Grading of clear cell renal cell carcinoma using diffusion MRI with a fractional order calculus model.

BACKGROUND: The fractional order calculus (FROC) model has been developed to describe restrained motion of water molecules as well as microstructural heterogeneity, providing a novel tool for non-invasive tumor grading. PURPOSE: To evaluate the role of the FROC model in characterizing clear cell renal cell carcinoma (ccRCC) grades. MATERIAL AND METHODS: A total of 59 patients diagnosed with ccRCC were included in this prospective study. The diffusion metrics derived from the mono-exponential model (apparent diffusion coefficient [ADC]), intra-voxel incoherent motion [IVIM] model [D, D*, f], and FROC model [Dfroc, ß, µ]) were calculated and compared between low- and high-grade ccRCCs. Binary logistic regression analysis was performed to establish the diagnostic models. Receiver operating characteristic (ROC) analysis and DeLong test were performed to evaluate and compare the diagnostic performance of metrics in grading ccRCC. RESULTS: All the metrics except D* and f exhibited statistical differences between low- and high-grade ccRCCs. ROC analysis showed individual FROC parameters, µ, Dfroc, and ß, outperformed ADC and IVIM parameters in grading ccRCC. For single parameter, µ demonstrated the highest AUC value, sensitivity, and diagnostic accuracy in discriminating the two ccRCC groups while ß exhibited the optimal specificity. Importantly, the combination of Dfroc, µ, and ß could further improve the diagnostic performance. CONCLUSION: The FROC parameters were superior to ADC and IVIM parameters in grading ccRCC, indicating the great potential of the FROC model in distinguishing low- and high-grade ccRCCs.

A Joint Extraction System Based on Conditional Layer Normalization for Health Monitoring.

Natural language processing (NLP) technology has played a pivotal role in health monitoring as an important artificial intelligence method. As a key technology in NLP, relation triplet extraction is closely related to the performance of health monitoring. In this paper, a novel model is proposed for joint extraction of entities and relations, combining conditional layer normalization with the talking-head attention mechanism to strengthen the interaction between entity recognition and relation extraction. In addition, the proposed model utilizes position information to enhance the extraction accuracy of overlapping triplets. Experiments on the Baidu2019 and CHIP2020 datasets demonstrate that the proposed model can effectively extract overlapping triplets, which leads to significant performance improvements compared with baselines.

POSTAR2: deciphering the post-transcriptional regulatory logics.

Post-transcriptional regulation of RNAs is critical to the diverse range of cellular processes. The volume of functional genomic data focusing on post-transcriptional regulation logics continues to grow in recent years. In the current database version, POSTAR2 (http://lulab.life.tsinghua.edu.cn/postar), we included the following new features and data: updated ∼500 CLIP-seq datasets (∼1200 CLIP-seq datasets in total) from six species, including human, mouse, fly, worm, Arabidopsis and yeast; added a new module 'Translatome', which is derived from Ribo-seq datasets and contains ∼36 million open reading frames (ORFs) in the genomes from the six species; updated and unified post-transcriptional regulation and variation data. Finally, we improved web interfaces for searching and visualizing protein-RNA interactions with multi-layer information. Meanwhile, we also merged our CLIPdb database into POSTAR2. POSTAR2 will help researchers investigate the post-transcriptional regulatory logics coordinated by RNA-binding proteins and translational landscape of cellular RNAs.

Ultrasound Shear Wave Elastography for Evaluation of Diaphragm Stiffness in Patients with Stable COPD: A Pilot Trial.

OBJECTIVES: Skeletal muscle dysfunction is one of the most common comorbidities in chronic obstructive pulmonary disease (COPD). The occurrence of respiratory failure in COPD is common and leads to the patient's death. The diaphragm is the most important muscle in the respiratory system and plays a key role in the onset of respiratory failure. This study explores the feasibility of ultrasound shear wave elastography (SWE) to measure diaphragmatic stiffness and evaluates its changes in COPD patients. METHODS: In total, 77 participants (43 patients with stable COPD and 34 healthy controls) were enrolled. All subjects underwent complete diaphragmatic ultrasound SWE measurements and pulmonary function tests. The diaphragmatic stiffness was indicated via diaphragmatic shear wave velocity (SWV) at functional residual capacity (FRC). A trained operator performed the ultrasound SWE examinations of the first 15 healthy controls thrice to assess the reliability of diaphragmatic SWE. RESULTS: A good to excellent reliability was found in diaphragmatic SWV at FRC (ICC = 0.93, 95%CI 0.82-0.98). As compared to the control group, the diaphragmatic SWV at FRC was considerably high in the COPD group (median 2.5 m/s versus 2.1 m/s, P = .008). Diaphragmatic SWV at FRC was linked to forced expiratory volume in one second (r = -0.30, P = .009), forced vital capacity (r = -0.33, P = .003), modified Medical Research Council score (r = 0.30, P = .001), and COPD assessment test score (r = 0.48, P < .001). CONCLUSIONS: Ultrasound SWE may be employed as an effective tool for quantitative evaluation of diaphragm stiffness and can help in personalized management of COPD, such as treatment guidance and follow-up monitoring.

Ribosome elongating footprints denoised by wavelet transform comprehensively characterize dynamic cellular translation events.

Translation is dynamically regulated during cell development and stress response. In order to detect actively translated open reading frames (ORFs) and dynamic cellular translation events, we have developed a computational method, RiboWave, to process ribosome profiling data. RiboWave utilizes wavelet transform to denoise the original signal by extracting 3-nt periodicity of ribosomes and precisely locate their footprint denoted as Periodic Footprint P-site (PF P-site). Such high-resolution footprint is found to capture the full track of actively elongating ribosomes, from which translational landscape can be explicitly characterized. We compare RiboWave with several published methods, like RiboTaper, ORFscore and RibORF, and found that RiboWave outperforms them in both accuracy and usage when defining actively translated ORFs. Moreover, we show that PF P-site derived by RiboWave shows superior performance in characterizing the dynamics and complexity of cellular translatome by accurately estimating the abundance of protein levels, assessing differential translation and identifying dynamic translation frameshift.

RNAex: an RNA secondary structure prediction server enhanced by high-throughput structure-probing data.

Several high-throughput technologies have been developed to probe RNA base pairs and loops at the transcriptome level in multiple species. However, to obtain the final RNA secondary structure, extensive effort and considerable expertise is required to statistically process the probing data and combine them with free energy models. Therefore, we developed an RNA secondary structure prediction server that is enhanced by experimental data (RNAex). RNAex is a web interface that enables non-specialists to easily access cutting-edge structure-probing data and predict RNA secondary structures enhanced by in vivo and in vitro data. RNAex annotates the RNA editing, RNA modification and SNP sites on the predicted structures. It provides four structure-folding methods, restrained MaxExpect, SeqFold, RNAstructure (Fold) and RNAfold that can be selected by the user. The performance of these four folding methods has been verified by previous publications on known structures. We re-mapped the raw sequencing data of the probing experiments to the whole genome for each species. RNAex thus enables users to predict secondary structures for both known and novel RNA transcripts in human, mouse, yeast and Arabidopsis The RNAex web server is available at http://RNAex.ncrnalab.org/.

Improved prediction of RNA secondary structure by integrating the free energy model with restraints derived from experimental probing data.

Recently, several experimental techniques have emerged for probing RNA structures based on high-throughput sequencing. However, most secondary structure prediction tools that incorporate probing data are designed and optimized for particular types of experiments. For example, RNAstructure-Fold is optimized for SHAPE data, while SeqFold is optimized for PARS data. Here, we report a new RNA secondary structure prediction method, restrained MaxExpect (RME), which can incorporate multiple types of experimental probing data and is based on a free energy model and an MEA (maximizing expected accuracy) algorithm. We first demonstrated that RME substantially improved secondary structure prediction with perfect restraints (base pair information of known structures). Next, we collected structure-probing data from diverse experiments (e.g. SHAPE, PARS and DMS-seq) and transformed them into a unified set of pairing probabilities with a posterior probabilistic model. By using the probability scores as restraints in RME, we compared its secondary structure prediction performance with two other well-known tools, RNAstructure-Fold (based on a free energy minimization algorithm) and SeqFold (based on a sampling algorithm). For SHAPE data, RME and RNAstructure-Fold performed better than SeqFold, because they markedly altered the energy model with the experimental restraints. For high-throughput data (e.g. PARS and DMS-seq) with lower probing efficiency, the secondary structure prediction performances of the tested tools were comparable, with performance improvements for only a portion of the tested RNAs. However, when the effects of tertiary structure and protein interactions were removed, RME showed the highest prediction accuracy in the DMS-accessible regions by incorporating in vivo DMS-seq data.

A Novel Energy Efficient Topology Control Scheme Based on a Coverage-Preserving and Sleep Scheduling Model for Sensor Networks.

In high-density sensor networks, scheduling some sensor nodes to be in the sleep mode while other sensor nodes remain active for monitoring or forwarding packets is an effective control scheme to conserve energy. In this paper, a Coverage-Preserving Control Scheduling Scheme (CPCSS) based on a cloud model and redundancy degree in sensor networks is proposed. Firstly, the normal cloud model is adopted for calculating the similarity degree between the sensor nodes in terms of their historical data, and then all nodes in each grid of the target area can be classified into several categories. Secondly, the redundancy degree of a node is calculated according to its sensing area being covered by the neighboring sensors. Finally, a centralized approximation algorithm based on the partition of the target area is designed to obtain the approximate minimum set of nodes, which can retain the sufficient coverage of the target region and ensure the connectivity of the network at the same time. The simulation results show that the proposed CPCSS can balance the energy consumption and optimize the coverage performance of the sensor network.

CT features of retroperitoneal solitary fibrous tumor: report of three cases and review of the literature.

CT findings in three cases with solitary fibrous tumors (SFTs) confirmed by histopathology and immunohistochemistry were reviewed retrospectively, and compared with pathological results. The three tumors were large, well-defined, and smooth contour masses and SFT consisted of solid components of two different densities. On enhanced CT scans, tumors were strongly enhancing, the multiple vascular shadows were seen within the tumor in the arterial phase. There is progressive enhancement from the arterial to the venous phase, and the tumor capsule can be observed. Histologically, the tumors are composed of spindle cells within a background of collagen stroma, and showed a wide range of growth patterns, alternating hypercellular (tumor cell-rich) and hypocellular (collagen-rich) areas. The diagnosis is confirmed by characteristic positive immunohistochemical staining for CD34.

3D dynamic landscape simulation of artificial intelligence in environmental landscape design.

Three-dimensional (3D) simulations and precise landscape visualizations are crucial for various applications, like landscape management and planning, computer and connection of the landscape, evaluation, and tracking of land use. The consequences of several plans and a large scene cannot be communicated using older methods of comprehensive environmental planning and development in a timely, rational, and coordinated manner. Architects have trouble incorporating ideas into other comprehensive planning implementation processes. Architects did not thoroughly investigate the neighbourhood's demographics and matching behavioural needs and lacked critical thinking. The 3D dynamic landscape simulation is a detailed computerized three-dimensional simulation of the environment that can be dynamically presented. With the aid of Artificial Intelligence (AI) technology, the system possesses a strong sense of reality, a user-friendly interface, and interactive features that can be tailored to the requirements of the contemporary urban environmental landscape. Regarding exterior publicity, domestic assistance, environmental land use planning, and information systems. The novelty of the proposed Interactive Design System based on AI (IDS-AI) is to create a 3D dynamic landscape model based on a real-life environmental scene, utilizing a Geographic Information System (GIS) to optimize landscape vision. Secondly, 3D environmental landscape design simulation was implemented using GIS spatial analysis in conjunction with the Fuzzy Analytical Hierarchical Process (FAHP) to reduce the data overlap rate and help make an accurate decision. Finally, the design incorporates the development of the interactive interface system application of landscape design and environmental resources for viewing the landscape, the factors that affect them, and the area coverage ratio of various land cover types. The experimental outcomes show that the suggested IDS model increases the gradient sensitivity level of 98.3 % and area coverage ratio of 93.4 % compared to other existing models.

Ultrasound-induced destruction of heparin-loaded microbubbles attenuates L-arginine-induced acute pancreatitis.

PURPOSE: Acute pancreatitis (AP) involves sudden inflammation caused by abnormal activation of pancreatic enzymes. The mechanisms underlying AP include oxidative stress, high levels of inflammatory mediators and inflammatory cell infiltration. Heparin, a key therapeutic drug, exerts anti-inflammatory, antioxidative, and anticoagulative effects. However, safe and effective drug delivery remains an obstacle. This study is the first to investigate the therapeutic effects of heparin-loaded microbubbles (HPMB) combined with ultrasound (UHPMB) and the role of heparin in acoustic cavitation. METHODS: The characteristics of the microbubbles, including particle size, concentration, release, stability, and development, were studied. Heparin concentration in the HPMB was measured, and heparin-induced anticoagulation was evaluated. Drug safety was explored using hemolysis and cell viability assessments. The ability of HPMB to alleviate oxidative stress and inflammation were investigated in vitro. L-arginine induces AP in vivo. UHPMB was used for AP treatment. Serum amylase levels were measured and pancreatic architecture and pathological features were evaluated to determine AP severity. In vivo efficacy was evaluated, and the underlying mechanism of heparin action during acoustic cavitation was explored. RESULTS: HPMB was spherical and presented as an emulsion-like solution without aggregation. HPMB was visible and stable and effectively released the drug under ultrasound (US). HPMB and UHPMB led to lower AP severity than in the untreated group. US-targeted microbubble destruction (UTMD) enhanced the therapeutic effect by decreasing oxidative stress and inflammation in AP models without injuring vital organs. UHPMB regulated VEGF/Flt-1 and SOD-1 expression. HPMB can also mitigate oxidative stress and inflammation in H2O2-pretreated cells. CONCLUSION: UHPMB exhibits a strong ability not only to selectively target pancreatic lesions and release heparin but also to provide efficient protection by inhibiting oxidative stress and inflammation.

Characterization of key aroma compounds in aged Qingxiangxing baijiu by comparative aroma extract dilution analysis, quantitative measurements, aroma recombination, and omission studies.

Aging aroma is an intriguing but an understudied phenomenon in baijiu. The aromatic characteristics of Qingxiangxing (QXX) baijiu stored for 45 years (aged) were investigated using the sensomics approach and were compared with those of young baijiu (0 year-old). Aroma extract dilution analysis revealed 59 odorants, with many long-retained components exhibiting greater flavor dilution values in the aged than the young sample. The analysis of variance of the quantitative data showed significant differences between the young and aged baijiu for 26 compounds. Based on the variations in the content of these compounds, QXX baijiu aged for 0, 5, 15, 20, and 45 years was analyzed and twelve compounds were significantly correlated with aging duration. Finally, eight positively correlated compounds were omitted from the recombinant model, and sotolon, methional, vanillin, dimethyl trisulfide, benzaldehyde, and 3-hydroxy-2-butanone were confirmed to primarily contribute toward the aging aroma of QXX baijiu.

A Novel and Effective Model to Predict Skip Metastasis in Papillary Thyroid Carcinoma Based on a Support Vector Machine.

Introduction: Skip metastasis, referred to as lymph node metastases to the lateral neck compartment without involvement of the central compartment, is generally unpredictable in papillary thyroid carcinoma (PTC). This study aims to establish an effective predictive model for skip metastasis in PTC. Meterials and Methods: Retrospective analysis was performed of clinical samples from 18192 patients diagnosed with thyroid cancer between 2016 to 2020. The First Affiliated Hospital of Wenzhou Medical University. The lateral lymph node metastasis was occureed in the training set (630 PTC patients) and validation set (189 PTC patients). The univariate and multivariate analyses were performed to detect the predictors of skip metastasis and the support vector machine (SVM) was used to establish a model to predict skip metastasis. Results: The rate of skip metastasis was 13.3% (84/631). Tumor size (≤10 mm), upper location, Hashimoto's thyroiditis, extrathyroidal extension, absence of BRAFV600E mutation, and less number of central lymph node dissection were considered as independent predictors of skip metastasis in PTC. For the training set, these predictors performed with 91.7% accuracy, 86.4% sensitivity, 92.2% specificity, 45.2% positive predictive value (PPV), and 98.9% negative predictive value (NPV) in the model. Meanwhile, these predictors showed 91.5% accuracy,71.4% sensitivity, 93.1% specificity, 45.5% PPV, and 97.6% NPV in validation set. Conclusion: This study screened the predictors of the skip lateral lymph node metastasis and to establish an effective and economic predictive model for skip metastasis in PTC. The model can accurately distinguish the skip metastasis in PTC using a simple and affordable method, which may have potential for daily clinical application in the future.

Construction of Enhanced Photostability Anthraquinone-Type Nanovesicles Based on a Novel Two-Step Supramolecular Assembly Strategy and Their Application on Multiband Laser-Responsive Composites.

The photostability and dispersity under aggregation states always become an obstacle for the development of small-molecular organic dye (SMOD) composites. Herein, a novel supramolecular assembly strategy with a two-step assembly method is implemented to encapsulate SMODs for improving their photostability and acquiring uniformly dispersed nanoaggregates in aqueous solution. By the novel assembly strategy, photodegradation rates of the anthraquinone-type dyes can decrease significantly, and the stability of dispersed nanoassembly bodies can be improved in solution. Based on the two-step supramolecular assembly strategy, a new kind of aqueous processing composite system can be developed for preparing multiband laser-responsive devices and in situ healing of optical composite films. This two-step supramolecular assembly strategy can provide a new template and reference for improving the defects of SMODs and fabricating high-performance optical devices.

Extracorporeal Shock Wave Combined with Teriparatide-Loaded Hydrogel Injection Promotes Segmental Bone Defects Healing in Osteoporosis.

BACKGROUND: Osteoporosis is a systemic bone disease characterized by decreased bone density and deterioration of bone microstructure, leading to an increased probability of fragility fractures. Once segmental bone defect occurs, it is easy to cause delayed union and nonunion. METHODS: The aim of this study is to investigate the efficacy of extracorporeal shock wave (ESW) and teriparatide-loaded hydrogel (T-Gel) combined strategy on the cell activity and differentiation of osteoporosis derived bone marrow mesenchymal stem cells (OP-BMSCs) in vitro and bone regeneration in osteoporotic segmental bone defects in vivo. RESULTS: In vitro, the strategy of combining ESW and T-Gel significantly enhanced OP-BMSCs proliferation, survival, migration, and osteogenic differentiation by up-regulating the alkaline phosphatase activity, mineralization, and expression of runt-related transcription factor-2, type I collagen, osteocalcin, and osteopontin. In the segmental bone defect models of osteoporotic rabbits, Micro-CT evaluation and histological observation demonstrated this ESW-combined with T-Gel injection significantly induced bone healing by enhancing the osteogenic activity of the local microenvironment in osteoporotic defects. CONCLUSION: In conclusion, ESW-combined with T-Gel injection could regulate the poor osteogenic microenvironment in osteoporotic defects and show potential for enhancing fragility fractures healing.

Clinical Research and Technique Note of TLIF by Wiltse Approach for the Treatment of Degenerative Lumbar.

OBJECTIVE: To assess the clinical efficacy and share the technique notes of Wiltse Approach TLIF for the treating single segment degenerative lumbar spinal disease. METHOD: In this retrospective controlled study, 780 patients with single segment degenerative lumbar disease who were operated in our hospital from January 2016 to December 2020 were analyzed retrospectively. The patients were randomly assigned to Wiltse approach group (group A, 410 cases) and conventional open approach group (group B, 370 cases). Patient's assessment of pain and disability were evaluated by the visual analogue scale (VAS) and the Oswestry disability index (ODI) before and after surgery. The incision length, operative time, exposure time, intraoperative blood loss, hidden blood loss, time to ambulation, total length of hospitalization, serum creatine kinase, X-rays, CT and MRI were also evaluated. RESULTS: There were no differences in sex, age, pre-operative ODI score, VAS score between the two groups (P > 0.05). The Wiltse approach group had a shorter incision length with 7.69 ± 0.44 cm compared to the conventional group with 11.13 ± 0.36 cm (P < 0.01). The average operative time was 119.20 ± 14.64 min with exposure time of 16.20 ± 3.42 min in the Wiltse approach group and 145.65 ± 16.98 min with 29.20 ± 3.42 min in the conventional group (P < 0.05, P < 0.01). Comparing the intraoperative blood loss, hidden blood loss, serum creatine kinase, time to ambulation, total length of hospitalization, the Wiltse approach group was less than the conventional open approach group (P < 0.05). The VAS score of the two groups decreased significantly with time, and the VAS score of the Wiltse group was significantly lower than that of the conventional open approach group (P < 0.05). At last investigation after operation, ODI scores of the two groups were significantly decreased compared with that before operation. Wiltse approach group was significantly lower than that of the conventional open approach group (P < 0.05). The multifidus of the two groups of patients had a certain degree of atrophy. But the Wiltse approach group multifidus muscle atrophy rate is significantly lower than the conventional open approach group. CONCLUSION: The Wiltse approach TLIF significantly reduces the damage to the paravertebral muscles and the postoperative incidence of chronic low back pain.

Application of deep learning in the detection of breast lesions with four different breast densities.

OBJECTIVE: This retrospective study evaluated the model from populations with different breast densities and showed the model's performance on malignancy prediction. METHODS: A total of 608 mammograms were collected from Northern Jiangsu People's Hospital in Yangzhou City. The data from this province have not been used in the training or evaluation data set. The model consists of three submodules, lesion detection (Mask-rcnn), lesion registration between craniocaudal view and mediolateral oblique view, malignancy prediction network (ResNet). The data set used to train the model was obtained from nine institutions across six cities. For normal cases, there were no annotations. Here, we adopted the free-response receiver operating characteristic (FROC) curve as the indicator to evaluate the detection performance of all cancers and triple-negative breast cancer (TNBC). The FROC curves are also shown for mass/distortion/asymmetry and typical benign calcification in two kinds of populations with four types of breast density. RESULTS: The sensitivity to mass/distortion/asymmetry for the four types of breast (A, B, C, D) are 0.94, 0.92, 0.89, and 0.72, respectively, when false positive per image is 0.25, while these values are 1.00, 0.95, 0.92, and 0.90, respectively, for the amorphous calcification lesions. The sensitivity for the cancer is 0.85 at the same false-positive rate. The TNBC accounts for about 10%-20% of all breast cancers and is more aggressive with poor prognosis than other breast cancers. Herein, we also evaluated performance on the TNBC cases. Our results show that Yizhun AI could detect 75% TNBC lesions at the same false-positive level mentioned above. CONCLUSION: The Yizhun AI model used in our work has good diagnostic efficiency for different types of breast, even for the extremely dense breast. It has a guiding role in the clinical diagnosis of breast cancer. The performance of Yizhun AI on mass/distortion/asymmetry is affected by breast density significantly compared to that on amorphous calcification.

Ultrasound-Induced Destruction of Nitric Oxide-Loaded Microbubbles in the Treatment of Thrombus and Ischemia-Reperfusion Injury.

Objectives: Early recanalization of large vessels in thromboembolism, such as myocardial infarction and ischemic stroke, is associated with improved clinical outcomes. Nitric oxide (NO), a biological gas signaling molecule, has been proven to protect against ischemia-reperfusion injury (IRI). However, the underlying mechanisms remain to be explored. This study investigated whether NO could mitigate IRI and the role of NO during acoustic cavitation. Methods: In vivo, thrombi in the iliac artery of rats were induced by 5% FeCl3. NO-loaded microbubbles (NO-MBs) and ultrasound (US) were used to treat thrombi. B-mode and Doppler US and histological analyses were utilized to evaluate the thrombolysis effect in rats with thrombi. Immunohistochemistry, immunofluorescence, and western blotting were conducted to investigate the underlying mechanisms of NO during acoustic cavitation. In vitro, hypoxia was used to stimulate cells, and NO-MBs were employed to alleviate oxidative stress and apoptosis. Results: We developed NO-MBs that significantly improve the circulation time of NO in vivo, are visible, and effectively release therapeutic gas under US. US-targeted microbubble destruction (UTMD) and NO-loaded UTMD (NO + UTMD) caused a significant decrease in the thrombus area and an increase in the recanalization rates and blood flow velocities compared to the control and US groups. We discovered that UTMD induced NO generation through activation of endothelial NO synthase (eNOS) in vivo. More importantly, we also observed significantly increased NO content and eNOS expression in the NO + UTMD group compared to the UTMD group. NO + UTMD can mitigate oxidative stress and apoptosis in the hind limb muscle without influencing blood pressure or liver and kidney functions. In vitro, NO-MBs alleviated oxidative stress and apoptosis in cells pretreated with hypoxia. Conclusion: Based on these data, UTMD affects the vascular endothelium by activating eNOS, and NO exerts a protective effect against IRI.

A comparison between deep learning convolutional neural networks and radiologists in the differentiation of benign and malignant thyroid nodules on CT images.

INTRODUCTION: We designed 5 convolutional neural network (CNN) models and ensemble models to differentiate malignant and benign thyroid nodules on CT, and compared the diagnostic performance of CNN models with that of radiologists. MATERIAL AND METHODS: We retrospectively included CT images of 880 patients with 986 thyroid nodules confirmed by surgical pathology between July 2017 and December 2019. Two radiologists retrospectively diagnosed benign and malignant thyroid nodules on CT images in a test set. Five CNNs (ResNet50, DenseNet121, DenseNet169, SE-ResNeXt50, and Xception) were trained-validated and tested using 788 and 198 thyroid nodule CT images, respectively. Then, we selected the 3 models with the best diagnostic performance on the test set for the model ensemble. We then compared the diagnostic performance of 2 radiologists with 5 CNN models and the integrated model. RESULTS: Of the 986 thyroid nodules, 541 were malignant, and 445 were benign. The area under the curves (AUCs) for diagnosing thyroid malignancy was 0.587-0.754 for 2 radiologists. The AUCs for diagnosing thyroid malignancy for the 5 CNN models and ensemble model was 0.901-0.947. There were significant differences in AUC between the radiologists' models and the CNN models (p < 0.05). The ensemble model had the highest AUC value. CONCLUSIONS: Five CNN models and an ensemble model performed better than radiologists in distinguishing malignant thyroid nodules from benign nodules on CT. The diagnostic performance of the ensemble model improved and showed good potential.

Utility of Point-of-Care Lung Ultrasound for Clinical Classification of COVID-19.

In this study, the utility of point-of-care lung ultrasound for clinical classification of coronavirus disease (COVID-19) was prospectively assessed. Twenty-seven adult patients with COVID-19 underwent bedside lung ultrasonography (LUS) examinations three times each within the first 2 wk of admission to the isolation ward. We divided the 81 exams into three groups (moderate, severe and critically ill). Lung scores were calculated as the sum of points. A rank sum test and bivariate correlation analysis were carried out to determine the correlation between LUS on admission and clinical classification of COVID-19. There were dramatic differences in LUS (p < 0.001) among the three groups, and LUS scores (r = 0.754) correlated positively with clinical severity (p < 0.01). In addition, moderate, severe and critically ill patients were more likely to have low (≤9), medium (9-15) and high scores (≥15), respectively. This study provides stratification criteria of LUS scores to assist in quantitatively evaluating COVID-19 patients.