Duanshu Li

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. Methods: A pathology-proven dataset was built from 279 cytological images of thyroid nodules. The images were cropped into fragmented images and divided into a training dataset and a test dataset. VGG-16 and Inception-v3 DCNNs were trained and tested to make differential diagnoses. The characteristics of tumor cell nucleus were quantified as contours, perimeter, area and mean of pixel intensity and compared using independent Student's t-tests. Results: In the test group, the accuracy rates of the VGG-16 model and Inception-v3 on fragmented images were 97.66% and 92.75%, respectively, and the accuracy rates of VGG-16 and Inception-v3 in patients were 95% and 87.5%, respectively. The contours, perimeter, area and mean of pixel intensity of PTC in fragmented images were more than the benign nodules, which were 61.0117.10 vs 47.0024.08, p=0.000, 134.9921.42 vs 62. 4029.15, p=0.000, 1770.89627.22 vs 1157.27722.23, p=0.013, 165.8426.33 vs 132.9428.73, p=0.000), respectively. Conclusion: In summary, after training with a large dataset, the DCNN VGG-16 model showed great potential in facilitating PTC diagnosis from cytological images. The contours, perimeter, area and mean of pixel intensity of PTC in fragmented images were more than the benign nodules.

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: 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.

// Shuwen Yang 1, 2, 3, * , Qinghai Ji 1, 2, 3, * , Bin Chang 2, 4 , Yan Wang 2, 3 , Yongxue Zhu 1, 2 , Duanshu Li 1, 2 , Caiping Huang 1, 2 , Yulong Wang 1, 2 , Guohua Sun 1, 2 , Ling Zhang 1, 2 , Qing Guan 1, 2 , Jun Xiang 1, 2 , Wenjun Wei 1, 2 , Zhongwu Lu 1, 2 , Tian Liao 1, 3 , Jiao Meng 3 , Ziliang Wang 3 , Ben Ma 1, 2, 3 , Li Zhou 1, 2, 3 , Yu Wang 1, 2, 3 , Gong Yang 2, 3, 5 1 Department of Head & Neck Surgery, Fudan University Shanghai Cancer Center, Shanghai 200032, China 2 Department of Oncology, Shanghai Medical College, Fudan University, Shanghai 200032, China 3 Cancer Research Institute, Fudan University Shanghai Cancer Center, Shanghai 200032, China 4 Department of Pathology, Fudan University Shanghai Cancer Center, Shanghai 200032, China 5 Central Laboratory, The Fifth People’s Hospital of Shanghai, Fudan University, Shanghai 200240, China * Co-First authors Correspondence to: Yu Wang, email: neck130@hotmail.com Gong Yang, email: yanggong@fudan.edu.cn Keywords: STC2, pAKT, Snail, HNSCC, metastasis Received: May 23, 2016     Accepted: October 14, 2016     Published: November 15, 2016 ABSTRACT The mammalian peptide hormone stanniocalcin 2 (STC2) plays an oncogenic role in many human cancers. However, the exact function of STC2 in human head and neck squamous cell carcinoma (HNSCC) is unclear. We aimed to examine the function and clinical significance of STC2 in HNSCC. Using in vitro and in vivo assays, we show that overexpression of STC2 suppressed cell apoptosis, promoted cell proliferation, migration, invasion, and cell cycle arrest at the G1/S transition. By contrast, silencing of STC2 inhibited these activities. We further show that STC2 upregulated the phosphorylation of AKT and enhanced HNSCC metastasis via Snail-mediated increase of vimentin and decrease of E-cadherin. These responses were blocked by silencing of STC2/Snail expression or inhibition of pAKT activity. Furthermore, clinical data indicate that high STC2 expression was associated with high levels of pAKT and Snail in tumor samples from HNSCC patients with regional lymph node metastasis (P < 0.01). Thus, we conclude that STC2 controls HNSCC metastasis via the PI3K/AKT/Snail signaling axis and that targeted therapy against STC2 may be a novel strategy to effectively treat patients with metastatic HNSCC.

BACKGROUND: tRNA-derived small noncoding RNAs (sncRNAs) are mainly categorized into tRNA halves (tiRNAs) and fragments (tRFs). Biological functions of tiRNAs in human solid tumor are attracting more and more attention, but researches concerning the mechanisms in tiRNAs-mediated tumorigenesis are rarely. The direct regulatory relationship between tiRNAs and splicing-related proteins remain elusive. METHODS: Papillary thyroid carcinoma (PTC) associated tRNA fragments were screened by tRNA fragments deep sequencing and validated by qRT-PCR and Northern Blot in PTC tissues. The biological function of tRNA fragments were assessed by cell counting kit, transwells and subcutaneous transplantation tumor of nude mice. For mechanistic study, tRNA fragments pull-down, RNA immunoprecipitation, Western Blot, Immunofluorescence, Immunohistochemical staining were performed. RESULTS: Herein, we have identified a 33 nt tiRNA-Gly significantly increases in papillary thyroid cancer (PTC) based on tRFs & tiRNAs sequencing. The ectopic expression of tiRNA-Gly promotes cell proliferation and migration, whereas down-regulation of tiRNA-Gly exhibits reverse effects. Mechanistic investigations reveal tiRNA-Gly directly bind the UHM domain of a splicing-related RNA-binding protein RBM17. The interaction with tiRNA-Gly could translocate RBM17 from cytoplasm into nucleus. In addition, tiRNA-Gly increases RBM17 protein expression via inhibiting its degradation in a ubiquitin/proteasome-dependent way. Moreover, RBM17 level in tiRNA-Gly high-expressing human PTC tissues is upregulated. In vivo mouse model shows that suppression of tiRNA-Gly decreases RBM17 expression. Importantly, tiRNA-Gly can induce exon 16 splicing of MAP4K4 mRNA leading to phosphorylation of downstream signaling pathway, which is RBM17 dependent. CONCLUSIONS: Our study firstly illustrates tiRNA-Gly can directly bind to RBM17 and display oncogenic effect via RBM17-mediated alternative splicing. This fully novel model broadens our understanding of molecular mechanism in which tRNA fragment in tumor cells directly bind RNA binding protein and play a role in alternative splicing.