Artificial intelligence in thyroid ultrasound.

Artificial intelligence (AI), particularly deep learning (DL) algorithms, has demonstrated remarkable progress in image-recognition tasks, enabling the automatic quantitative assessment of complex medical images with increased accuracy and efficiency. AI is widely used and is becoming increasingly popular in the field of ultrasound. The rising incidence of thyroid cancer and the workload of physicians have driven the need to utilize AI to efficiently process thyroid ultrasound images. Therefore, leveraging AI in thyroid cancer ultrasound screening and diagnosis cannot only help radiologists achieve more accurate and efficient imaging diagnosis but also reduce their workload. In this paper, we aim to present a comprehensive overview of the technical knowledge of AI with a focus on traditional machine learning (ML) algorithms and DL algorithms. We will also discuss their clinical applications in the ultrasound imaging of thyroid diseases, particularly in differentiating between benign and malignant nodules and predicting cervical lymph node metastasis in thyroid cancer. Finally, we will conclude that AI technology holds great promise for improving the accuracy of thyroid disease ultrasound diagnosis and discuss the potential prospects of AI in this field.

Corrigendum: Association between diagnostic efficacy of acoustic radiation force impulse for benign and malignant thyroid nodules and the presence or absence of non-papillary thyroid cancer: A meta-analysis.

[This corrects the article DOI: 10.3389/fonc.2023.1007464.].

Association between diagnostic efficacy of acoustic radiation force impulse for benign and malignant thyroid nodules and the presence or absence of non-papillary thyroid cancer: A meta-analysis.

Purpose: The aim of this study was to investigate the diagnostic efficacy of Acoustic Radiation Force Impulse (ARFI) for benign and malignant thyroid nodules in the presence and absence of non-papillary thyroid cancer (NPTC) and to determine the cut-off values of Shear Wave Velocity (SWV) for the highest diagnostic efficacy of Virtual Touch Quantification (VTQ) and Virtual Touch Tissue Imaging and Quantification (VTIQ). Methods: The diagnostic accuracy of ARFI for benign and malignant thyroid nodules was assessed by pooling sensitivity, specificity and area under the curve (AUC) in each group in the presence and absence of both non-papillary thyroid glands, using histology and cytology as the gold standard. All included studies were divided into two groups according to VTQ and VTIQ, and each group was ranked according to the magnitude of the SWV cutoff value to determine the SWV cutoff interval with the highest diagnostic efficacy for VTQ and VTIQ. Results: A total of 57 studies were collected on the evaluation of ARFI for the diagnosis of benign and malignant thyroid nodules. The results showed that the presence of non-papillary thyroid carcinoma led to differences in the specificity of VTIQ for the identification of benign and malignant thyroid nodules, and the differences were statistically significant. In addition, the diagnostic efficacy of VTQ was best when the cutoff value of SWV was in the interval of 2.48-2.55 m/s, and the diagnostic efficacy of VTIQ was best when the cutoff value of SWV was in the interval of 3.01-3.15 m/s. Conclusion: VTQ and VTIQ have a high diagnostic value for benign and malignant thyroid nodules; however, when the malignant nodules in the study contain non-papillary thyroid carcinoma occupying the thyroid gland, the findings should be viewed in a comprehensive manner.

Corrigendum: Ultrasound-based deep learning using the VGGNet model for the differentiation of benign and malignant thyroid nodules: A meta-analysis.

[This corrects the article DOI: 10.3389/fonc.2022.944859.].

Diagnostic value of multiple diagnostic methods for lymph node metastases of papillary thyroid carcinoma: A systematic review and meta-analysis.

Objective: This study compared the diagnostic value of various diagnostic methods for lymph node metastasis (LNM) of papillary thyroid carcinoma (PTC) through network meta-analysis. Methods: In this experiment, databases such as CNKI, Wanfang, PubMed, and Web of Science were retrieved according to the Cochrane database, Prisma, and NMAP command manual. A meta-analysis was performed using STATA 15.0, and the value of the surface under the cumulative ranking curve (SUCRA) was used to determine the most effective diagnostic method. Quality assessments were performed using the Cochrane Collaboration's risk of bias tool, and publication bias was assessed using Deeks' funnel plot. Results: A total of 38 articles with a total of 6285 patients were included. A total of 12 diagnostic methods were used to study patients with LNM of PTC. The results showed that 12 studies were direct comparisons and 8 studies were indirect comparisons. According to the comprehensive analysis of the area of SUCRA, US+CT(86.8) had the highest sensitivity, FNAC had the highest specificity (92.4) and true positive predictive value (89.4), and FNAC+FNA-Tg had higher negative predictive value (99.4) and accuracy (86.8). In the non-invasive method, US+CT had the highest sensitivity, and the sensitivity (SEN) was [OR=0.59, 95% confidence interval (CI): (0.30, 0.89]. Among the invasive methods, the combined application of FNAC+FNA-Tg had higher diagnostic performance. The sensitivity was [OR=0.62, 95% CI: (0.26, 0.98)], the specificity (SPE) was [OR=1.12, 95% CI: (0.59, 1.64)], the positive predictive value was [OR=0.98, 95% CI: (0.59, 1.37)], the negative predictive value was [OR=0.64, 95% CI (0.38, 0.90)], and the accuracy was [OR=0.71, 95% CI: (0.31, 1.12)]. Conclusion: In the non-invasive method, the combined application of US+CT had good diagnostic performance, and in the invasive method, the combined application of FNAC+FNA-Tg had high diagnostic performance, and the above two methods were recommended.

Ultrasound-based deep learning using the VGGNet model for the differentiation of benign and malignant thyroid nodules: A meta-analysis.

Objective: The aim of this study was to evaluate the accuracy of deep learning using the convolutional neural network VGGNet model in distinguishing benign and malignant thyroid nodules based on ultrasound images. Methods: Relevant studies were selected from PubMed, Embase, Cochrane Library, China National Knowledge Infrastructure (CNKI), and Wanfang databases, which used the deep learning-related convolutional neural network VGGNet model to classify benign and malignant thyroid nodules based on ultrasound images. Cytology and pathology were used as gold standards. Furthermore, reported eligibility and risk bias were assessed using the QUADAS-2 tool, and the diagnostic accuracy of deep learning VGGNet was analyzed with pooled sensitivity, pooled specificity, diagnostic odds ratio, and the area under the curve. Results: A total of 11 studies were included in this meta-analysis. The overall estimates of sensitivity and specificity were 0.87 [95% CI (0.83, 0.91)] and 0.85 [95% CI (0.79, 0.90)], respectively. The diagnostic odds ratio was 38.79 [95% CI (22.49, 66.91)]. The area under the curve was 0.93 [95% CI (0.90, 0.95)]. No obvious publication bias was found. Conclusion: Deep learning using the convolutional neural network VGGNet model based on ultrasound images performed good diagnostic efficacy in distinguishing benign and malignant thyroid nodules. Systematic Review Registration: https://www.crd.york.ac.nk/prospero, identifier CRD42022336701.

The value of contrast-enhanced ultrasound for the diagnosis of metastatic cervical lymph nodes of papillary thyroid carcinoma: A systematic review and meta-analysis.

To investigate the diagnostic efficiency of contrast-enhanced ultrasound (CEUS) for the diagnosis of cervical lymph nodes metastasis (CLNM) of papillary thyroid carcinoma (PTC), eight available datasets of seven qualified articles before March 31, 2021 were included after a comprehensive search. Meta-analysis results showed that CEUS demonstrated acceptable diagnostic performance in the diagnosis of CLNM of PTC. Furthermore, meta-regression analysis was conducted to identify the reasons for heterogeneity and the results indicated that the criteria of CEUS for the diagnosis of CLNM in PTC need to be unified.

Diagnostic value of multiple ultrasound diagnostic techniques for axillary lymph node metastases in breast cancer: A systematic analysis and network meta-analysis.

Background: Early diagnosis of axillary lymph node metastasis is very important for the recurrence and prognosis of breast cancer. Currently, Lymph node biopsy is one of the important methods to detect lymph node metastasis in breast cancer, however, its invasiveness might bring complications to patients. Therefore, this study investigated the diagnostic performance of multiple ultrasound diagnostic methods for axillary lymph node metastasis of breast cancer. Materials and methods: In this study, we searched PubMed, Web of Science, CNKI and Wan Fang databases, conducted Bayesian network meta-analysis (NMA) on the studies that met the inclusion criteria, and evaluated the consistency of five different ultrasound imaging techniques in axillary lymph node metastasis of breast cancer. Funnel graph was used to evaluate whether it had publication bias. The diagnostic performance of each ultrasound imaging method was ranked using SUCRA. Results: A total of 22 papers were included, US+CEUS showed the highest SUCRA values in terms of sensitivity (SEN) (0.874), specificity (SPE) (0.911), positive predictive value (PPV) (0.972), negative predictive value (NPV) (0.872) and accuracy (ACC) (0.990). Conclusion: In axillary lymph node metastasis of breast cancer, the US+CEUS combined diagnostic method showed the highest SUCRA value among the five ultrasound diagnostic methods. This study provides a theoretical basis for preoperative noninvasive evaluation of axillary lymph node metastases in breast cancer patients and clinical treatment decisions. Systematic review registration: https://www.crd.york.ac.uk/PROSPERO/, identifier CRD42022351977.

HLA-B*58:01 allele is associated with augmented risk for both mild and severe cutaneous adverse reactions induced by allopurinol in Han Chinese.

AIM: Allopurinol is widely used as an effective urate-lowering drug and is one of the most frequent causes of cutaneous adverse drug reactions (cADRs). Recently, a strong association of HLA-B*58:01 with allopurinol-induced severe cADRs was identified. This study investigated the predisposition to different types of allopurinol-cADRs conferred by HLA-B*5801 in a Han population from mainland China. PATIENTS & METHODS: HLA-B genotyping was performed on 38 Chinese patients with different types of allopurinol-cADRs from 2008 to 2011. RESULTS: All the allopurinol-cADR patients carried HLA-B*58:01, in contrast with only 11.11% (7/63) in the allopurinol-tolerant patients (odds ratio [OR] = 580.07; p < 0.0001) and 13.99% (80/572) in a Han Chinese population from the human MHC database (dbMHC; OR: 471.09; p < 0.0001) carried the genotype. Each type of allopurinol cADRs revealed a statistically significant association with HLA-B*58:01. In particular, the risk of allopurinol-induced maculopapular eruption was significantly higher in patients with HLA-B*58:01 (OR: 339.00; p < 0.0001). CONCLUSION: The strong association of both the mild and severe types of allopurinol cADRs with the HLA-B*58:01 allele were observed. The results indicated that the prospective use of a genetic test of HLA-B*58:01 might reduce the prevalence of allopurinol-induced cADRs. Original submitted 7 March 2012; Revision submitted 21 May 2012.

[Study on the extraction of polyphenol from Scindapsus officinalis with ultrasonic wave technology optimized by central composite design-response surface method].

OBJECTIVE: To optimize ultrasonic extraction technology process conditions of polyphenol from Scindapsus officinalis by the response surface method. METHODS: Based on ethanol concentration, ultrasonic time, the liquid-solid ratio of single factor experiment, the principle of design for 3 star factor 3 level response surface methodology was applied. With FC extraction method for determination of polyphenols, the response surface optimization extraction conditions were studied. RESULTS: The ethanol concentration of 61.14%, ultrasonic wave extracting time of 59.73 min and the ratio of solvent volume of 27.72:1 (Extract 3 times) were selected as the optimum conditions,the extraction yield of polyphenols was 1.352%, with the theoretical 1.361% for the relative error of -0.66%. CONCLUSION: Ultrasonic extraction is a good method for saving time, energy and material,and can be applied to the polyphenols extraction. Central composite design-response surface optimization can get better ecasting results.