Multiple serous cavity effusion screening based on smear images using vision transformer.

Serous cavity effusion is a prevalent pathological condition encountered in clinical settings. Fluid samples obtained from these effusions are vital for diagnostic and therapeutic purposes. Traditionally, cytological examination of smears is a common method for diagnosing serous cavity effusion, renowned for its convenience. However, this technique presents limitations that can compromise its efficiency and diagnostic accuracy. This study aims to overcome these challenges and introduce an improved method for the precise detection of malignant cells in serous cavity effusions. We have developed a transformer-based classification framework, specifically employing the vision transformer (ViT) model, to fulfill this objective. Our research involved collecting smear images and corresponding cytological reports from 161 patients who underwent serous cavity drainage. We meticulously annotated 4836 patches from these images, identifying regions with and without malignant cells, thus creating a unique dataset for smear image classification. The findings of our study reveal that deep learning models, particularly the ViT model, exhibit remarkable accuracy in classifying patches as malignant or non-malignant. The ViT model achieved an impressive area under the receiver operating characteristic curve (AUROC) of 0.99, surpassing the performance of the convolutional neural network (CNN) model, which recorded an AUROC of 0.86. Additionally, we validated our models using an external cohort of 127 patients. The ViT model sustained its high-level screening performance, achieving an AUROC of 0.98 at the patient level, compared to the CNN model's AUROC of 0.84. The visualization of our ViT models confirmed their capability to precisely identify regions containing malignant cells in multiple serous cavity effusion smear images. In summary, our study demonstrates the potential of deep learning models, particularly the ViT model, in automating the screening process for serous cavity effusions. These models offer significant assistance to cytologists in enhancing diagnostic accuracy and efficiency. The ViT model stands out for its advanced self-attention mechanism, making it exceptionally suitable for tasks that necessitate detailed analysis of small, sparsely distributed targets like cellular clusters in serous cavity effusions.

Ankle rehabilitation robot training for stroke patients with foot drop: Optimizing intensity and frequency.

BACKGROUND: Robotic solutions for ankle joint physical therapy have extensively been researched. The optimal frequency and intensity of training for patients when using the ankle robot is not known which can affect rehabilitation outcome. OBJECTIVE: To explore the optimal ankle robot training protocol on foot drop in stroke subjects. METHODS: Subjects were randomly divided into four groups, with 9 in each group. The subjects received different intensities (low or high intensity) with frequencies (1 session/day or 2 sessions/day) of robot combination training. Each session lasted 20 minutes and all subjects were trained 5 days a week for 3 weeks. RESULTS: After 3 weeks of treatment, all groups showed an improvement in passive and active ankle dorsiflexion range of motion (PROM and AROM) and Fugl-Meyer Assessment for lower extremity (FMA-LE) compared to pre-treatment. When training at the same level of intensity, patients who received 2 sessions/day of training had better improvement in ankle dorsiflexion PROM than those who received 1 session/day. In terms of the improvement in dorsiflexion AROM and FMA-LE, patients who received 2 sessions/day with high intensity training improved better than other protocols. CONCLUSION: High frequency and high intensity robot training can be more effective in improving ankle dysfunction.

Research on an ankle rehabilitation robot for hemiplegic patients after stroke.

This paper proposes an ankle rehabilitation robot to assist hemiplegic patients with movement training. The robot consists of two symmetric mechanisms, allowing stroke survivors to execute ankle rehabilitation training based on physiological differences. LPMS-B sensors measure the range of movement (ROM) of the human ankle joint, and the results are used for control parameters of the robot. Control strategies for constant speed training mode, constant torque training mode, and combination training mode are put forth based on the hardware system of the robot. Experiments verify the feasibility of the robot for ankle rehabilitation training. Results show a maximum mean error of 0.3364° between the trajectory of the intact side and the affected side, a maximum mean error of 0.0335°/s between target speed and experimental speed, and a maximum mean error of 0.0775 N m between target torque and experimental torque. The ankle joint rehabilitation robot proposed in this paper can help patients complete the training well under the three control modes.

MG-Trans: Multi-Scale Graph Transformer With Information Bottleneck for Whole Slide Image Classification.

Multiple instance learning (MIL)-based methods have become the mainstream for processing the megapixel-sized whole slide image (WSI) with pyramid structure in the field of digital pathology. The current MIL-based methods usually crop a large number of patches from WSI at the highest magnification, resulting in a lot of redundancy in the input and feature space. Moreover, the spatial relations between patches can not be sufficiently modeled, which may weaken the model's discriminative ability on fine-grained features. To solve the above limitations, we propose a Multi-scale Graph Transformer (MG-Trans) with information bottleneck for whole slide image classification. MG-Trans is composed of three modules: patch anchoring module (PAM), dynamic structure information learning module (SILM), and multi-scale information bottleneck module (MIBM). Specifically, PAM utilizes the class attention map generated from the multi-head self-attention of vision Transformer to identify and sample the informative patches. SILM explicitly introduces the local tissue structure information into the Transformer block to sufficiently model the spatial relations between patches. MIBM effectively fuses the multi-scale patch features by utilizing the principle of information bottleneck to generate a robust and compact bag-level representation. Besides, we also propose a semantic consistency loss to stabilize the training of the whole model. Extensive studies on three subtyping datasets and seven gene mutation detection datasets demonstrate the superiority of MG-Trans.

Spatiotemporal evolution and trend prediction of regional water-energy-food-ecology system vulnerability: a case study of the Yangtze River Economic Belt, China.

The vulnerability study of the water-energy-food-ecosystem (WEFE) system in the Yangtze River Economic Belt (YREB) is of great significance for ensuring water, energy, food, and ecological security. In this study, the concept of vulnerability of WEFE systems is explained, and then its mechanism is analyzed. Based on the vulnerability concept of the WEFE system and combined with the pressure-state-response (PSR) framework and the exposure-sensitivity-adaptability (ESA) capacity model, this study constructs an evaluation index system for the vulnerability of the WEFE system. The cross-efficiency data envelopment analysis (CE-DEA) method, which considers both self-evaluation and peer evaluation efficiency, is used to calculate the vulnerability from 2005 to 2020. The spatiotemporal dynamic characteristics are explored using the hot spot analysis and spatial autocorrelation model. The results show that the overall vulnerability of the WEFE system in the YREB has fluctuated and increased during the study period, with a spatial pattern characterized by "high in the middle and low on both sides." Over time, the spatial evolution tends to be centralized and non-equilibrium, forming a relatively independent spatial pattern.

A survey of artificial intelligence in tongue image for disease diagnosis and syndrome differentiation.

The rapid development of artificial intelligence technology has gradually extended from the general field to all walks of life, and intelligent tongue diagnosis is the product of a miraculous connection between this new discipline and traditional disciplines. We reviewed the deep learning methods and machine learning applied in tongue image analysis that have been studied in the last 5 years, focusing on tongue image calibration, detection, segmentation, and classification of diseases, syndromes, and symptoms/signs. Introducing technical evolutions or emerging technologies were applied in tongue image analysis; as we have noticed, attention mechanism, multiscale features, and prior knowledge were successfully applied in it, and we emphasized the value of combining deep learning with traditional methods. We also pointed out two major problems concerned with data set construction and the low reliability of performance evaluation that exist in this field based on the basic essence of tongue diagnosis in traditional Chinese medicine. Finally, a perspective on the future of intelligent tongue diagnosis was presented; we believe that the self-supervised method, multimodal information fusion, and the study of tongue pathology will have great research significance.

Validation of the Longshi scale in assessment of function in patients from different economic, educational and regional backgrounds in China.

OBJECTIVE: This study validates performance of the Longshi scale against modified Barthel index in assessment of function in patients from different economic, educational and regional backgrounds in China. DESIGN: This is a cross-sectional study. SETTING: One hundred and three hospitals and rehabilitation institutions across China. PARTICIPANTS: A total of 14,752 patients with physical and cognitive impairments were enrolled and classified to five educational levels and five family income levels; 8060 participants were further selected from five regions to study regional influences. MAIN MEASURES: Activities of daily living were assessed with two evaluation tools, the Longshi scale and modified Barthel index. Results of evaluation with Longshi scale performed by non-healthcare workers were validated against modified Barthel index performed by healthcare workers using Pearson's correlation test. RESULTS: There were significant positive correlations between the findings of the Longshi scale administered by non-healthcare professionals and the modified Barthel index administered by healthcare professionals. This was true for level of education (correlations ranged from 0.697 to 0.822), family income level (correlations ranged from 0.724 to 0.761) and region (correlations ranged from 0.737 to 0.776). CONCLUSION: In a large dataset of 14,752 patients, positive correlations were found between Longshi scale and modified Barthel index measures of function. Positive correlations persisted across subgroup analysis of groups from different social, economic and regional backgrounds, and with administration by non-healthcare professionals. CLINICAL TRIAL REGISTRATION: ChiCTR2000034067, www.chictr.org.cn.

A Structure-Aware Hierarchical Graph-Based Multiple Instance Learning Framework for pT Staging in Histopathological Image.

Pathological primary tumor (pT) stage focuses on the infiltration degree of the primary tumor to surrounding tissues, which relates to the prognosis and treatment choices. The pT staging relies on the field-of-views from multiple magnifications in the gigapixel images, which makes pixel-level annotation difficult. Therefore, this task is usually formulated as a weakly supervised whole slide image (WSI) classification task with the slide-level label. Existing weakly-supervised classification methods mainly follow the multiple instance learning paradigm, which takes the patches from single magnification as the instances and extracts their morphological features independently. However, they cannot progressively represent the contextual information from multiple magnifications, which is critical for pT staging. Therefore, we propose a structure-aware hierarchical graph-based multi-instance learning framework (SGMF) inspired by the diagnostic process of pathologists. Specifically, a novel graph-based instance organization method is proposed, namely structure-aware hierarchical graph (SAHG), to represent the WSI. Based on that, we design a novel hierarchical attention-based graph representation (HAGR) network to capture the critical patterns for pT staging by learning cross-scale spatial features. Finally, the top nodes of SAHG are aggregated by a global attention layer for bag-level representation. Extensive studies on three large-scale multi-center pT staging datasets with two different cancer types demonstrate the effectiveness of SGMF, which outperforms state-of-the-art up to 5.6% in the F1 score.

Hydrogel armed with Bmp2 mRNA-enriched exosomes enhances bone regeneration.

BACKGROUND: Sustained release of bioactive BMP2 (bone morphogenetic protein-2) is important for bone regeneration, while the intrinsic short half-life of BMP2 at protein level cannot meet the clinical need. In this study, we aimed to design Bmp2 mRNA-enriched engineered exosomes, which were then loaded into specific hydrogel to achieve sustained release for more efficient and safe bone regeneration. RESULTS: Bmp2 mRNA was enriched into exosomes by selective inhibition of translation in donor cells, in which NoBody (non-annotated P-body dissociating polypeptide, a protein that inhibits mRNA translation) and modified engineered BMP2 plasmids were co-transfected. The derived exosomes were named ExoBMP2+NoBody. In vitro experiments confirmed that ExoBMP2+NoBody had higher abundance of Bmp2 mRNA and thus stronger osteogenic induction capacity. When loaded into GelMA hydrogel via ally-L-glycine modified CP05 linker, the exosomes could be slowly released and thus ensure prolonged effect of BMP2 when endocytosed by the recipient cells. In the in vivo calvarial defect model, ExoBMP2+NoBody-loaded GelMA displayed great capacity in promoting bone regeneration. CONCLUSIONS: Together, the proposed ExoBMP2+NoBody-loaded GelMA can provide an efficient and innovative strategy for bone regeneration.

Semi-Supervised Pixel Contrastive Learning Framework for Tissue Segmentation in Histopathological Image.

Accurate tissue segmentation in histopathological images is essential for promoting the development of precision pathology. However, the size of the digital pathological image is great, which needs to be tiled into small patches containing limited semantic information. To imitate the pathologist's diagnosis process and model the semantic relation of the whole slide image, We propose a semi-supervised pixel contrastive learning framework (SSPCL) which mainly includes an uncertainty-guided mutual dual consistency learning module (UMDC) and a cross image pixel-contrastive learning module (CIPC). The UMDC module enables efficient learning from unlabeled data through mutual dual-consistency and consensus-based uncertainty. The CIPC module aims at capturing the cross-patch semantic relationship by optimizing a contrastive loss between pixel embeddings. We also propose several novel domain-related sampling methods by utilizing the continuous spatial structure of adjacent image patches, which can avoid the problem of false sampling and improve the training efficiency. In this way, SSPCL significantly reduces the labeling cost on histopathological images and realizes the accurate quantitation of tissues. Extensive experiments on three tissue segmentation datasets demonstrate the effectiveness of SSPCL, which outperforms state-of-the-art up to 5.0% in mDice.

A semi-supervised multi-task learning framework for cancer classification with weak annotation in whole-slide images.

Cancer region detection (CRD) and subtyping are two fundamental tasks in digital pathology image analysis. The development of data-driven models for CRD and subtyping on whole-slide images (WSIs) would mitigate the burden of pathologists and improve their accuracy in diagnosis. However, the existing models are facing two major limitations. Firstly, they typically require large-scale datasets with precise annotations, which contradicts with the original intention of reducing labor effort. Secondly, for the subtyping task, the non-cancerous regions are treated as the same as cancerous regions within a WSI, which confuses a subtyping model in its training process. To tackle the latter limitation, the previous research proposed to perform CRD first for ruling out the non-cancerous region, then train a subtyping model based on the remaining cancerous patches. However, separately training ignores the interaction of these two tasks, also leads to propagating the error of the CRD task to the subtyping task. To address these issues and concurrently improve the performance on both CRD and subtyping tasks, we propose a semi-supervised multi-task learning (MTL) framework for cancer classification. Our framework consists of a backbone feature extractor, two task-specific classifiers, and a weight control mechanism. The backbone feature extractor is shared by two task-specific classifiers, such that the interaction of CRD and subtyping tasks can be captured. The weight control mechanism preserves the sequential relationship of these two tasks and guarantees the error back-propagation from the subtyping task to the CRD task under the MTL framework. We train the overall framework in a semi-supervised setting, where datasets only involve small quantities of annotations produced by our minimal point-based (min-point) annotation strategy. Extensive experiments on four large datasets with different cancer types demonstrate the effectiveness of the proposed framework in both accuracy and generalization.

Quantitative evaluation of China's basin ecological compensation policies based on the PMC index model.

Policy evaluation is the premise of the scientific formulation and effective implementation of a basin ecological compensation policy. However, whether the formulation of the basin ecological compensation policy (BECP) is reasonable or not lacks theoretical and technical support. This study constructed a model based on the PMC and text mining methods. PMC index model enables decision-makers to determine the level of consistency and the strengths and weaknesses of any policy from multiple angles and makes the evaluation results more targeted and operable. By establishing an evaluation system for BECP and building a multi-input-output table, the score of each policy is calculated. Based on this, the rationality of nine ecological compensation policies in the Yangtze and Yellow River basins was then examined. The results show that the average value of the PMC index for the nine policies is 7.23, which indicate the formulation of the basin ecological compensation policy in China is generally reasonable. Ranking of policy scores from high to low is P2 > P1 > P5 > P7 > P3 > P4 > P6 > P9 > P8. However, deficiencies exist in policy timeliness, incentive measures, and policy receptors. In addition, there is a large gap in the formulation of policies at different levels. Moreover, the level of local policies is uneven.

HistoML, a markup language for representation and exchange of histopathological features in pathology images.

The study of histopathological phenotypes is vital for cancer research and medicine as it links molecular mechanisms to disease prognosis. It typically involves integration of heterogenous histopathological features in whole-slide images (WSI) to objectively characterize a histopathological phenotype. However, the large-scale implementation of phenotype characterization has been hindered by the fragmentation of histopathological features, resulting from the lack of a standardized format and a controlled vocabulary for structured and unambiguous representation of semantics in WSIs. To fill this gap, we propose the Histopathology Markup Language (HistoML), a representation language along with a controlled vocabulary (Histopathology Ontology) based on Semantic Web technologies. Multiscale features within a WSI, from single-cell features to mesoscopic features, could be represented using HistoML which is a crucial step towards the goal of making WSIs findable, accessible, interoperable and reusable (FAIR). We pilot HistoML in representing WSIs of kidney cancer as well as thyroid carcinoma and exemplify the uses of HistoML representations in semantic queries to demonstrate the potential of HistoML-powered applications for phenotype characterization.

Unsupervised Representation Learning for Tissue Segmentation in Histopathological Images: From Global to Local Contrast.

Tissue segmentation is an essential task in computational pathology. However, relevant datasets for such a pixel-level classification task are hard to obtain due to the difficulty of annotation, bringing obstacles for training a deep learning-based segmentation model. Recently, contrastive learning has provided a feasible solution for mitigating the heavy reliance of deep learning models on annotation. Nevertheless, applying contrastive loss to the most abstract image representations, existing contrastive learning frameworks focus on global features, therefore, are less capable of encoding finer-grained features (e.g., pixel-level discrimination) for the tissue segmentation task. Enlightened by domain knowledge, we design three contrastive learning tasks with multi-granularity views (from global to local) for encoding necessary features into representations without accessing annotations. Specifically, we construct: (1) an image-level task to capture the difference between tissue components, i.e., encoding the component discrimination; (2) a superpixel-level task to learn discriminative representations of local regions with different tissue components, i.e., encoding the prototype discrimination; (3) a pixel-level task to encourage similar representations of different tissue components within a local region, i.e., encoding the spatial smoothness. Through our global-to-local pre-training strategy, the learned representations can reasonably capture the domain-specific and fine-grained patterns, making them easily transferable to various tissue segmentation tasks in histopathological images. We conduct extensive experiments on two tissue segmentation datasets, while considering two real-world scenarios with limited or sparse annotations. The experimental results demonstrate that our framework is superior to existing contrastive learning methods and can be easily combined with weakly supervised and semi-supervised segmentation methods.

A Novel Multi-Omics Analysis Model for Diagnosis and Survival Prediction of Lower-Grade Glioma Patients.

Background: Lower-grade gliomas (LGGs) are characterized by remarkable genetic heterogeneity and different clinical outcomes. Classification of LGGs is improved by the development of molecular stratification markers including IDH mutation and 1p/19q chromosomal integrity, which are used as a hallmark of survival and therapy sensitivity of LGG patients. However, the reproducibility and sensitivity of the current classification remain ambiguous. This study aimed to construct more accurate risk-stratification approaches. Methods: According to bioinformatics, the sequencing profiles of methylation and transcription and imaging data derived from LGG patients were analyzed and developed predictable risk score and radiomics score. Moreover, the performance of predictable models was further validated. Results: In this study, we determined a cluster of 6 genes that were correlated with IDH mutation/1p19q co-deletion status. Risk score model was calculated based on 6 genes and showed gratifying sensitivity and specificity for survival prediction and therapy response of LGG patients. Furthermore, a radiomics risk score model was established to noninvasively assist judgment of risk score in pre-surgery. Taken together, a predictable nomogram that combined transcriptional signatures and clinical characteristics was established and validated to be preferable to the histopathological classification. Our novel multi-omics nomograms showed a satisfying performance. To establish a user-friendly application, the nomogram was further developed into a web-based platform: https://drw576223193.shinyapps.io/Nomo/, which could be used as a supporting method in addition to the current histopathological-based classification of gliomas. Conclusions: Our novel multi-omics nomograms showed the satisfying performance of LGG patients and assisted clinicians to draw up individualized clinical management.

A Novel Encoding and Decoding Calibration Guiding Pathway for Pathological Image Analysis.

Diagnostic pathology is the foundation and gold standard for identifying carcinomas, and the accurate quantification of pathological images can provide objective clues for pathologists to make more convincing diagnosis. Recently, the encoder-decoder architectures (EDAs) of convolutional neural networks (CNNs) are widely used in the analysis of pathological images. Despite the rapid innovation of EDAs, we have conducted extensive experiments based on a variety of commonly used EDAs, and found them cannot handle the interference of complex background in pathological images, making the architectures unable to focus on the regions of interest (RoIs), thus making the quantitative results unreliable. Therefore, we proposed a pathway named GLobal Bank (GLB) to guide the encoder and the decoder to extract more features of RoIs rather than the complex background. Sufficient experiments have proved that the architecture remoulded by GLB can achieve significant performance improvement, and the quantitative results are more accurate.

Endoscopic submucosal dissection treatment of metachronous early gastric adenocarcinoma in a case with diffuse large B-cell lymphoma.

Cases of primary gastric lymphoma complicated with gastric adenocarcinoma are rare in clinical practice. We report a case of metachronous early gastric adenocarcinoma diagnosed eight years after the onset of gastric diffuse large B-cell lymphoma (DLBCL) and treated with endoscopic submucosal dissection (ESD).

Inhibitory effect of pirfenidone on pulmonary fibrosis in patients with acute paraquat poisoning.

OBJECTIVE: To study the efficacy of pirfenidone (PFD) on patients with pulmonary fibrosis caused by acute paraquat (PQ) poisoning. METHODS: A total of 86 patients with pulmonary fibrosis caused by acute PQ poisoning admitted to our hospital were analyzed retrospectively. All of them successfully received the standard 21-day treatment based on "Taishan Consensus", and they were assigned to the PFD group or the NO-PFD group according to whether they received PFD treatment (at 200 mg/time, 3 times/day) for 6 months after discharge. The two groups were compared in effective treatment rate, mortality and incidence of adverse reactions such as liver and kidney function damage, pulmonary fibrosis-associated indexes, pulmonary function-associated indexes, and arterial blood gas indexes before and after therapy. RESULTS: The PFD group showed a notably higher effective treatment rate than the NO-PFD group (P<0.05). Additionally, the PFD group showed notably lower levels of serum hyaluronic acid (HA), laminin (LN), type IV collagen (CIV), and type III procollagen (PCIII), and notably higher levels of forced expiratory volume in 1 second (FEV1), forced vital capacity (FVC), and FEV1/FVC than the NO-PFD group (all P<0.001), and the PFD group also showed significantly higher levels of arterial blood gas indexes including arterial partial pressure of oxygen (PaO2) and PaO2/inspired oxygen (FIO2) than the NO-PFD group (both P<0.001). Moreover, the Kaplan-Meier survival curves showed that the survival rate of the patients in PFD group was significantly higher than that in the NO-PFD group (P<0.05). CONCLUSION: With a high safety, PFD can effectively improve the treatment efficacy in patients with pulmonary fibrosis caused by acute PQ poisoning. PFD can improve the pulmonary function and arterial blood gas status of patients, without causing obvious liver and kidney damage.

An Adaptive Method for Gait Event Detection of Gait Rehabilitation Robots.

Accurate gait event detection is necessary for control strategies of gait rehabilitation robots. However, due to personal diversity between individuals, it is a challenge for robots to detect a gait event at various stride frequencies. This paper proposes a novel method for gait event detection of a gait rehabilitation robot using a single inertial sensor mounted on the thigh. A self-adaptive threshold for detecting heel strike is obtained in real time via a linear regression model. Observable thresholds for toe off detection are constant at various stride frequencies. Experiments are conducted based on 20 healthy subjects and six hemiplegic patients wearing a gait rehabilitation robot and walking at various kinds of stride frequencies. The experimental results show that the proposed method can detect heel strike and toe off gait events within an average 2% gait cycle temporal errors and never miss two-gait event detection. Compared to the conventional thresholding method, this work presents a simple and robust application for gait event detection in healthy and hemiplegic subjects by one inertial sensor. The linear regression model can be applicable to different subjects walking at various stride frequencies.