Background: Recurrent laryngeal nerve (RLN) injury is one of the most frequent postoperative complications of esophageal squamous cell carcinoma (ESCC) radical resection. This study aims to develop a novel scoring system to predict recurrent laryngeal nerve lymph node (RLNLN) metastases in early ESCC and explore the indications for precise RLN lymphadenectomy. Results: A total of 311 cases selected from 1,466 ESCC patients were divided into the dissection group and the control group. Age, tumor length, macroscopic tumor type, T stage, tumor location and tumor differentiation were independent predictors of RLNLN metastases. The weighted scoring system included age (+2 for <56 years), tumor length (+2 for over 4.45 cm), tumor location (+4 for upper thoracic, +2 for mid-thoracic) and macroscopic tumor type (+1 for advanced type). The total number of points estimated the probability of RLNLN metastases [low-risk (0-2 point), 0%; moderate-risk (3-4 points), 9.8%; and highrisk (>4 points), 43.4%]. Besides, the dissection group had more complications and similar survival rate when compared with the control group.Conclusions: We developed a novel scoring system that accurately estimated the risk of RLNLN metastases in early ESCC patients. RLN lymphadenectomy may be safely omitted for the patients in the low-risk subgroup.
Pseudo bounding boxes from the self-training paradigm are inevitably noisy for semi-supervised object detection. To cope with that, a dual decoupling training framework is proposed in the present study, i.e. clean and noisy data decoupling, and classification and localization task decoupling. In the first decoupling, two-level thresholds are used to categorize pseudo boxes into three groups, i.e. clean backgrounds, noisy foregrounds and clean foregrounds. With a specially designed noise-bypass head focusing on noisy data, backbone networks can extract coarse but diverse information; and meanwhile, an original head learns from clean samples for more precise predictions. In the second decoupling, we take advantage of the two-head structure for better evaluation of localization quality, thus the category label and location of a pseudo box can remain independent of each other during training. The approach of two-level thresholds is also applied to group pseudo boxes into three sections of different location accuracy. We outperform existing works by a large margin on VOC datasets, reaching 54.8 mAP(+1.8), and even up to 55.9 mAP(+1.5) by leveraging MS-COCO train2017 as extra unlabeled data. On MS-COCO benchmark, our method also achieves about 1.0 mAP improvements averaging across protocols compared with the prior state-of-the-art.
The present paper introduces sparsely supervised instance segmentation, with the datasets being fully annotated bounding boxes and sparsely annotated masks. A direct solution to this task is self-training, which is not fully explored for instance segmentation yet. In this paper, we propose MaskBooster for sparsely supervised instance segmentation (SpSIS) with comprehensive usage of pseudo masks. MaskBooster is featured with (1) dynamic and progressive pseudo masks from an online updating teacher model, (2) refining binary pseudo masks with the help of bounding box prior, (3) learning inter-class prediction distribution via knowledge distillation for soft pseudo masks. As an end-to-end and universal self-training framework, MaskBooster can empower fully supervised algorithms and boost their segmentation performance on SpSIS. Abundant experiments are conducted on COCO and BDD100K datasets and validate the effectiveness of MaskBooster. Specifically, on different COCO protocols and BDD100K, we surpass sparsely supervised baseline by a large margin for both Mask RCNN and ShapeProp. MaskBooster on SpSIS also outperforms weakly and semi-supervised instance segmentation state-of-the-art on the datasets with similar annotation budgets.
The lateral free vibration of micro-rods initially subjected to axial loads based on a nonlocal continuum theory is considered. The effects of nonlocal long-range interaction fields on the natural frequencies and vibration modes are examined. A simply supported micro-rod is taken as an example; the linear vibration responses are observed by two different methods, including the separation of variables and multiple scales analysis. The relations between the vibration mode and dimensionless coordinate and the relations between natural frequencies and nonlocal parameters are analyzed and discussed in detail. The numerical comparison shows that the theoretical results by two different approaches have a good agreement, which validates the present micro-rod model that can be used as a component of the micro-electromechanical system.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.