Convolutional Neural Network-Based Computer-Assisted Diagnosis of Hashimoto's Thyroiditis on Ultrasound

Transcriptomic profiling and machine learning reveal novel RNA signatures for enhanced molecular characterization of Hashimoto's thyroiditis

While ultrasonography effectively diagnoses Hashimoto's thyroiditis (HT), exploring its transcriptomic landscape could reveal valuable insights into disease mechanisms. This study aimed to identify HT-associated RNA signatures and investigate their potential for enhanced molecular characterization. Samples comprising 31 HT patients and 30 healthy controls underwent RNA sequencing of peripheral blood. Differential expression analysis identified transcriptomic features, which were integrated using multi-omics factor analysis. Pathway enrichment, co-expression, and regulatory network analyses were performed. A novel machine-learning model was developed for HT molecular characterization using stacking techniques. HT patients exhibited increased thyroid volume, elevated tissue hardness, and higher antibody levels despite being in the early subclinical stage. Analysis identified 79 HT-associated transcriptomic features (3 mRNA, 6 miRNA, 64 lncRNA, 6 circRNA). Co-expression (77 nodes, 266 edges) and regulatory (18 nodes, 45 edges) networks revealed significant hub genes and modules associated with HT. Enrichment analysis highlighted dysregulation in immune system, cell adhesion and migration, and RNA/protein regulation pathways. The novel stacking-model achieved 95% accuracy and 97% AUC for HT molecular characterization. This study demonstrates the value of transcriptome analysis in uncovering HT-associated signatures, providing insights into molecular changes and potentially guiding future research on disease mechanisms and therapeutic strategies.

Deblurring masked image modeling for ultrasound image analysis

Recently, large pretrained vision foundation models based on masked image modeling (MIM) have attracted unprecedented attention and achieved remarkable performance across various tasks. However, the study of MIM for ultrasound imaging remains relatively unexplored, and most importantly, current MIM approaches fail to account for the gap between natural images and ultrasound, as well as the intrinsic imaging characteristics of the ultrasound modality, such as the high noise-to-signal ratio. In this paper, motivated by the unique high noise-to-signal ratio property in ultrasound, we propose a deblurring MIM approach specialized to ultrasound, which incorporates a deblurring task into the pretraining proxy task. The incorporation of deblurring facilitates the pretraining to better recover the subtle details within ultrasound images that are vital for subsequent downstream analysis. Furthermore, we employ a multi-scale hierarchical encoder to extract both local and global contextual cues for improved performance, especially on pixel-wise tasks such as segmentation. We conduct extensive experiments involving 280,000 ultrasound images for the pretraining and evaluate the downstream transfer performance of the pretrained model on various disease diagnoses (nodule, Hashimoto's thyroiditis) and task types (classification, segmentation). The experimental results demonstrate the efficacy of the proposed deblurring MIM, achieving state-of-the-art performance across a wide range of downstream tasks and datasets. Overall, our work highlights the potential of deblurring MIM for ultrasound image analysis, presenting an ultrasound-specific vision foundation model.

FBENet: Feature-Level Boosting Ensemble Network for Hashimoto's Thyroiditis Ultrasound Image Classification

Distinguishing Hashimoto's thyroiditis (HT) lesions from ordinary thyroid tissues is difficult with ultrasound images. Challenges in achieving high performance of HT ultrasound image classification include the low resolution, blurred features and large area of irrelevant noise. To address these problems, we propose a Feature-level Boosting Ensemble Network (FBENet) for HT ultrasound image classification. Specifically, to capture the features of suspicious HT lesions efficiently, an Ensemble Feature Boosting Module (EFBM) is introduced into the feature-level ensemble to boost the blurred features. Then, the spatial attention mechanism is adopted in backbone models to improve the feature focusing performance and representation ability. Furthermore, feature-level ensemble technique is employed in the training process to achieve more comprehensive feature representation ability. Experimentally, FBENet was trained on 6,503 HT ultrasound images, and tested on 1,626 HT ultrasound images with 82.92% accuracy and 89.24% AUC on average.

Association of autoimmune thyroid disease with type 1 diabetes mellitus and its ultrasonic diagnosis and management

As a common hyperglycemic disease, type 1 diabetes mellitus (T1DM) is a complicated disorder that requires a lifelong insulin supply due to the immune-mediated destruction of pancreatic β cells. Although it is an organ-specific autoimmune disorder, T1DM is often associated with multiple other autoimmune disorders. The most prevalent concomitant autoimmune disorder occurring in T1DM is autoimmune thyroid disease (AITD), which mainly exhibits two extremes of phenotypes: hyperthyroidism [Graves' disease (GD)] and hypo-thyroidism [Hashimoto's thyroiditis, (HT)]. However, the presence of comorbid AITD may negatively affect metabolic management in T1DM patients and thereby may increase the risk for potential diabetes-related complications. Thus, routine screening of thyroid function has been recommended when T1DM is diagnosed. Here, first, we summarize current knowledge regarding the etiology and pathogenesis mechanisms of both diseases. Subsequently, an updated review of the association between T1DM and AITD is offered. Finally, we provide a relatively detailed review focusing on the application of thyroid ultrasonography in diagnosing and managing HT and GD, suggesting its critical role in the timely and accurate diagnosis of AITD in T1DM.

HT-RCM: Hashimoto's Thyroiditis Ultrasound Image Classification Model Based on Res-FCT and Res-CAM

The early lesions of Hashimoto's thyroiditis are inconspicuous, and the ultrasonic features of these early lesions are indistinguishable from other thyroid diseases. This paper proposes a Hashimoto Thyroiditis ultrasound image classification model HT-RCM which consists of a Residual Full Convolution Transformer (Res-FCT) model and a Residual Channel Attention Module (Res-CAM). To collect the low-order information caused by hypoechoic signals accurately, the residual connection is injected between FCTs to form Res-FCT which helps HT-RCM superimpose the low-order input information and high-order output information together. Res-FCT can make HT-RCM focus more on hypoechoic information while avoiding gradient dispersion. The initial feature map is inserted into Res-FCT again through a down-sampling component, which further helps HT-RCM exact multi-level original semantic information in the ultrasound image. Res-CAM is constructed by implementing a residual connection between a channel attention module and a convolution layer. Res-CAM can effectively increase the weights of the lesion channels while suppressing the weights of the noise channels, which makes HT-RCM focus more on the lesion regions. The experimental results on our collected dataset show that HT-RCM outperforms the mainstream models and obtains state-of-the-art performance in HT ultrasound image classification.

A narrative review of deep learning in thyroid imaging: current progress and future prospects

Background and Objective

Deep learning (DL) has contributed substantially to the evolution of image analysis by unlocking increased data and computational power. These DL algorithms have further facilitated the growing trend of implementing precision medicine, particularly in areas of diagnosis and therapy. Thyroid imaging, as a routine means to screening for thyroid diseases on large-scale populations, is a massive data source for the development of DL models. Thyroid disease is a global health problem and involves structural and functional changes. The objective of this study was to evaluate the general rules and future directions of DL networks in thyroid medical image analysis through a review of original articles published between 2018 and 2023.

Methods

We searched for English-language articles published between April 2018 and September 2023 in the databases of PubMed, Web of Science, and Google Scholar. The keywords used in the search included artificial intelligence or DL, thyroid diseases, and thyroid nodule or thyroid carcinoma.

Key Content and Findings

The computer vision tasks of DL in thyroid imaging included classification, segmentation, and detection. The current applications of DL in clinical workflow were found to mainly include management of thyroid nodules/carcinoma, risk evaluation of thyroid cancer metastasis, and discrimination of functional thyroid diseases.

Conclusions

DL is expected to enhance the quality of thyroid images and provide greater precision in the assessment of thyroid images. Specifically, DL can increase the diagnostic accuracy of thyroid diseases and better inform clinical decision-making.

Automatic detection of thyroid nodules with a real-time artificial intelligence system in a real clinical scenario and the associated influencing factors

BACKGROUND

At present, most articles mainly focused on the diagnosis of thyroid nodules by using artificial intelligence (AI), and there was little research on the detection performance of AI in thyroid nodules.

OBJECTIVE

To explore the value of a real-time AI based on computer-aided diagnosis system in the detection of thyroid nodules and to analyze the factors influencing the detection accuracy.

METHODS

From June 1, 2022 to December 31, 2023, 224 consecutive patients with 587 thyroid nodules were prospective collected. Based on the detection results determined by two experienced radiologists (both with more than 15 years experience in thyroid diagnosis), the detection ability of thyroid nodules of radiologists with different experience levels (junior radiologist with 1 year experience and senior radiologist with 5 years experience in thyroid diagnosis) and real-time AI were compared. According to the logistic regression analysis, the factors influencing the real-time AI detection of thyroid nodules were analyzed.

RESULTS

The detection rate of thyroid nodules by real-time AI was significantly higher than that of junior radiologist (P = 0.013), but lower than that of senior radiologist (P = 0.001). Multivariate logistic regression analysis showed that nodules size, superior pole, outside (near carotid artery), close to vessel, echogenicity (isoechoic, hyperechoic, mixed-echoic), morphology (not very regular, irregular), margin (unclear), ACR TI-RADS category 4 and 5 were significant independent influencing factors (all P < 0.05). With the combination of real-time AI and radiologists, junior and senior radiologist increased the detection rate to 97.4% (P < 0.001) and 99.1% (P = 0.015) respectively.

CONCLUSONS

The real-time AI has good performance in thyroid nodule detection and can be a good auxiliary tool in the clinical work of radiologists.

The diagnostic value of ultrasound on different-sized thyroid nodules based on ACR TI-RADS

OBJECTIVES

The thyroid nodule is one of the most common endocrine system diseases. Risk classification models based on ultrasonic features have been created by multiple professional societies, including the American College of Radiology (ACR), which published the Thyroid Imaging Reporting and Data System (TI-RADS) in 2017. The effect of the size in the diagnostic value of ultrasound remains not well defined. The purposes of our study aims to explore diagnostic value of the ACR TI-RADS on different-sized thyroid nodules.

METHODS

A total of 1183 thyroid nodules were selected from 952 patients with thyroid nodules confirmed by surgical pathology from January 2021 to October 2022. Based on the maximum diameters of the nodules, they were stratified into groups A ( ≤ 10 mm), B ( > 10 mm, < 20 mm) and C ( ≥ 20 mm). The ultrasonic features of the thyroid nodules in each group were evaluated and scored based on ACR TI-RADS, and the receiver operating characteristic curve (ROC) was plotted to determine the optimal cut-off value for the ACR TI-RADS scores and categories in each group. Finally, the diagnostic efficacy of ACR TI-RADS on different-sized thyroid nodules was analyzed.

RESULTS

Among the 1183 thyroid nodules, 340 were benign, 10 were low-risk and 833 were malignant. For the convenience of statistical analysis, low-risk thyroid nodules were classified as malignant in this study. The ACR TI-RADS scores and categorical levels of malignant thyroid nodules in each group were higher than those of benign ones (p < 0.05). The areas under the ROCs (AUCs) plotted based on scores were 0.741, 0.907, and 0.904 respectively in the three groups, and the corresponding optimal cut-off values were > 6 points, > 5 points and > 4 points respectively. While the AUCs of the ACR TI-RADS categories were 0.668, 0.855, and 0.887 respectively in each group, with the optimal cut-off values were all > TR4. Besides, for thyroid nodules of larger sizes, ACR TI-RADS exhibited weaker sensitivity with lower positive prediction value (PPV), but the specificity and negative prediction value (NPV) were both higher, presenting with statistically significant differences (p < 0.05).

CONCLUSION

For thyroid nodules of different sizes, the diagnostic efficacy of ACR TI-RADS varies as well. The system shows better diagnostic efficacy on thyroid nodules of > 10 mm than on those ≤ 10 mm. Considering the favorable prognosis of thyroid microcarcinoma and the low diagnostic efficacy of ACR TI-RADS on it, the scoring and classification of thyroid micro-nodules can be left out in appropriate cases, so as to avoid the over-diagnosis and over-treatment of thyroid microcarcinoma to a certain extent.

HTC-Net: Hashimoto's thyroiditis ultrasound image classification model based on residual network reinforced by channel attention mechanism

Convolutional neural network (CNN) is efficient in extracting and aggregating local features in the spatial dimension of the images. However, obtaining the inapparent texture information of the low-echo area in the ultrasound images is not easy, and it is especially challenging for the early lesion recognition in Hashimoto's thyroiditis (HT) ultrasound images. In this paper, a HT ultrasound image classification model HTC-Net based on residual network reinforced by channel attention mechanism is proposed. HTC-Net strengthens the features of the important channels by reinforced channel attention mechanism through which the high-level semantic information is enchanced and the low-level semantic information is suppressed. Residual network assists HTC-Net focus on the key local areas of the ultrasound images while pay attention to the global semantic information. Furthermore, in order to solve the problem of uneven distribution caused by large amount of difficult-to-classify samples in the data sets, a new feature loss function TanCELoss with weight factor dynamically adjusting is constructed. TanCELoss function can better assist HTC-Net to transform difficult-to-classify samples into easy-to-classify samples gradually, and improve the balancing distribution of the samples. The experiments are implemented based on data sets collected by the Endocrinology Department of four branches from Guangdong Provincial Hospital of Chinese Medicine. Both quantitative testing and visualization results show that HTC-Net obtains STOA performance for early lesions recognition in HT ultrasound images. HTC-Net has great application value especially under the condition of owning only small data samples.

Computer image analysis with artificial intelligence: a practical introduction to convolutional neural networks for medical professionals

Artificial intelligence tools, particularly convolutional neural networks (CNNs), are transforming healthcare by enhancing predictive, diagnostic, and decision-making capabilities. This review provides an accessible and practical explanation of CNNs for clinicians and highlights their relevance in medical image analysis. CNNs have shown themselves to be exceptionally useful in computer vision, a field that enables machines to 'see' and interpret visual data. Understanding how these models work can help clinicians leverage their full potential, especially as artificial intelligence continues to evolve and integrate into healthcare. CNNs have already demonstrated their efficacy in diverse medical fields, including radiology, histopathology, and medical photography. In radiology, CNNs have been used to automate the assessment of conditions such as pneumonia, pulmonary embolism, and rectal cancer. In histopathology, CNNs have been used to assess and classify colorectal polyps, gastric epithelial tumours, as well as assist in the assessment of multiple malignancies. In medical photography, CNNs have been used to assess retinal diseases and skin conditions, and to detect gastric and colorectal polyps during endoscopic procedures. In surgical laparoscopy, they may provide intraoperative assistance to surgeons, helping interpret surgical anatomy and demonstrate safe dissection zones. The integration of CNNs into medical image analysis promises to enhance diagnostic accuracy, streamline workflow efficiency, and expand access to expert-level image analysis, contributing to the ultimate goal of delivering further improvements in patient and healthcare outcomes.

Comparison of Diagnostic Values of ACR TI-RADS versus C-TIRADS Scoring and Classification Systems for the Elderly Thyroid Cancers

Purpose

To compare the diagnostic value of the Thyroid Imaging Reporting and Data System (TI-RADS) of the American College of Radiology (ACR) versus the Chinese Thyroid Imaging Reporting and Data System (C-TIRADS) scoring and classification system for elderly thyroid cancers.

Patients and Methods

A total of 512 nodules from 465 patients aged ≥60 with surgical pathology-proven thyroid nodules were enrolled in our study. The ultrasound features of thyroid nodules were independently evaluated by the ACR TI-RADS and C-TIRADS classification systems, and the receiver operating characteristic curve (ROC) was plotted. The optimal cut-off values of the ACR TI-RADS and C-TIRADS scoring and classification systems for diagnosing elderly thyroid nodules were estimated, and the diagnostic efficacy was analyzed.

Results

The ACR TI-RADS and C-TIRADS scores and classifications of thyroid cancers were both higher than benign nodules (both P < 0.05). The area under the curve (AUC) of ACR TI-RADS and C-TIRADS scoring and classification systems were 0.861, 0.897, 0.879, and 0.900, respectively, and the AUC of the scoring system was greater than the classification system for both criteria. When the Youden index was the highest, the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the ACR TI-RADS scoring and classification systems were consistent, ie, they were 89.66%, 41.70%, 89.93%, and 59.00%, respectively; the sensitivity, specificity, PPV, and NPV of the C-TIRADS scoring and classification systems were also consistent, ie, they were 88.71%, 44.26%, 90.23%, 59.69%, respectively. The diagnostic efficacy between the two systems was not statistically significant.

Conclusion

ACR TI-RADS and C-TIRADS systems had relatively high diagnostic efficacy for elderly thyroid cancer. The diagnostic efficiency of the scoring systems of ACR TI-RADS and C-TIRADS were higher than the classification systems. In addition, the two systems had high clinical practical values, while there is still a significant risk of missed diagnosis.

Deep Learning-Based Evaluation of Ultrasound Images for Benign Skin Tumors

In this study, a combined convolutional neural network for the diagnosis of three benign skin tumors was designed, and its effectiveness was verified through quantitative and statistical analysis. To this end, 698 sonographic images were taken and diagnosed at the Department of Dermatology at Severance Hospital in Seoul, Korea, between 10 November 2017 and 17 January 2020. Through an empirical process, a convolutional neural network combining two structures, which consist of a residual structure and an attention-gated structure, was designed. Five-fold cross-validation was applied, and the train set for each fold was augmented by the Fast AutoAugment technique. As a result of training, for three benign skin tumors, an average accuracy of 95.87%, an average sensitivity of 90.10%, and an average specificity of 96.23% were derived. Also, through statistical analysis using a class activation map and physicians' findings, it was found that the judgment criteria of physicians and the trained combined convolutional neural network were similar. This study suggests that the model designed and trained in this study can be a diagnostic aid to assist physicians and enable more efficient and accurate diagnoses.

Comparing Diagnostic Efficacy of C-TIRADS Positive Features on Different Sizes of Thyroid Nodules

Purpose

To explore the diagnostic value of positive features in the Chinese Thyroid Imaging Reporting and Data System (C-TIRADS) for thyroid nodules of different sizes.

Patients and Methods

A total of 1864 patients with 2347 thyroid nodules were selected from January 2021 to December 2022 and assessed according to C-TIRADS. According to the maximum diameter, nodules were divided into the A1 group (≤10 mm), A2 group (>10 mm,<20 mm), and A3 group (≥20 mm). With surgical pathology as the golden standard, the receiver operating characteristic curves (ROC) were constructed, and each group's area under the curve (AUC) was calculated. The diagnostic value of positive features in C-TIRADS for different sizes of thyroid nodules was analyzed.

Results

In all groups, malignant thyroid nodules had a higher incidence of positive features than benign nodules (P < 0.05). In A1 group, the diagnostic efficiency of C-TIRADS positive features for thyroid nodules was vertical orientation> ill-defined/irregular margin or extrathyroidal extension> solid composition> markedly hypoechoic> microcalcifications. The AUCs were 0.718, 0.675, 0.609, 0.558, and 0.581, respectively. In A2 group, the diagnostic efficacy of each positive features for thyroid nodules was ill-defined/irregular margins or extra-thyroid invasion> solid composition> microcalcifications> markedly hypoechoic> vertical orientation. The AUCs were 0.854, 0.730, 0.719, 0.670, and 0.609, respectively. In A3 group, the diagnostic efficacy of each positive features for thyroid nodules was ill-defined/irregular margin or extrathyroidal extension> microcalcifications> solid composition> vertical orientation> markedly hypoechoic. The AUCs were 0.847, 0.778, 0.767, 0.584, and 0.560, respectively.

Conclusion

C-TIRADS positive features exhibited different diagnostic efficacy for thyroid nodules of various sizes, especially for thyroid nodules ≤10 mm, for which all positive features had low diagnostic efficacy.

Diagnosis of thyroid disease using deep convolutional neural network models applied to thyroid scintigraphy images: a multicenter study

Objectives

The aim of this study was to improve the diagnostic performance of nuclear medicine physicians using a deep convolutional neural network (DCNN) model and validate the results with two multicenter datasets for thyroid disease by analyzing clinical single-photon emission computed tomography (SPECT) image data.

Methods

In this multicenter retrospective study, 3194 SPECT thyroid images were collected for model training (n=2067), internal validation (n=514) and external validation (n=613). First, four pretrained DCNN models (AlexNet, ShuffleNetV2, MobileNetV3 and ResNet-34) for were tested multiple medical image classification of thyroid disease types (i.e., Graves' disease, subacute thyroiditis, thyroid tumor and normal thyroid). The best performing model was then subjected to fivefold cross-validation to further assess its performance, and the diagnostic performance of this model was compared with that of junior and senior nuclear medicine physicians. Finally, class-specific attentional regions were visualized with attention heatmaps using gradient-weighted class activation mapping.

Results

Each of the four pretrained neural networks attained an overall accuracy of more than 0.85 for the classification of SPECT thyroid images. The improved ResNet-34 model performed best, with an accuracy of 0.944. For the internal validation set, the ResNet-34 model showed higher accuracy (p < 0.001) when compared to that of the senior nuclear medicine physician, with an improvement of nearly 10%. Our model achieved an overall accuracy of 0.931 for the external dataset, a significantly higher accuracy than that of the senior physician (0.931 vs. 0.868, p < 0.001).

Conclusion

The DCNN-based model performed well in terms of diagnosing thyroid scintillation images. The DCNN model showed higher sensitivity and greater specificity in identifying Graves' disease, subacute thyroiditis, and thyroid tumors compared to those of nuclear medicine physicians, illustrating the feasibility of deep learning models to improve the diagnostic efficiency for assisting clinicians.

Leveraging deep learning to identify calcification and colloid in thyroid nodules

Background

Both calcification and colloid in thyroid nodules are reflected as echogenic foci in ultrasound images. However, calcification and colloid have significantly different probabilities of malignancy. We explored the performance of a deep learning (DL) model in distinguishing the echogenic foci of thyroid nodules as calcification or colloid.

Methods

We conducted a retrospective study using ultrasound image sets. The DL model was trained and tested on 30,388 images of 1127 nodules. All nodules were pathologically confirmed. The area under the receiver-operator characteristic curve (AUC) was employed as the primary evaluation index.

Results

The YoloV5 (You Only Look Once Version 5) transfer learning model for thyroid nodules based on DL detection showed that the average sensitivity, specificity, and accuracy of distinguishing echogenic foci in the test 1 group (n = 192) was 78.41%, 91.36%, and 77.81%, respectively. The average sensitivity, specificity, and accuracy of the three radiologists were 51.14%, 82.58%, and 61.29%, respectively. The average sensitivity, specificity, and accuracy of distinguishing small echogenic foci in the test 2 group (n = 58) was 70.17%, 77.14%, and 73.33%, respectively. Correspondingly, the average sensitivity, specificity, and accuracy of the radiologists were 57.69%, 63.29%, and 59.38%.

Conclusions

The study demonstrated that DL performed far better than radiologists in distinguishing echogenic foci of thyroid nodules as calcifications or colloid.

Artificial Intelligence in Thyroidology: A Narrative Review of the Current Applications, Associated Challenges, and Future Directions

Background: The use of artificial intelligence (AI) in health care has grown exponentially with the promise of facilitating biomedical research and enhancing diagnosis, treatment, monitoring, disease prevention, and health care delivery. We aim to examine the current state, limitations, and future directions of AI in thyroidology. Summary: AI has been explored in thyroidology since the 1990s, and currently, there is an increasing interest in applying AI to improve the care of patients with thyroid nodules (TNODs), thyroid cancer, and functional or autoimmune thyroid disease. These applications aim to automate processes, improve the accuracy and consistency of diagnosis, personalize treatment, decrease the burden for health care professionals, improve access to specialized care in areas lacking expertise, deepen the understanding of subtle pathophysiologic patterns, and accelerate the learning curve of less experienced clinicians. There are promising results for many of these applications. Yet, most are in the validation or early clinical evaluation stages. Only a few are currently adopted for risk stratification of TNODs by ultrasound and determination of the malignant nature of indeterminate TNODs by molecular testing. Challenges of the currently available AI applications include the lack of prospective and multicenter validations and utility studies, small and low diversity of training data sets, differences in data sources, lack of explainability, unclear clinical impact, inadequate stakeholder engagement, and inability to use outside of the research setting, which might limit the value of their future adoption. Conclusions: AI has the potential to improve many aspects of thyroidology; however, addressing the limitations affecting the suitability of AI interventions in thyroidology is a prerequisite to ensure that AI provides added value for patients with thyroid disease.

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.

Investigating the diagnostic efficiency of a computer-aided diagnosis system for thyroid nodules in the context of Hashimoto's thyroiditis

Objectives

This study aims to investigate the efficacy of a computer-aided diagnosis (CAD) system in distinguishing between benign and malignant thyroid nodules in the context of Hashimoto's thyroiditis (HT) and to evaluate the role of the CAD system in reducing unnecessary biopsies of benign lesions.

Methods

We included a total of 137 nodules from 137 consecutive patients (mean age, 43.5 ± 11.8 years) who were histopathologically diagnosed with HT. The two-dimensional ultrasound images and videos of all thyroid nodules were analyzed by the CAD system and two radiologists with different experiences according to ACR TI-RADS. The diagnostic cutoff values of ACR TI-RADS were divided into two categories (TR4 and TR5), and then the sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) of the CAD system and the junior and senior radiologists were compared in both cases. Moreover, ACR TI-RADS classification was revised according to the results of the CAD system, and the efficacy of recommended fine-needle aspiration (FNA) was evaluated by comparing the unnecessary biopsy rate and the malignant rate of punctured nodules.

Results

The accuracy, sensitivity, specificity, PPV, and NPV of the CAD system were 0.876, 0.905, 0.830, 0.894, and 0.846, respectively. With TR4 as the cutoff value, the AUCs of the CAD system and the junior and senior radiologists were 0.867, 0.628, and 0.722, respectively, and the CAD system had the highest AUC (P < 0.0001). With TR5 as the cutoff value, the AUCs of the CAD system and the junior and senior radiologists were 0.867, 0.654, and 0.812, respectively, and the CAD system had a higher AUC than the junior radiologist (P < 0.0001) but comparable to the senior radiologist (P = 0.0709). With the assistance of the CAD system, the number of TR4 nodules was decreased by both junior and senior radiologists, the malignant rate of punctured nodules increased by 30% and 22%, and the unnecessary biopsies of benign lesions were both reduced by nearly half.

Conclusions

The CAD system based on deep learning can improve the diagnostic performance of radiologists in identifying benign and malignant thyroid nodules in the context of Hashimoto's thyroiditis and can play a role in FNA recommendations to reduce unnecessary biopsy rates.

Convolutional Neural Network for Predicting Thyroid Cancer Based on Ultrasound Elastography Image of Perinodular Region

We aimed to develop deep learning models based on perinodular regions' shear-wave elastography (SWE) images and ultrasound (US) images of thyroid nodules (TNs) and determine their performances in predicting thyroid cancer. A total of 1747 American College of Radiology Thyroid Imaging Reporting & Data System 4 (TR4) thyroid nodules (TNs) in 1582 patients were included in this retrospective study. US images, SWE images, and 2 quantitative SWE parameters (maximum elasticity of TNs; 5-point average maximum elasticity of TNs) were obtained. Based on US and SWE images of TNs and perinodular tissue, respectively, 7 single-image convolutional neural networks (CNN) models [US, internal SWE, 0.5 mm SWE, 1.0 mm SWE, 1.5 mm SWE, 2.0 mm SWE of perinodular tissue, and whole SWE region of interest (ROI) image] and another 6 fusional-image CNN models (US + internal SWE, US + 0.5 mm SWE, US + 1.0 mm SWE, US + 1.5 mm SWE, US + 2.0 mm SWE, US + ROI SWE) were established using RestNet18. All of the CNN models and quantitative SWE parameters were built on a training cohort (1247 TNs) and evaluated on a validation cohort (500 TNs). In predicting thyroid cancer, US + 2.0 mm SWE image CNN model obtained the highest area under the curve in 10 mm < TNs ≤ 20 mm (0.95 for training; 0.92 for validation) and TNs > 20 mm (0.95 for training; 0.92 for validation), while US + 1.0 mm SWE image CNN model obtained the highest area under the curve in TNs ≤ 10 mm (0.95 for training; 0.92 for validation). The CNN models based on the fusion of SWE segmentation images and US images improve the radiological diagnostic accuracy of thyroid cancer.

Deep learning to diagnose Hashimoto's thyroiditis from sonographic images

Hashimoto's thyroiditis (HT) is the main cause of hypothyroidism. We develop a deep learning model called HTNet for diagnosis of HT by training on 106,513 thyroid ultrasound images from 17,934 patients and test its performance on 5051 patients from 2 datasets of static images and 1 dataset of video data. HTNet achieves an area under the receiver operating curve (AUC) of 0.905 (95% CI: 0.894 to 0.915), 0.888 (0.836-0.939) and 0.895 (0.862-0.927). HTNet exceeds radiologists' performance on accuracy (83.2% versus 79.8%; binomial test, p < 0.001) and sensitivity (82.6% versus 68.1%; p < 0.001). By integrating serologic markers with imaging data, the performance of HTNet was significantly and marginally improved on the video (AUC, 0.949 versus 0.888; DeLong's test, p = 0.004) and static-image (AUC, 0.914 versus 0.901; p = 0.08) testing sets, respectively. HTNet may be helpful as a tool for the management of HT.