Objective: In this study, we exploited a VGG-16 deep convolutional neural network (DCNN) model to differentiate papillary thyroid carcinoma (PTC) from benign thyroid nodules using cytological images.

Background: To explore whether deep convolutional neural networks (DCNNs) have the potential to improve diagnostic efficiency and increase the level of interobserver agreement in the classification of thyroid nodules in histopathological slides. Methods: A total of 11,715 fragmented images from 806 patients’ original histological images were divided into a training dataset and a test dataset. Inception-ResNet-v2 and VGG-19 were trained using the training dataset and tested using the test dataset to determine the diagnostic efficiencies of different histologic types of thyroid nodules, including normal tissue, adenoma, nodular goiter, papillary thyroid carcinoma (PTC), follicular thyroid carcinoma (FTC), medullary thyroid carcinoma (MTC) and anaplastic thyroid carcinoma (ATC). Misdiagnoses were further analyzed. Results: The total 11,715 fragmented images were divided into a training dataset and a test dataset for each pathology type at a ratio of 5:1. Using the test set, VGG-19 yielded a better average diagnostic accuracy than did Inception-ResNet-v2 (97.34% vs. 94.42%, respectively). The VGG-19 model applied to 7 pathology types showed a fragmentation accuracy of 88.33% for normal tissue, 98.57% for ATC, 98.89% for FTC, 100% for MTC, 97.77% for PTC, 100% for nodular goiter and 92.44% for adenoma. It achieved excellent diagnostic efficiencies for all the malignant types. Normal tissue and adenoma were the most challenging histological types to classify. Conclusions: The DCNN models, especially VGG-19, achieved satisfactory accuracies on the task of differentiating thyroid tumors by histopathology. Analysis of the misdiagnosed cases revealed that normal tissue and adenoma were the most challenging histological types for the DCNN to differentiate, while all the malignant classifications achieved excellent diagnostic efficiencies. The results indicate that DCNN models may have potential for facilitating histopathologic thyroid disease diagnosis.

Background: To explore the ability of the deep learning network Inception-v3 to differentiate between papillary thyroid carcinomas (PTCs) and benign nodules in ultrasound images. Methods: A total of 2,836 thyroid ultrasound images from 2,235 patients were divided into a training dataset and a test dataset. Inception-v3 was trained and tested to crop the margin of the images of nodules and provide a differential diagnosis. The sizes and sonographic features of nodules were further analysed to identify the factors that may influence diagnostic efficiency. Statistical analyses included χ² and Fisher’s exact tests and univariate and multivariate analyses. Results: There were 1,275 PTCs and 1,162 benign nodules in the training group and 209 PTCs and 190 benign nodules in the test group. A margin size of 50 pixels and an input size of 384×384 showed the best outcome after training, and these parameters were selected for the test group. In the test group, the sensitivity and specificity for Inception-v3 were 93.3% (195/209) and 87.4% (166/190), respectively. Inception-v3 displayed the highest accuracy for 0.5–1.0 cm nodules. The accuracy differed according to the margin description (P=0.024). Taller nodules were more accurately diagnosed than were wider nodules (P=0.015). Microcalcification [odds ratio (OR) =0.254, 95% confidence interval (CI): 0.076–0.847, P=0.026] and taller shape (OR =0.243, 95% CI: 0.073–0.810, P=0.021) were negatively associated with misdiagnosis rate. Conclusions: Inception-v3 can achieve an excellent diagnostic efficiency. Nodules that are 0.5–1.0 cm in size and have microcalcification and a taller shape can be more accurately diagnosed by Inception-v3.

The status of central lymph node metastasis (CLNM) is crucial to determining the surgical strategies for papillary thyroid micro carcinoma (PTMC). The objective of the study was to establish a nomogram to predict the possibility of CLNM in PTMC patients. A total of 8668 PTMC patients who underwent total thyroidectomy or lobectomy with central lymph node dissection (CLND) from 2006 to 2016 were retrospectively retrieved. Univariate and multivariate analysis were performed to examine risk factors associated with CLNM. A nomogram for predicting CLNM was established and internally validated. Six variables significantly associated with CLNM were included in our model, these are age <55 years (odd ratio [OR] = 1.779, 95% confidence interval [CI],1.569-2.017; P < 0.001), male sex (OR = 1.718, 95%CI,1.543-1.913; P < 0.001), tumor size 0.5-1.0 cm (OR = 1.985,95%CI,1.761-2.238; P < 0.001), multifocal lesions (OR = 1.263, 95%CI,1.118-1.426; P < 0.001), extra thyroidal extension (ETE) (OR = 1.348, 95%CI,1.092-1.664; P = 0.005) and lateral lymph node metastasis (LLNM) (OR = 5.135, 95%CI, 4.236-6.225; P < 0.001). The discrimination of the prediction model was 0.711 (95%CI, 0.700-0.722; P < 0.001). Based on the quantified risk stratification offered by our nomogram, clinicians might have a thorough discussion with PTMC patients during the both pre- and postoperative period. Prophylactic CLND and strict postoperative evaluation may be indicated when the patients have a high nomogram score.

Objective: Extensive lateral neck dissection (LND), especially routine level V dissection, may potentially increase morbidity and have no significant benefit to survival in patients with papillary thyroid carcinoma (PTC). Our study aimed to determine the pattern and risk factors for level V lymph node metastasis (LNM) and to develop an assessment model based on a nomogram for predicting whether level V dissection is necessary.

Background: The techniques of DNA microarray and bioinformatic analysis have exhibited efficiency in identifying dysregulated gene expression in human cancers. In this study, we used integrated bioinformatics analysis to improve our understanding of the pathogenesis of papillary thyroid cancer (PTC). Methods: In this study, we integrated four Gene Expression Omnibus (GEO) datasets, GSE33630, GSE35570, GSE60542 and GSE29265, including 136 normal samples and 157 PTC specimens. The contents of the four datasets are based on GPL570, an Affymetrix Human Genome U133 Plus 2.0 array. Gene ontology (GO) analysis was used to identify characteristic the biological attributes of differentially expressed genes (DEGs) between PTC and normal samples. GO annotation was performed on the DEGs obtained, and the process relied on the DAVID online tool. Kyoto Encyclopedia of Genes and Genomes (KEGG) approach enrichment analyses were adopted to obtain the basic functions of the DEGs. The KOBAS online analysis database was used to complete DEG KEGG pathway comparison and analysis. The search tool (STRING) database was mainly used to search for interacting genes and complete the construction of protein-protein interaction (PPI) networks. Results: Five hundred-ninety DEGs were consistently expressed in the four datasets; 327 of them were upregulated, while 263 were downregulated. Ten DEGs, including five upregulated (ENTPD1, THRSP, KLK10, ADAMTS9, MIR31HG) and five downregulated (SCARA5, EPHB1, CHRDL1, LOC440934, FOXP2) genes, were randomly selected for q-PCR in our own tissue samples to validate the integrated data. The most highly enriched GO terms were extracellular exosome (GO:0070062), cell adhesion (GO:0070062), positive regulation of gene expression (GO:0010628), and extracellular matrix (ECM) organization (GO:0030198). KEGG pathway analysis was performed, and it was found that abnormally expressed genes effectively participated in pathways such as tyrosine metabolism, complement and coagulation cascades, cell adhesion molecules (CAMs), transcriptional misregulation and ECM-receptor interaction pathways. Conclusions: Five hundred-ninety DEGs were identified in PTC by integrated microarray analysis. The GO and KEGG analyses presented here suggest that the DEGs were enriched in extracellular exosome, tyrosine metabolism, CAMs, complement and coagulation cascades, transcriptional misregulation and ECM-receptor interaction pathways. Functional studies of PTC should focus on these pathways.