Background: Pleural disease is a prevalent condition. As precision therapy advances, noninvasive imaging modalities play even more important roles in the evaluation of pleural diseases. This study investigated the diagnostic capabilities of high-frequency B-mode ultrasound (US) and contrast-enhanced US (CEUS) in terms of differentiating between benign and malignant pleural diseases.Methods: Patients with unexplained thickened pleurae were prospectively analyzed via transthoracic US.High-frequency B-mode US was used to derive the pleural thickness, morphology, and echogenicity. We analyzed the high-frequency CEUS data including the enhancement mode and time intensity curve (TIC).The cause of pleural thickening was confirmed by pleural biopsy and follow-up after the biopsy. We analyzed the differences in various ultrasonic features between the malignant and benign groups. Moreover, we plotted receiver operator curves (ROCs) and obtained the area under the curves, sensitivities, and specificities of all significant continuous variables. Multivariate logistic regression was used to assess the combined usefulness of multiple US indicators in terms of predicting malignant pleurae.Results: Thirty malignant and twenty benign thickened pleurae were finally diagnosed via pleural biopsy and at least six months of follow-up. The pleural morphology and enhancement modes significantly differed between the two groups (all P<0.05). The thickness derived from B-mode US and CEUS were significantly thicker in the malignant group (both P<0.05). Arrival time (AT) and the time to peak (TTP) of TIC were significantly shorter in the malignant group, whereas peak intensity and the area under the TIC were significantly higher in the malignant group (all P<0.05). The area under the ROC for pleural thickness derived from B-mode US was 0.819; pleural thickness derived from CEUS was 0.848; AT was 0.804; TTP was 0.750; peak intensity was 0.745; the area under the TIC was 0.743; and the combined various B-mode and CEUS parameter was 0.975.Conclusions: Pleural thickness, morphology, enhancement mode, and the TIC of high-frequency US aided the differentiation of benign from malignant pleural diseases. Combined analysis of US indicators further improved the diagnostic capability.

BackgroundImmunohistochemical microvessel density (MVD) is an early indicator of angiogenesis and it could be used to evaluate the therapeutic efficacy of non-small cell lung cancer (NSCLC). We sought to identify the ability of contrast-enhanced ultrasound (CEUS) in evaluating MVD of subpleural NSCLC.MethodsWe prospectively collected CEUS data of NSCLC confirmed by ultrasound-guided transthoracic needle biopsy from October 2019 to February 2021, The MVD of NSCLC counted by CD34-positive vessels of immunohistochemical staining. Microflow enhancement (MFE) of CEUS was divided into “dead wood”, “cotton”, and “vascular” patterns. Pathology subgroup and MVD between different MFE patterns were analyzed, respectively. The arrival time, time to peak, peak intensity (PI), and area under curve (AUC) derivefrom time-intensity curve of CEUS with MVD in NSCLC and its pathological subgroups (adenocarcinoma and squamous cell carcinoma) were subjected to correlation analysis.ResultsA total of 87 patients were included in this study, consisting of 53 cases of adenocarcinoma and 34 cases of squamous cell carcinoma with a mean MVD of 27.8 ± 12.2 mm–1. There was a significant statistical difference in MFE patterns between two pathological subgroups (p &lt; 0.05). Besides, the MVD of “cotton” and “vascular” patterns were significantly higher than that of “dead wood” pattern (both of p &lt; 0.05), whereas there was no significant difference in MVD between “cotton” pattern and “vascular” pattern. PI and AUC of CEUS were positively correlated with the MVD of NSCLC (r = 0.497, p &lt; 0.001, and r = 0.367, p &lt; 0.001, respectively). Besides, PI and AUC of CEUS were positively correlated with the MVD of squamous cell carcinoma (r = 0.802, and r = 0.663, respectively; both of p &lt; 0.001). Only the PI was positively correlated with the MVD of lung adenocarcinoma (r = 0.288, p = 0.037).ConclusionsMFE patterns and quantitative parameters of CEUS had good correlation with MVD of NSCLC, especially in squamous cell carcinoma.

No abstract

With the growth of knowledge graphs, entity descriptions are becoming extremely lengthy. Entity summarization task, aiming to generate diverse, comprehensive and representative summaries for entities, has received an increasing interest recently. In most previous methods, features are usually extracted by the hand-crafted templates. Then the feature selection and multi-user preference simulation take place, depending too much on human expertise. In this paper, a novel integration method called AutoSUM is proposed for automatic feature extraction and multi-user preference simulation to overcome the drawbacks of previous methods. There are two modules in AutoSUM: extractor and simulator. The extractor module operates automatic feature extraction based on a BiLSTM with a combined input representation including word embeddings and graph embeddings. Meanwhile, the simulator module automates multi-user preference simulation based on a well-designed twophase attention mechanism (i.e., entity-phase attention and user-phase attention). Experimental results demonstrate that AutoSUM produces the state-of-the-art performance on two widely used datasets (i.e., DBpedia and LinkedMDB) in both F-measure and MAP.