Long non-coding RNA LINC00152 promotes cell growth and invasion of papillary thyroid carcinoma by regulating the miR-497/BDNF axis.

Long intergenic non-coding RNA 152 (LINC00152) was reported to be tightly linked to tumorigenesis and progression in multiple cancers. However, its biological role and modulatory mechanism in papillary thyroid carcinoma (PTC) has not been elucidated. In this study, we determined the expression levels of LINC00152 in PTC tissues and cell lines by quantitative real time polymerase chain reaction (qRT-PCR). Cell proliferation, colony formation, migration, and invasion were measured by a Cell Counting Kit-8 assay, colony formation analysis, wound healing, and transwell invasion assay, respectively. A luciferase reporter assay and qRT-PCR were used to determine whether LINC00152 interacts with miR-497 directly. We established a xenograft mouse model to examine the underlying molecular mechanism and effect of LINC00152 on tumor growth in vivo. We found that LINC00152 expression was significantly increased in PTC tissues and derived cell lines. LINC00152 knockdown significantly inhibited proliferation, colony formation, migration, and invasion in vitro, and impaired tumor growth in vivo. We revealed that LINC00152 functioned as a competing endogenous RNA to the miR-497 sponge, downregulating its downstream target brain-derived neurotrophic factor (BDNF), which is an oncogene in thyroid cancer. These findings suggest that LINC00152 is responsible for PTC cell proliferation and invasion and exerts its function by regulating the miR-497/BDNF axis.

Folate Supplementation for Methotrexate Therapy in Patients With Rheumatoid Arthritis: A Systematic Review.

OBJECTIVE: To review the evidence for benefits and harms of folate (folic acid or folinic acid) supplementation on methotrexate (MTX) treatment for rheumatoid arthritis (RA), to assess whether or not folate supplementation would reduce MTX toxicity or reduce MTX benefits, and to decide whether a higher MTX dosage is essential. METHODS: We performed a sensitive search strategy and searched systematically the Medline, Embase, Web of Science and Cochrane Library databases from inception to 2 June 2016. Abstracts from major rheumatology meetings and major trial registers were also searched to retrieve all randomized controlled trials that interested us. RESULTS: Seven studies with 709 patients were included. No significant heterogeneity was found between these trials. For RA patients treated with MTX, those supplied with folate were less likely to have elevated transaminase (odds ratio [OR] 0.15; 95% confidence interval [95% CI] 0.10, 0.23 [p < 0.00001]) and gastrointestinal side-effects such as nausea and vomiting (OR 0.71; 95% CI 0.51, 0.99 [p = 0.04]). Folate appeared to promote compliance to MTX as it reduced patient withdrawal compared to placebo (OR 0.29; 95% CI 0.21, 0.42 [p < 0.00001]). There was no statistical difference for mouth sores between folate and placebo (OR 0.83; 95% CI 0.57, 1.22 [p = 0.35]). As the markers of disease activity in those trials were not consistent, it was impossible to decide whether folate supplementation reduced MTX efficacy. Besides, we compared high-dose folate (≥25 mg per week) and low-dose folate (≤10 mg per week) on MTX efficacy, finding no statistical difference (OR 2.07; 95% CI 0.81, 5.30 [p = 0.13]), nor on MTX toxicity (OR 1.56; 95% CI 0.80,3.04 [p = 0.19]). CONCLUSION: Folate supplementation can reduce the incidence of hepatotoxicity and gastrointestinal side-effects of MTX in patients with RA. It can also reduce patient withdrawal from MTX treatment. Although it tended to reduce mouth sores, it had no statistical significance. No significant difference was found between high-dose folate and low-dose folate on MTX efficacy or toxicity.

Nanoparticles based on glycyrrhetinic acid modified porphyrin for photodynamic therapy of cancer.

Self-assembled small molecules, as a novel form of drug presentation, have splendid capabilities for water stability and cell endocytosis. Photodynamic therapy (PDT) is regarded as a promising cancer treatment because it is less invasive and has fewer side effects. In this work, an amphiphilic glycyrrhetinic acid-porphyrin conjugate (TPP-GA) was designed and synthesized, and this could self-assemble into nanoparticles (TPP-GA NPs) in water. The morphology of the TPP-GA NPs was observed using transmission electron microscopy (TEM) and their stability in water was revealed by dynamic light scattering (DLS). The PDT effect of the TPP-GA NPs was confirmed by means of in vitro experiment, after photostability verification. It could be conceived that the application of TPP-GA NPs for PDT is a feasible choice and this work shows progress in the self-assembly of amphiphilic small organic molecules.

Radioactive Iodine Ablation Decrease Recurrences in Papillary Thyroid Microcarcinoma with Lateral Lymph Node Metastasis in Chinese Patients.

INTRODUCTION: We aimed to carry out a retrospective study from a single institution to determine whether radioactive iodine (RAI) ablation decreases the risk of recurrence of papillary thyroid microcarcinoma (PTMC) patients that presented with lateral lymph node metastasis (LLNM). METHODS: We retrospectively analyzed a total of 6361 consecutive PTMC patients who initially underwent surgery for the treatment of thyroid carcinoma at the First Hospital of Jilin University, between January 2005 and February 2015. Altogether, 137 patients with PTMC with LLNM have been enrolled in our study. RESULTS: The disease-free survival period was significantly shorter for the RAI (-) patients than for the RAI (+) patients (p = 0.0074 by the log-rank test). The disease-free survival rates at 5 and 10 years were 93.02 and 87.5%, respectively, in the RAI (-) group and 100 and 96.81%, respectively, in the RAI (+) group. CLNM ratio and LLNM ratio were factors identified for multivariate analysis by Cox's proportional hazards method yielding risk ratios of 7.281 [CI 1.804-17.554; p = 0.010] and 1.157 [CI 1.0125-9.381; p = 0.048] in the RAI (-) group. CONCLUSION: Taken together, RAI may be beneficial for PTMC with LLNM, especially when CLNM ratio or LLNM ratio was greater than 0.5.

Total thyroidectomy may be more reasonable as initial surgery in unilateral multifocal papillary thyroid microcarcinoma: a single-center experience.

BACKGROUND: The extent of surgery in patients with unilateral multifocal papillary thyroid microcarcinoma (UMPTMC) remains to be controversial. Aimed to improve surgical management of UMPTMC, this study was performed to identify the recurrence of UMPTMC and analyze its predictive factors. METHODS: This study was approved by the Ethical Committee of The First Hospital of the Jilin University, and written informed consent was given by participants for their clinical records to be used in this study. We retrospectively analyzed a total of 97 consecutive patients who underwent initial surgery for the treatment of UMPTMC at The First Hospital of Jilin University, between October 2005 and October 2006. RESULTS: Altogether, 97 patients of UMPTMC have been enrolled in our study, in which 57 cases were performed with hemithyroidectomy (HT) while other 40 cases with total thyroidectomy (TT). The sum diameter of all tumors >1 cm was more frequent in HT group than in TT group (40.35 vs 20%; p = 0.046). Positive central lymph nodes were found more frequently in the TT patients than in the HT patients (80 vs 59.65%; p = 0.046). Tumor recurrence was seen more frequently in the HT cases than in the TT cases (26 vs 5%; p = 0.007). The disease-free survival period was significantly shorter for the HT patients than for the TT patients (p = 0.0059 by the log-rank test). The disease-free survival rates at 5 and 10 years were 91.23 and 73.68%, respectively, in the HT group and 100 and 92.5%, respectively, in the TT group. Univariate analysis by Cox's proportional hazards method showed male gender, sum diameter of all tumors >1 cm, and central lymph node metastases (CLNM) to be risk factors for recurrence of HT patients. Male gender and sum diameter >1 cm were factors identified for multivariate analysis by Cox's proportional hazards method which yielded risk ratios of 3.037 [CI 1.026-8.988; p = 0.045] and 5.475 [CI 1.389-21.572; p = 0.015] in the HT group. CONCLUSIONS: In summary, with an increased risk of recurrence, TT may be more reasonable as initial surgery in UMPTMC, especially with male gender and total tumor diameter greater than 1 cm.

The role of parathyroid autotransplantation for hypoparathyroidism following total thyroidectomy with bilateral central neck dissection.

Background: Hypoparathyroidism is the most common complication for patients who undergo total thyroidectomy (TT) with bilateral central lymph node dissection (BCLND). The objective of this retrospective study was to investigate the relationship between parathyroid autotransplantation (PA) and postoperative hypoparathyroidism. Materials and Methods: Four hundred and sixty-five patients with papillary thyroid carcinoma (PTC) who underwent TT with BCLND (including prophylactic and therapeutic BCLND) by the same surgeon were enrolled in this retrospective study. They were divided into five groups based on the number of PAs. Group 0 was defined as no PA, while Group 1, 2, 3, and 4 were considered as 1, 2, 3, and 4 PAs during TT with BCLND, respectively. Results: Transient and permanent hypoparathyroidism occurred in 83 (17.8%) and 2 (0.4%) patients who underwent TT and BCLND, respectively. The incidence of transient hypoparathyroidism increased gradually with an increase in the number of PAs. Compared with the previous group, Groups 2 and 3 had significantly more cases of transient hypoparathyroidism (p=0.03 and p=0.04, respectively). All cases of permanent hypoparathyroidism occurred in the patients without PA. Compared with Group 0, there were more removed central lymph nodes (RCLNs) in patients with one PA. Furthermore, Group 2 had more metastatic central lymph nodes(MCLNs) and RCLNs than Group 1.The number of PAs was the only identified risk factor for transient hypoparathyroidism after the multivariate logistic regression analysis. The median parathyroid hormone level recovered to the normal range within 1 month after surgery. Conclusion: With an increasing number of PAs, the possibility of transient hypoparathyroidism also increases in patients with PTC who undergo TT and BCLND. Considering the rapid recovery of transient hypoparathyroidism in 1 month, two PAs during TT and BCLND could be a good choice, leading to an increase in the central lymph node yield and no permanent hypoparathyroidism. However, this conclusion should be validated in future multicenter prospective studies.

Analgesic effects of pectoserratus plane block and intermediate cervical plexus block for transaxillary endoscopic thyroidectomy: a prospective randomized controlled trial.

BACKGROUND: Transaxillary endoscopic thyroidectomy (TAET) is favored for its favorable cosmetic outcomes and the preservation of anterior cervical function. Despite these benefits, postoperative analgesia has traditionally relied on pharmacological interventions, and regional anesthetic procedures may be an alternative method. This study aimed to evaluate the efficacy of an ultrasound-guided pectoserratus plane block (PSPB) combined with an intermediate cervical plexus block (ICPB) for TAET. METHODS: Forty patients undergoing TAET were randomized into two groups: the nerve block group (N.=20) received ultrasound guided PSPB with 20 ml of 0.375% ropivacaine and ICPB with 8 mL of 0.3% ropivacaine, while the control group (N.=20) received no block. The primary outcome was the Visual Analog Scale (VAS) scores for postoperative neck and axillary pain at different time points (1, 6, 12, 24 h) during rest and movement post-TAET. The secondary outcomes included intraoperative remifentanil consumption, incidence of postoperative nausea and vomiting (PONV), number of remedial analgesic requirements, and patient satisfaction postoperatively. RESULTS: Compared to the control group, patients in the nerve block had significantly lower VAS scores of the neck and axilla whether at rest or movement, and 1, 6, 12, and 24 h postoperatively (P<0.0125). The nerve block group showed higher patient satisfaction (P<0.001). No difference was observed in intraoperative remifentanil consumption, need for rescue analgesics, or other adverse effects 48 h postoperatively. CONCLUSIONS: Ultrasound-guided PSPB with ICPB significantly alleviated postoperative pain and improved patient satisfaction with TAET.

Extremely Sparse Networks via Binary Augmented Pruning for Fast Image Classification.

Network pruning and binarization have been demonstrated to be effective in neural network accelerator design for high speed and energy efficiency. However, most existing pruning approaches achieve a poor tradeoff between accuracy and efficiency, which on the other hand, has limited the progress of neural network accelerators. At the same time, binary networks are highly efficient, however, a large accuracy gap exists between binary networks and their full-precision counterparts. In this article, we investigate the merits of extremely sparse networks with binary connections for image classification through software-hardware codesign. More specifically, we first propose a binary augmented extremely pruning method that can achieve ~98% sparsity with small accuracy degradation. Then we design the hardware architecture based on the resulting sparse and binary networks, which extensively explores the benefits of extreme sparsity with negligible resource consumption introduced by binary branch. Experiments on large-scale ImageNet classification and field-programmable gate array (FPGA) demonstrate that the proposed software-hardware architecture can achieve a prominent tradeoff between accuracy and efficiency.

A novel nomogram for identifying high-risk patients among active surveillance candidates with papillary thyroid microcarcinoma.

Objective: Active surveillance (AS) has been recommended as the first-line treatment strategy for low-risk (LR) papillary thyroid microcarcinoma (PTMC) according to the guidelines. However, preoperative imaging and fine-needle aspiration could not rule out a small group of patients with aggressive PTMC with large-volume lymph node micro-metastasis, extrathryoidal invasion to surrounding soft tissue, or high-grade malignancy from the AS candidates. Methods: Among 2,809 PTMC patients, 2,473 patients were enrolled in this study according to the inclusion criteria. Backward stepwise multivariate logistic regression analysis was used to filter clinical characteristics and ultrasound features to identify independent predictors of high-risk (HR) patients. A nomogram was developed and validated according to selected risk factors for the identification of an HR subgroup among "LR" PTMC patients before operation. Results: For identifying independent risk factors, multivariable logistic regression analysis was performed using the backward stepwise method and revealed that male sex [3.91 (2.58-5.92)], older age [0.94 (0.92-0.96)], largest tumor diameter [26.7 (10.57-69.22)], bilaterality [1.44 (1.01-2.3)], and multifocality [1.14 (1.01-2.26)] were independent predictors of the HR group. Based on these independent risk factors, a nomogram model was developed for predicting the probability of HR. The C index was 0.806 (95% CI, 0.765-0.847), which indicated satisfactory accuracy of the nomogram in predicting the probability of HR. Conclusion: Taken together, we developed and validated a nomogram model to predict HR of PTMC, which could be useful for patient counseling and facilitating treatment-related decision-making.

Optimization-Based Post-Training Quantization With Bit-Split and Stitching.

Deep neural networks have shown great promise in various domains. Meanwhile, problems including the storage and computing overheads arise along with these breakthroughs. To solve these problems, network quantization has received increasing attention due to its high efficiency and hardware-friendly property. Nonetheless, most existing quantization approaches rely on the full training dataset and the time-consuming fine-tuning process to retain accuracy. Post-training quantization does not have these problems, however, it has mainly been shown effective for 8-bit quantization. In this paper, we theoretically analyze the effect of network quantization and show that the quantization loss in the final output layer is bounded by the layer-wise activation reconstruction error. Based on this analysis, we propose an Optimization-based Post-training Quantization framework and a novel Bit-split optimization approach to achieve minimal accuracy degradation. The proposed framework is validated on a variety of computer vision tasks, including image classification, object detection, instance segmentation, with various network architectures. Specifically, we achieve near-original model performance even when quantizing FP32 models to 3-bit without fine-tuning.

Comparisons of different approaches and incisions of thyroid surgery and selection strategy.

To date, the traditional open thyroid surgery via a low collar incision remains the standard approach for patients undergoing thyroidectomy. However, this conventional approach will inevitably leave patients a neck scar and even cause a variety of complications such as paresthesia, hypesthesia, and other uncomfortable sensations. With the progress in surgical techniques, especially in endoscopic surgery, and the increasing desire for cosmetic and functional outcomes, various new approaches for thyroidectomy have been developed to avoid or decrease side effects. Some of these alternative approaches have obvious advantages compared with traditional surgery and have already been widely used in the treatment of thyroid disease, but each has its limitations. This review aims to evaluate and compare the different approaches to thyroidectomy to help surgeons make the proper treatment strategy for different individuals.

High throughput sequencing revealed enhanced cell cycle signaling in SLE patients.

The multi-system involvement and high heterogeneity of systemic lupus erythematosus (SLE) pose great challenges to its diagnosis and treatment. The purpose of the current study is to identify genes and pathways involved in the pathogenesis of SLE. High throughput sequencing was performed on the PBMCs from SLE patients. We conducted differential gene analysis, gene ontology (GO) analysis, kyoto encyclopedia of genes and genomes (KEGG) analysis, and quantitative real-time PCR (qRT-PCR) verification. Protein-protein interaction (PPI) analysis, alternative splicing analysis, and disease correlation analysis were conducted on some key pathogenic genes as well. Furthermore, si-CDC6 was used for transfection and cell proliferation was monitored using a cell counting kit-8 (CCK-8) assay. We identified 2495 differential genes (1494 upregulated and 1001 downregulated) in SLE patients compared with healthy controls. The significantly upregulated genes were enriched in the biological process-related GO terms of the cell cycle, response to stress, and chromosome organization. KEGG enrichment analysis revealed 7 significantly upregulated pathways including SLE, alcoholism, viral carcinogenesis, cell cycle, proteasome, malaria, and transcriptional misregulation in cancer. We successfully verified some differential genes on the SLE pathway and the cell cycle pathway. CDC6, a key gene in the cell cycle pathway, had remarkably higher MXE alternative splicing events in SLE patients than that in controls, which may explain its significant upregulation in SLE patients. We found that CDC6 participates in the pathogenesis of many proliferation-related diseases and its levels are positively correlated with the severity of SLE. Knockdown of CDC6 suppressed the proliferation of Hela cells and PBMCs from SLE patients in vitro. We identified SLE-related genes and their alternative splicing events. The cell cycle pathway and the cell cycle-related biological processes are over-activated in SLE patients. We revealed a higher incidence of MXE events of CDC6, which may lead to its high expression in SLE patients. Upregulated cell cycle signaling and CDC6 may be related to the hyperproliferation and pathogenesis of SLE.

Toward Accurate Binarized Neural Networks With Sparsity for Mobile Application.

While binarized neural networks (BNNs) have attracted great interest, popular approaches proposed so far mainly exploit the symmetric sign function for feature binarization, i.e., to binarize activations into -1 and +1 with a fixed threshold of 0. However, whether this option is optimal has been largely overlooked. In this work, we propose the Sparsity-inducing BNN (Si-BNN) to quantize the activations to be either 0 or +1, which better approximates ReLU using 1-bit. We further introduce trainable thresholds into the backward function of binarization to guide the gradient propagation. Our method dramatically outperforms the current state-of-the-art, lowering the performance gap between full-precision networks and BNNs on mainstream architectures, achieving the new state-of-the-art on binarized AlexNet (Top-1 50.5%), ResNet-18 (Top-1 62.2%), and ResNet-50 (Top-1 68.3%). At inference time, Si-BNN still enjoys the high efficiency of bit-wise operations. In our implementation, the running time of binary AlexNet on the CPU can be competitive with the popular GPU-based deep learning framework.

Prednisolone induces sleep disorders via inhibition of melatonin secretion by the circadian rhythm in zebrafish.

Prednisolone (PN) is a glucocorticoid (GC) analog that is clinically used to treat allergic inflammation and autoimmune diseases. However, the long-term use of GC-like drugs results in many side effects, among which sleep disorders caused by PN have attracted much attention. Many studies have showed that GCs indirectly cause sleep disorders by disrupting the circadian rhythm of the peripheral biological clock. However, the detailed mechanism of this effect in zebrafish remains unclear. In the present study, we aimed to study the pharmacology and toxicology of PN by analyzing the sleep phenotype and internal circadian oscillation of zebrafish. Exposure of zebrafish to PN resulted in decreased melatonin secretion and shortened sleep time. Additionally, analysis of the internal circadian rhythm of the zebrafish revealed that the expression of per and cry was significantly upregulated, resulting in a significant delay in the phase of the zebrafish behavioral rhythm. A dual-luciferase reporter assay further revealed that PN repressed per2 and cry1aa expression via the GC receptor (GR), which inhibited aanat2 expression. This caused a decrease in melatonin secretion and led to sleep disorders. The findings of this study highlight the mechanisms underlying the effects of GCs on sleep.

Retraction: MicroRNA-497 inhibits thyroid cancer tumor growth and invasion by suppressing BDNF.

[This retracts the article DOI: 10.18632/oncotarget.13747.].

[Retracted] Silencing NOB1 enhances doxorubicin antitumor activity of the papillary thyroid carcinoma in vitro and in vivo.

Following the publication of this paper, it was drawn to the Editors' attention by a concerned reader that certain of the western blotting assay data shown in Figs. 2 and 5, and the tumour images shown in Fig. 6A, were strikingly similar to data appearing in different form in other articles by different authors. Owing to the fact that the contentious data in the above article had already been published elsewhere, or were already under consideration for publication, prior to its submission to Oncology Reports, the Editor has decided that this paper should be retracted from the Journal. After having been in contact with the authors, they agreed with the decision to retract the paper. The Editor apologizes to the readership for any inconvenience caused. [the original article was published in Oncology Reports 33: 1551-1559, 2015; DOI: 10.3892/or.2015.3730].

Supraclavicular Approach of Lobectomy Improves Quality of Life for Patients With Unilateral Papillary Thyroid Microcarcinoma: A Prospective Cohort Study.

Objective: Postoperative neck symptoms, including pain, swelling, uncomfortable feelings during swallowing, and incision adhesion formation, are common in patients after lobectomy through the traditional middle neck approach. A new unilateral supraclavicular approach is proposed to protect the anterior cervical region and reduce related complications. The aim of this study is to investigate the efficacy, safety, and advantages of the supraclavicular approach in lobectomy for unilateral papillary thyroid microcarcinoma (PTMC). Methods: Two hundred sixty-three patients were recruited into either a conventional middle group (CM) or a new supraclavicular (NS) group. Clinicopathological features, surgically related variables, and postoperative symptoms were recorded. Quality of life (QOL) of all patients was assessed by the 12-item short-form health survey (SF-12) and thyroid cancer-specific QOL (THYCA-QoL) questionnaire in 3 and 12 months. Results: There were no statistically significant differences in clinicopathological features (including sex, age, multifocality, extrathyroidal extension, histological variants, largest tumor diameter, Hashimoto's thyroiditis, metastasized central lymph node, removed central lymph node, surgeon, BRAF mutation, and follow-up duration), hospitalization (including hospital cost, surgery time, and blood loss during surgery), and complications between the two groups. Patients who underwent lobectomy through the NS approach had significantly better SF-12 physical, mental, and THYCA-QoL than the CM group patients in both 3 and 12 months (all p < 0.001). Moreover, the NS group had a shorter hospitalization time. Conclusion: In conclusion, the NS approach for lobectomy is a safe and effective method for reducing postoperative symptoms and increasing QOL in patients with unilateral PTMC in both 3 and 12 months' follow-up.

Isolation, purification, gene cloning and expression of antifungal protein from Bacillus amyloliquefaciens MG-3.

The antifungal protein MG-3A was isolated from Bacillus amyloliquefaciens MG-3, and was purified and identified. The results showed that antifungal protein MG-3A was likely a serine protease with a molecular weight of ~48 kDa. The serine protease exhibited a broad antifungal spectrum and effectively extended the shelf-life of loquat fruit up to 25 d. The antifungal protein MG-3A showed good stabilities to temperature, pH and protease K. Primers were designed according to the mass spectrum of antifungal protein and the comparison with proteins in the NCBI database. The serine protease gene MG-3A from B. amyloliquefaciens genome was isolated and cloned using PCR. The prokaryotic expression plasmid pET28a-MG-3A was constructed and used to express the antimicrobial protein in vitro. The SDS-PAGE results showed that the recombinant protein expressed in Escherichia coli BL21 (DE3) was highly soluble. Affinity chromatography was used to purify the recombinant protein and its antifungal activity was evaluated.

Modeling and Optimization of Electrostatic Film Actuators Based on the Method of Moments.

Electrostatic film actuators represent a promising new approach to drive a soft robot, but they lack a comprehensive model to link the design parameters and actuation performance, making actuator design and parameter optimization challenging. To solve this problem, we build a mathematical model based on the method of moments by assuming that each electrode consists of a large number of line charges. This model can directly deduce fluctuation in thrust and adhesive forces during actuator movement, as well as the distribution of electric potential and field strength, for analysis and optimization. It consumes shorter computing time and fewer computing resources, but with comparable accuracy, in comparison with previous indirect means. It is validated by results from both previous studies and on-site experiments. Based on this model, we generate numerous values of actuator output force for different structural parameters. By analyzing the tendency, we summarize a parameter optimization workflow and write an open-sourced program as an example to facilitate the parameter selection for actuator design starting from scratch.

Unsupervised Network Quantization via Fixed-Point Factorization.

The deep neural network (DNN) has achieved remarkable performance in a wide range of applications at the cost of huge memory and computational complexity. Fixed-point network quantization emerges as a popular acceleration and compression method but still suffers from huge performance degradation when extremely low-bit quantization is utilized. Moreover, current fixed-point quantization methods rely heavily on supervised retraining using large amounts of the labeled training data, while the labeled data are hard to obtain in the real-world applications. In this article, we propose an efficient framework, namely, fixed-point factorized network (FFN), to turn all weights into ternary values, i.e., {-1, 0, 1}. We highlight that the proposed FFN framework can achieve negligible degradation even without any supervised retraining on the labeled data. Note that the activations can be easily quantized into an 8-bit format; thus, the resulting networks only have low-bit fixed-point additions that are significantly more efficient than 32-bit floating-point multiply-accumulate operations (MACs). Extensive experiments on large-scale ImageNet classification and object detection on MS COCO show that the proposed FFN can achieve about more than 20× compression and remove most of the multiply operations with comparable accuracy. Codes are available on GitHub at https://github.com/wps712/FFN.