superior in both the training and testing groups. Furthermore, the calibration curve demonstrated good agreement between predicted probabilities and actual outcomes. Decision curve analysis further confirmed that the nomogram provided greater net benefits. Ultimately, the YOLOv3 model and nomogram successfully improved the accuracy of distinguishing between benign and malignant TI-RADS category 4 thyroid nodules, which is crucial for proper management and improved patient outcomes.

Blood leukocytes as a non-invasive diagnostic tool for thyroid nodules: a prospective cohort study.

BACKGROUND: Thyroid nodule (TN) patients in China are subject to overdiagnosis and overtreatment. The implementation of existing technologies such as thyroid ultrasonography has indeed contributed to the improved diagnostic accuracy of TNs. However, a significant issue persists, where many patients undergo unnecessary biopsies, and patients with malignant thyroid nodules (MTNs) are advised to undergo surgery therapy. METHODS: This study included a total of 293 patients diagnosed with TNs. Differential methylation haplotype blocks (MHBs) in blood leukocytes between MTNs and benign thyroid nodules (BTNs) were detected using reduced representation bisulfite sequencing (RRBS). Subsequently, an artificial intelligence blood leukocyte DNA methylation (BLDM) model was designed to optimize the management and treatment of patients with TNs for more effective outcomes. RESULTS: The DNA methylation profiles of peripheral blood leukocytes exhibited distinctions between MTNs and BTNs. The BLDM model we developed for diagnosing TNs achieved an area under the curve (AUC) of 0.858 in the validation cohort and 0.863 in the independent test cohort. Its specificity reached 90.91% and 88.68% in the validation and independent test cohorts, respectively, outperforming the specificity of ultrasonography (43.64% in the validation cohort and 47.17% in the independent test cohort), albeit with a slightly lower sensitivity (83.33% in the validation cohort and 82.86% in the independent test cohort) compared to ultrasonography (97.62% in the validation cohort and 100.00% in the independent test cohort). The BLDM model could correctly identify 89.83% patients whose nodules were suspected malignant by ultrasonography but finally histological benign. In micronodules, the model displayed higher specificity (93.33% in the validation cohort and 92.00% in the independent test cohort) and accuracy (88.24% in the validation cohort and 87.50% in the independent test cohort) for diagnosing TNs. This performance surpassed the specificity and accuracy observed with ultrasonography. A TN diagnostic and treatment framework that prioritizes patients is provided, with fine-needle aspiration (FNA) biopsy performed only on patients with indications of MTNs in both BLDM and ultrasonography results, thus avoiding unnecessary biopsies. CONCLUSIONS: This is the first study to demonstrate the potential of non-invasive blood leukocytes in diagnosing TNs, thereby making TN diagnosis and treatment more efficient in China.

Predictive values of ultrasound characters associated with malignant thyroid nodules in Yaoundé: a cross-sectional study.

Introduction: most ultrasound criteria are defined in developed countries and commonly used in practice to assess the malignancy risk of thyroid nodules. This practice does not take into consideration some aspects of our context as delay of consultation and insufficient iodine intake. The objective of this study was to determine the predictive values of ultrasound characters associated with malignant thyroid nodules in our environment. Methods: we conducted a cross-sectional, prospective, and analytical study in three hospitals in Yaoundé over a six-month period in 2022. Our sample consisted of thyroid nodules with ultrasound, cytopathological, and histopathological data. The ultrasound characters and histology status of category III thyroid nodules and higher in Bethesda score were analysed in univariate and multivariate statistics to determine their predictive values. Results: eighty-nine nodules were obtained according to our inclusion criteria. The sex ratio was 0.46 and the average age of the patients was 46 years (IQR=42-59). The cancer prevalence in our sample was 22.47%. On ultrasound assessment, the characters associated to malignant histology (p<0.05) were nodules count, echogenicity, echostructure, presence or absence of microcalcifications, margins, and type of vascularization. Positive predictive values ranged from 26.15 to 57.14%, while negative predictive values ranged from 12.5 to 33.3%. Conclusion: taken alone, the ultrasound characters of suspected thyroid nodules have poor predictive values. There was a high variability in sensitivity but that was generally good (60-95%) while specificity was low. The prediction of malignant thyroid nodules is correlated with the association of at least two ultrasound criteria supported by clinical arguments.

Long-Term Follow-up of Patients With Autonomous Thyroid Nodules Treated With Radioiodine.

PURPOSE: This article aims to describe the presentation of Plummer disease and its evolution after radioiodine treatment and determine factors that may influence treatment efficacy. PATIENTS AND METHODS: The sample included retrospective medical records of 165 adult patients with toxic nodular goiter treated with radioiodine between 1997 and 2017, followed up at a single thyroid center. RESULTS: The efficacy of treatment with a single dose of radioiodine was higher than 90%. The mean radioiodine activity was 28.9 ± 3.4 mCi. The mean time between radioiodine performance and hyperthyroidism resolution was 3.6 ± 3.0 months, ranging from 1-12 months. After the first year, 33.9% of the patients were under hypothyroidism, 59.4% under euthyroidism, and 6.7% under hyperthyroidism. Among the nonresponders, the variables that showed statistical difference were the presence of multinodular goiter and the radioiodine activity (mean, 25.5 ± 6.5 mCi; median, 30 [15-30 mCi]). The cumulative rate of hypothyroidism was 48.9% over 20 years of follow-up. CONCLUSIONS: Radioiodine therapy is an effective and safe treatment. In Plummer disease, high rates of euthyroidism are expected after the radioiodine treatment. Therapeutic failure was observed mainly in patients with larger multinodular goiters treated with lower doses of radioiodine. The evolution to hypothyroidism was mostly observed in younger patients with larger and uninodular goiters.

Comparison of ultrasound-guided microwave ablation, laser ablation, and radiofrequency ablation for the treatment of elderly patients with benign thyroid nodules: A meta-analysis.

BACKGROUND: A new minimally invasive technique, ultrasound-guided thermal ablation has become one of the treatment methods for benign thyroid nodules. This study aims to evaluate the efficacy and safety of laser ablation (LA), radiofrequency ablation (RFA), and microwave ablation (MWA) in the treatment of elderly patients with benign thyroid nodules. METHODS: PubMed, Web of Science, and Cochrane Library were searched for qualified randomized controlled studies (RCTs) issued from establishing databases to March 2022. After screening and evaluating the article quality, the data on nodular volume reduction rate (VRR) and the incidence of complications after thermal ablation were extracted and analyzed by RevMan 5.3 and Stata l4.0. RESULTS: The meta-analysis included seven articles with 3055 participants. We found that LA, RFA, and MWA could markedly reduce the volume of benign thyroid nodules. LA was superior to RFA and MWA in reducing the volume of benign thyroid nodules in 6 months of follow-up (all P < 0.05). LA, RFA, and MWA can be safely implemented in patients with benign thyroid nodules. The incidence of significant complications after the RFA group was enhanced compared with that in the MWA (P < 0.05), and the incidence of secondary complications after RFA was slightly higher than that of LA (P < 0.05). CONCLUSION: LA, RFA, and MWA can markedly reduce the volume of benign thyroid nodules in elderly patients and can safely treat benign thyroid nodules.

The value of a neural network based on multi-scale feature fusion to ultrasound images for the differentiation in thyroid follicular neoplasms.

OBJECTIVE: The objective of this research was to create a deep learning network that utilizes multiscale images for the classification of follicular thyroid carcinoma (FTC) and follicular thyroid adenoma (FTA) through preoperative US. METHODS: This retrospective study involved the collection of ultrasound images from 279 patients at two tertiary level hospitals. To address the issue of false positives caused by small nodules, we introduced a multi-rescale fusion network (MRF-Net). Four different deep learning models, namely MobileNet V3, ResNet50, DenseNet121 and MRF-Net, were studied based on the feature information extracted from ultrasound images. The performance of each model was evaluated using various metrics, including sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy, F1 value, receiver operating curve (ROC), area under the curve (AUC), decision curve analysis (DCA), and confusion matrix. RESULTS: Out of the total nodules examined, 193 were identified as FTA and 86 were confirmed as FTC. Among the deep learning models evaluated, MRF-Net exhibited the highest accuracy and area under the curve (AUC) with values of 85.3% and 84.8%, respectively. Additionally, MRF-Net demonstrated superior sensitivity and specificity compared to other models. Notably, MRF-Net achieved an impressive F1 value of 83.08%. The curve of DCA revealed that MRF-Net consistently outperformed the other models, yielding higher net benefits across various decision thresholds. CONCLUSION: The utilization of MRF-Net enables more precise discrimination between benign and malignant thyroid follicular tumors utilizing preoperative US.

Multimodal ultrasonography evaluation in thyroid nodule characterization: what is the ideal algorithm?

AIMS: The aim of this study is to investigate the diagnostic performances of Ultrasonography (US), Shear-wave Elastography (SWE), and Superb Microvascular Imaging (SMI) findings in the diagnosis of malignant thyroid nodules (MTNs) and to determine the US algorithm with the best diagnostic performance. MATERIAL AND METHODS: Eighty-one nodules in 77 patients who had underwent multimodal US with biopsy results, were evaluated. Echogenicity, nodule components, contours, presence and type of calcification, and size were analyzed with US. Nodule stiffness and vascular index (VI) measurements were performed via SWE and SMI. The power of the US algorithm in predicting malignancy was evaluated. RESULTS: Hypoechogenicity, irregular contour, aspect ratio (anteroposterior (AP)/transvers diameter) >1, and >43.9 kPa were the characteristicshad significant efficacy in the diagnosis of MTNs. Sensitivity, specificity, and AUC values were respectively 100%, 48.5%, and 0.742 for hypoechogenicity; 80%, 90.1%, and 0.855 for irregular contour; 60%, 71.2%, and 0.656 for aspect ratio >1; 60%, 72.7%, and 0.671 for >43.9 kPa; and 93.3%, 90.9%, and 0.921 for the US algorithm. VI did not show significant efficacy in diagnosis. CONCLUSION: Some B-mode and SWE findings showed sufficient efficacy in differentiating benign and malign nodules on their own. However, diagnostic accuracy increased significantly when the US algorithm was applied.

Comparative evaluation of shear wave elastography elasticity values in thyroid nodules with cytology results and TI-RADS scoring in differentiation of benign-malignant nodules.

PURPOSE: The aim of this prospective study was to investigate the diagnostic performance of shear wave elastography (SWE) in differentiating benign and malignant thyroid nodules and their correlation with the American College of Radiology Thyroid Imaging Reporting and Data System (ACR TI-RADS). METHODS: This prospective study included 370 thyroid nodules in 308 patients aged 18-70 years. All the patients underwent B-mode ultrasound (US), Doppler examination, and SWE and were given an ACR TI-RADS risk score before fine needle aspiration biopsy (FNAB) and/or surgery. The correlation between SWE parameters and ACR TI-RADS categories was investigated statistically and compared with histopathologic results. Additionally, the diagnostic performance of SWE was evaluated to distinguish malignant and benign thyroid nodules. RESULTS: One hundred and thirty-five of the 370 thyroid nodules were malignant, and 235 nodules were benign. The mean shear wave velocity (SWV) value of the malignant nodules (3.70 ± 0.98 m/s) was statistically higher than that of the benign nodules (2.70 ± 0.37 m/s). The best cutoff value of the mean SWV for differentiating benign and malignant nodules was found to be 2.94 m/s (sensitivity 90.4%, specificity 89.9%, positive predictive value 81.3%, negative predictive value 94.1%, p < 0.001). The average score of the nodules according to the ACR TI-RADS was 3.57 ± 1.83 in benign nodules and 7.38 ± 2.69 in malignant nodules (p ≤ 0.001). CONCLUSION: This study showed that combining SWE and TI-RADS improves the specificity of TI-RADS alone in differentiating benign and malignant nodules.

Localization and Risk Stratification of Thyroid Nodules in Ultrasound Images Through Deep Learning.

OBJECTIVE: Deep learning algorithms have commonly been used for the differential diagnosis between benign and malignant thyroid nodules. The aim of the study described here was to develop an integrated system that combines a deep learning model and a clinical standard Thyroid Imaging Reporting and Data System (TI-RADS) for the simultaneous segmentation and risk stratification of thyroid nodules. METHODS: Three hundred four ultrasound images from two independent sites with TI-RADS 4 thyroid nodules were collected. The edge connection and Criminisi algorithm were used to remove manually induced markers in ultrasound images. An integrated system based on TI-RADS and a mask region-based convolution neural network (Mask R-CNN) was proposed to stratify subclasses of TI-RADS 4 thyroid nodules and to segment thyroid nodules in the ultrasound images. Accuracy and the precision-recall curve were used to evaluate stratification performance, and the Dice similarity coefficient (DSC) between the segmentation of Mask R-CNN and the radiologist's contour was used to evaluate the segmentation performance of the model. RESULTS: The combined approach could significantly enhance the performance of the proposed integrated system. Overall stratification accuracy of TI-RADS 4 thyroid nodules, mean average precision and mean DSC of the proposed model in the independent test set was 90.79%, 0.8579 and 0.83, respectively. Specifically, stratification accuracy values for TI-RADS 4a, 4b and 4c thyroid nodules were 95.83%, 84.21% and 77.78%, respectively. CONCLUSION: An integrated system combining TI-RADS and a deep learning model was developed. The system can provide clinicians with not only diagnostic assistance from TI-RADS but also accurate segmentation of thyroid nodules, which improves the applicability of the system in clinical practice.

Collaborative Enhancement of Consistency and Accuracy in US Diagnosis of Thyroid Nodules Using Large Language Models.

Background Large language models (LLMs) hold substantial promise for medical imaging interpretation. However, there is a lack of studies on their feasibility in handling reasoning questions associated with medical diagnosis. Purpose To investigate the viability of leveraging three publicly available LLMs to enhance consistency and diagnostic accuracy in medical imaging based on standardized reporting, with pathology as the reference standard. Materials and Methods US images of thyroid nodules with pathologic results were retrospectively collected from a tertiary referral hospital between July 2022 and December 2022 and used to evaluate malignancy diagnoses generated by three LLMs-OpenAI's ChatGPT 3.5, ChatGPT 4.0, and Google's Bard. Inter- and intra-LLM agreement of diagnosis were evaluated. Then, diagnostic performance, including accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC), was evaluated and compared for the LLMs and three interactive approaches: human reader combined with LLMs, image-to-text model combined with LLMs, and an end-to-end convolutional neural network model. Results A total of 1161 US images of thyroid nodules (498 benign, 663 malignant) from 725 patients (mean age, 42.2 years ± 14.1 [SD]; 516 women) were evaluated. ChatGPT 4.0 and Bard displayed substantial to almost perfect intra-LLM agreement (κ range, 0.65-0.86 [95% CI: 0.64, 0.86]), while ChatGPT 3.5 showed fair to substantial agreement (κ range, 0.36-0.68 [95% CI: 0.36, 0.68]). ChatGPT 4.0 had an accuracy of 78%-86% (95% CI: 76%, 88%) and sensitivity of 86%-95% (95% CI: 83%, 96%), compared with 74%-86% (95% CI: 71%, 88%) and 74%-91% (95% CI: 71%, 93%), respectively, for Bard. Moreover, with ChatGPT 4.0, the image-to-text-LLM strategy exhibited an AUC (0.83 [95% CI: 0.80, 0.85]) and accuracy (84% [95% CI: 82%, 86%]) comparable to those of the human-LLM interaction strategy with two senior readers and one junior reader and exceeding those of the human-LLM interaction strategy with one junior reader. Conclusion LLMs, particularly integrated with image-to-text approaches, show potential in enhancing diagnostic medical imaging. ChatGPT 4.0 was optimal for consistency and diagnostic accuracy when compared with Bard and ChatGPT 3.5. © RSNA, 2024 Supplemental material is available for this article.

Enhancing the fairness of AI prediction models by Quasi-Pareto improvement among heterogeneous thyroid nodule population.

Artificial Intelligence (AI) models for medical diagnosis often face challenges of generalizability and fairness. We highlighted the algorithmic unfairness in a large thyroid ultrasound dataset with significant diagnostic performance disparities across subgroups linked causally to sample size imbalances. To address this, we introduced the Quasi-Pareto Improvement (QPI) approach and a deep learning implementation (QP-Net) combining multi-task learning and domain adaptation to improve model performance among disadvantaged subgroups without compromising overall population performance. On the thyroid ultrasound dataset, our method significantly mitigated the area under curve (AUC) disparity for three less-prevalent subgroups by 0.213, 0.112, and 0.173 while maintaining the AUC for dominant subgroups; we also further confirmed the generalizability of our approach on two public datasets: the ISIC2019 skin disease dataset and the CheXpert chest radiograph dataset. Here we show the QPI approach to be widely applicable in promoting AI for equitable healthcare outcomes.

An Improved k-Nearest Neighbor Algorithm for Recognition and Classification of Thyroid Nodules.

OBJECTIVES: To complete the task of automatic recognition and classification of thyroid nodules and solve the problem of high classification error rates when the samples are imbalanced. METHODS: An improved k-nearest neighbor (KNN) algorithm is proposed and a method for automatic thyroid nodule classification based on the improved KNN algorithm is established. In the improved KNN algorithm, we consider not only the number of class labels for various classes of data in KNNs, but also the corresponding weights. And we use the Minkowski distance measure instead of the Euclidean distance measure. RESULTS: A total of 508 ultrasound images of thyroid nodules, including 415 benign nodules and 93 malignant nodules, were used in the paper. Experimental results show the improved KNN has 0.872549 accuracy, 0.867347 precision, 1 recall, and 0.928962 F1-score. At the same time, we also considered the influence of different distance weights, the value of k, different distance measures on the classification results. CONCLUSIONS: A comparison result shows that our method has a better performance than the traditional KNN and other classical machine learning methods.

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.