An integrated model incorporating deep learning, hand-crafted radiomics and clinical and US features to diagnose central lymph node metastasis in patients with papillary thyroid cancer.

OBJECTIVE: To evaluate the value of an integrated model incorporating deep learning (DL), hand-crafted radiomics and clinical and US imaging features for diagnosing central lymph node metastasis (CLNM) in patients with papillary thyroid cancer (PTC). METHODS: This retrospective study reviewed 613 patients with clinicopathologically confirmed PTC from two institutions. The DL model and hand-crafted radiomics model were developed using primary lesion images and then integrated with clinical and US features selected by multivariate analysis to generate an integrated model. The performance was compared with junior and senior radiologists on the independent test set. SHapley Additive exPlanations (SHAP) plot and Gradient-weighted Class Activation Mapping (Grad-CAM) were used for the visualized explanation of the model. RESULTS: The integrated model yielded the best performance with an AUC of 0.841. surpassing that of the hand-crafted radiomics model (0.706, p < 0.001) and the DL model (0.819, p = 0.26). Compared to junior and senior radiologists, the integrated model reduced the missed CLNM rate from 57.89% and 44.74-27.63%, and decreased the rate of unnecessary central lymph node dissection (CLND) from 29.87% and 27.27-18.18%, respectively. SHAP analysis revealed that the DL features played a primary role in the diagnosis of CLNM, while clinical and US features (such as extrathyroidal extension, tumour size, age, gender, and multifocality) provided additional support. Grad-CAM indicated that the model exhibited a stronger focus on thyroid capsule in patients with CLNM. CONCLUSION: Integrated model can effectively decrease the incidence of missed CLNM and unnecessary CLND. The application of the integrated model can help improve the acceptance of AI-assisted US diagnosis among radiologists.

A Survey on Non-Autoregressive Generation for Neural Machine Translation and Beyond.

Non-autoregressive (NAR) generation, which is first proposed in neural machine translation (NMT) to speed up inference, has attracted much attention in both machine learning and natural language processing communities. While NAR generation can significantly accelerate inference speed for machine translation, the speedup comes at the cost of sacrificed translation accuracy compared to its counterpart, autoregressive (AR) generation. In recent years, many new models and algorithms have been designed/proposed to bridge the accuracy gap between NAR generation and AR generation. In this paper, we conduct a systematic survey with comparisons and discussions of various non-autoregressive translation (NAT) models from different aspects. Specifically, we categorize the efforts of NAT into several groups, including data manipulation, modeling methods, training criterion, decoding algorithms, and the benefit from pre-trained models. Furthermore, we briefly review other applications of NAR models beyond machine translation, such as grammatical error correction, text summarization, text style transfer, dialogue, semantic parsing, automatic speech recognition, and so on. In addition, we also discuss potential directions for future exploration, including releasing the dependency of KD, reasonable training objectives, pre-training for NAR, and wider applications, etc. We hope this survey can help researchers capture the latest progress in NAR generation, inspire the design of advanced NAR models and algorithms, and enable industry practitioners to choose appropriate solutions for their applications.

The Predictive Value of ACR TI-RADS Classification for Central Lymph Node Metastasis of Papillary Thyroid Carcinoma: A Retrospective Study.

BACKGROUND: The aim of this retrospective study was to evaluate the risk factors for central lymph node metastasis (CLNM) in papillary thyroid carcinoma (PTC), according to the guidelines of the 2017 Thyroid Imaging Report and Data System (TI-RADS) published by the American College of Radiology (ACR). METHODS: This study included a retrospective analysis of 844 patients with PTC who were pathologically diagnosed, treated with central lymph node dissection, and divided into CLNM and nonmetastatic groups. Univariate and multivariate analyses were performed to determine the relationship between the TI-RADS score and CLNM. RESULTS: Among 844 patients, 439 developed CLNM, with a metastasis rate of 52% and a TI-RADS score of 9.42 ± 2.262, which were higher than those of the non-CLNM group (P < 0.05). Univariate analysis demonstrated that the sex, location, maximum diameter of the nodule, multifocality, margin, shape, calcification, and TI-RADS score were related to CLNM (P < 0.05 for all). However, multivariate logistic regression analysis demonstrated that female, maximum diameter of the nodule, multifocality, a taller-than-wide shape, and high TI-RADS score were the independent risk factors for CLNM (P < 0.05 for all). CONCLUSION: The TI-RADS score combined with sex, nodule size, shape, and multifocality has a certain predictive effect on CLNM, which can provide a reference to the clinicians for further treatment strategies.

CircFADS2: A potential prognostic biomarker of colorectal cancer.

IMPACT STATEMENT: Colorectal cancer (CRC) is the third most common malignancy worldwide with the second highest mortality rate. Although multidisciplinary cooperative therapies are helpful for improving the survival of CRC patients, the prognosis remains poor. Therefore, it is imperative to seek new biomarkers for the development of individualized treatment for each CRC patient. Circular RNA, an endogenous transcript with specific covalent closed loop, exhibits higher stability, conservation and expression abundance than the corresponding linear component and thus may be utilized as a promised biomarker. Although the majority of studies have focused on circular RNA expression profiling in various types of cancers, evidence supporting their critical role in the diagnosis and prognosis of CRC is limited. This study aimed to screen and identify novel circular RNA biomarkers of CRC by chip analysis and qPCR verification, and to highlight their potential as targets for CRC prognosis, and therapy.

Formin-like 3 regulates RhoC/FAK pathway and actin assembly to promote cell invasion in colorectal carcinoma.

AIM: To clarify the underlying mechanism of formin-like 3 (FMNL3) in the promotion of colorectal carcinoma (CRC) cell invasion. METHODS: The in vitro biological function analyses of FMNL3 were performed by gain- and loss-of function approaches. Changes in the F-actin cytoskeleton were detected by the technologies of phalloidin-TRITC labeling and confocal microscopy. The signaling pathway mediated by FMNL3 was explored by western blot, gelatin zymograph assay, co-immunoprecipitation (co-IP), immunofluorescence co-localization, and glutathione S-transferase (GST) pull-down assay. RESULTS: The in vitro experimental results showed that FMNL3 significantly promoted the proliferation, invasion, and migration of CRC cells (P < 0.05 and P < 0.01). Moreover, FMNL3 regulated the remodeling of actin-based protrusions such as filopodia and lamellipodia in a RhoC-dependent manner. The western blot and gelatin zymograph assay results indicated that FMNL3 was involved in the RhoC/ focal adhesion kinase (FAK) pathway and acted as an effector of RhoC to activate the downstream signaling of p-FAK as well as p-MAPK and p-AKT. This resulted in the increased expression of matrix metalloproteinase 2 (MMP2), matrix metalloproteinase 9 (MMP9) and vascular endothelial growth factor (VEGF), and the subsequent promotion of CRC cell invasion. The results of TAE226, U0126 or Ly294002 treatment confirmed an essential role of FMNL3 in activation of the RhoC/FAK pathway and the subsequent promotion of CRC invasion. Co-IP, co-localization and GST pull-down assays showed the direct interaction of FMNL3 with RhoC in vivo and in vitro. CONCLUSION: FMNL3 regulates the RhoC/FAK signaling pathway and RhoC-dependent remodeling of actin-based protrusions to promote CRC invasion.

Increased expression of formin-like 3 contributes to metastasis and poor prognosis in colorectal carcinoma.

Formin-like 3 (FMNL3), a member of diaphanous-related formins subfamily, plays an important role in cytoskeleton reorganization, cell adhesion and cancer cell invasion in vitro. This study aimed to explore the expression of FMNL3 in colorectal carcinoma (CRC) cell-lines and tissues, and further evaluate its prognostic value and correlation with the clinicopathological parameters, and also investigate the effects of FMNL3 gene silencing on the growth and metastasis of CRC in vivo. Immunohistochemical analysis showed that FMNL3 protein was distributed in a punctuate aggregation pattern and located mainly in the cytoplasm of glandular cavity side, close to the nucleus of CRC cells. The positive rate of FMNL3 expression was 87.5% (84/96) in CRC, which was significantly higher than that in adjacent normal mucosa (30%, 9/30). Moreover, FMNL3 protein expressed far more in primary CRC with metastasis and corresponding lymph nodes metastatic CRC than in primary CRC without metastasis. Increased expression of FMNL3 was closely correlated with tumor size, differentiation, serosal invasion, and both lymph node metastasis and distant metastasis. However, it was not correlated with patients' age and gender. According to Kaplan-Meier survival analyses, patients with FMNL3 high expression level had lower overall survival rate than that with FMNL3 low expression level. Univariate and multivariate analyses revealed that high FMNL3 expression was a significant and independent prognostic predictor of patients with CRC. In addition, FMNL3 mRNA and protein levels were substantially up-regulated in CRC-metastasis-derived cell lines, as compared to those in primary-CRC-derived ones. FMNL3 gene silencing suppressed the growth and metastasis of CRC in vivo. In conclusion, FMNL3 plays an important role in the progression and metastasis of CRC and may be a novel potential prognostic predictor and therapeutic target for patients with CRC.