Explainable survival analysis with uncertainty using convolution-involved vision transformer.

Image-based precision medicine research is able to help doctors make better decisions on treatments. Among all kinds of medical images, a special form is called Whole Slide Image (WSI), which is used for diagnosing patients with cancer, aiming to enable more accurate survival prediction with its high resolution. However, One unique challenge of the WSI-based prediction models is processing the gigabyte-size or even terabyte-size WSIs, which would make most models computationally infeasible. Although existing models mostly use a pre-selected subset of key patches or patch clusters as input, they might discard some important morphology information, making the prediction inferior. Another challenge is improving the prediction models' explainability, which is crucial to help doctors understand the predictions given by the models and make faithful decisions with high confidence. To address the above two challenges, in this work, we propose a novel explainable survival prediction model based on Vision Transformer. Specifically, we adopt dual-channel convolutional layers to utilize the complete WSIs for more accurate predictions. We also introduce the aleatoric uncertainty into our model to understand its limitation and avoid overconfidence in using the prediction results. Additionally, we present a post-hoc explainable method to identify the most salient patches and distinct morphology features as supporting evidence for predictions. Evaluations of two large cancer datasets show that our proposed model is able to make survival predictions more effectively and has better explainability for cancer diagnosis.

Highly thermal conductive shape-stabilized composite phase change materials based on boron nitride and expanded graphite for solar thermal applications.

Phase change heat storage technology is a good way to solve the problem that the temperature of solar hot water outlet is affected by the time domain. A stearic acid (SA)-benzamide (BA) eutectic mixture is a potential phase change material (PCM), but it still has the disadvantages of low thermal conductivity and liquid leakage. In this work, a new high thermal conductive shape-stabilized composite PCM was prepared by adding boron nitride (BN) and expanded graphite (EG) to a melted SA-BA eutectic mixture using an ultrasonic and melt adsorption method, and its phase change temperature, latent heat, crystal structure, morphology, thermal conductivity, chemical stability, thermal stability, cycle stability and leakage characteristics were characterized. The results indicates that the addition of BN and EG significantly improved the thermal conductivity of the SA-BA eutectic mixture, and they efficiently adsorbed the melted SA-BA eutectic mixture. Besides, when the mass fractions of BN and EG are 15 wt% and 20 wt%, respectively, the 15BN20EG composite has almost no liquid phase leakage. When the melting enthalpy and temperature of 15BN20EG are 132.35 J g-1 and 65.21 °C, respectively, the thermal conductivity of the 15BN20EG is 6.990 W m-1 K-1, which is 20.601 times that of the SA-BA eutectic mixture. Moreover, 15BN20EG shows good thermal stability after 100 cycles and good chemical stability below 100 °C. Therefore, the 15BN20EG composite is considered as a potential candidate for solar thermal applications.

A convolutional neural network with pixel-wise sparse graph reasoning for COVID-19 lesion segmentation in CT images.

The COVID-19 pandemic has extremely threatened human health, and automated algorithms are needed to segment infected regions in the lung using computed tomography (CT). Although several deep convolutional neural networks (DCNNs) have proposed for this purpose, their performance on this task is suppressed due to the limited local receptive field and deficient global reasoning ability. To address these issues, we propose a segmentation network with a novel pixel-wise sparse graph reasoning (PSGR) module for the segmentation of COVID-19 infected regions in CT images. The PSGR module, which is inserted between the encoder and decoder of the network, can improve the modeling of global contextual information. In the PSGR module, a graph is first constructed by projecting each pixel on a node based on the features produced by the encoder. Then, we convert the graph into a sparsely-connected one by keeping K strongest connections to each uncertainly segmented pixel. Finally, the global reasoning is performed on the sparsely-connected graph. Our segmentation network was evaluated on three publicly available datasets and compared with a variety of widely-used segmentation models. Our results demonstrate that (1) the proposed PSGR module can capture the long-range dependencies effectively and (2) the segmentation model equipped with this PSGR module can accurately segment COVID-19 infected regions in CT images and outperform all other competing models.

Contrastive Brain Network Learning via Hierarchical Signed Graph Pooling Model.

Recently, brain networks have been widely adopted to study brain dynamics, brain development, and brain diseases. Graph representation learning techniques on brain functional networks can facilitate the discovery of novel biomarkers for clinical phenotypes and neurodegenerative diseases. However, current graph learning techniques have several issues on brain network mining. First, most current graph learning models are designed for unsigned graph, which hinders the analysis of many signed network data (e.g., brain functional networks). Meanwhile, the insufficiency of brain network data limits the model performance on clinical phenotypes' predictions. Moreover, few of the current graph learning models are interpretable, which may not be capable of providing biological insights for model outcomes. Here, we propose an interpretable hierarchical signed graph representation learning (HSGPL) model to extract graph-level representations from brain functional networks, which can be used for different prediction tasks. To further improve the model performance, we also propose a new strategy to augment functional brain network data for contrastive learning. We evaluate this framework on different classification and regression tasks using data from human connectome project (HCP) and open access series of imaging studies (OASIS). Our results from extensive experiments demonstrate the superiority of the proposed model compared with several state-of-the-art techniques. In addition, we use graph saliency maps, derived from these prediction tasks, to demonstrate detection and interpretation of phenotypic biomarkers.

TNFAIP6 Promotes Gastric Carcinoma Cell Invasion via Upregulating PTX3 and Activating the Wnt/ß-Catenin Signaling Pathway.

Tumor metastasis is a fundamental cause of the poor prognosis of gastric carcinoma (GC). In order to study the problems affecting metastasis and recurrence of gastric cancer, the paper expose that TNF alpha induced protein 6 (TNFAIP6) is aberrantly overexpressed in GC, and patients with high-TNFAIP6 levels exhibited inferior overall survival. Mechanistically, overexpression of TNFAIP6 raised ß-catenin ectopic nuclear distribution and activated the Wnt/ß-catenin signal pathway. The experimental results show that TNFAIP6 facilitates the aggressive potential of GC cells through modulating PTX3 expression.

A gas-pressure-assisted ratiometric atomic flame assay for the point-of-care testing of tumor-cell-derived exosomes.

The multicolor-based point-of-care testing (POCT) of tumor cell-derived exosomes is of vital importance for understanding tumor growth and metastasis. Multicolor-based ratiometric signals most often rely on molecular optics, such as fluorescence resonance energy transfer (FRET)-dependent molecular fluorescence and localized surface plasmon resonance (LSPR)-related molecular colorimetry. However, finding acceptable FRET donor-acceptor fluorophore pairs and the kinetically slow color responses during size-related molecular colorimetry have greatly impeded POCT applications. Herein, an atomic flame was used to develop a visual sensing platform for the POCT of tumor-cell-derived exosomes. In comparison with common molecular optics, the atomic flame possessed the advantages of providing both a variety of ratiometric flame signals and fast response sensitivity. The integration of a gas-pressure-assisted flame reaction and dual-aptamer recognition guaranteed the sensitive and selective analysis of exosomes with a low limit of detection (LOD) of 7.6 × 102 particles per mL. Such a novel optical signal will inspire the development of more user-friendly POCT approaches.

CommPOOL: An interpretable graph pooling framework for hierarchical graph representation learning.

Recent years have witnessed the emergence and flourishing of hierarchical graph pooling neural networks (HGPNNs) which are effective graph representation learning approaches for graph level tasks such as graph classification. However, current HGPNNs do not take full advantage of the graph's intrinsic structures (e.g., community structure). Moreover, the pooling operations in existing HGPNNs are difficult to be interpreted. In this paper, we propose a new interpretable graph pooling framework - CommPOOL, that can capture and preserve the hierarchical community structure of graphs in the graph representation learning process. Specifically, the proposed community pooling mechanism in CommPOOL utilizes an unsupervised approach for capturing the inherent community structure of graphs in an interpretable manner. CommPOOL is a general and flexible framework for hierarchical graph representation learning that can further facilitate various graph-level tasks. Evaluations on five public benchmark datasets and one synthetic dataset demonstrate the superior performance of CommPOOL in graph representation learning for graph classification compared to the state-of-the-art baseline methods, and its effectiveness in capturing and preserving the community structure of graphs.

Peritoneal dialysis catheter insertion using a very-low-site approach: a 5-year experience.

PURPOSE: Peritoneal dialysis (PD) catheter tip migration accounts for the majority of cases of PD catheter malfunction. In this case series, we described our experiences of using a modified PD catheter implantation approach through a site that is lower than the site that is conventionally used, to reduce catheter malfunction. METHODS: We retrospectively identified 76 patients who received PD catheter implantation at the Affiliated Wujin Hospital of Jiangsu University, among whom 39 received the traditional approach of low-site insertion and 37 received a modified approach of very-low-site insertion. All participants were followed up for at least 2 years after PD catheter implantation, and the development of catheter dysfunction or death during this period was monitored. RESULTS: We found that the survival rate of the initially inserted catheter was 75.68% among the very-low-site group. This survival rate was significantly better than that observed among the low-site group (48.72%; p = 0.029). Kaplan-Meier curves of the initial catheter survival also showed that the catheter survival was significantly higher in the patients in the very-low-site group than those in the low-site group (log rank p = 0.012). Complications, such as catheter tip migration, were not observed in the very-low-site group, while tip migration occurred in 15.38% of the patients in the low-site group (very-low-site group vs low-site group: p = 0.039). CONCLUSIONS: A safe and simple PD catheter implantation can be performed either through the low-site approach or the very-low-site approach.

[The study of hBMP-4 gene modified tissue-engineered bone].

OBJECTIVE: Bone marrow stromal cells (bMSCs) of rabbits transferred with mammalian hBMP-4 expression plasmid were used to construct tissue-engineered bone. Gene therapy combined with tissue-engineering technique was explored to further improve osteogenesis. METHODS: pEGFP-hBMP-4 plasmid was constructed by subcloning technique. bMSCs were then transferred with either pEGFP-hBMP-4, pEGFP plasmid by lipofectamine or left uninfected in vitro. The cells from the 3 groups were combined with natural non-organic bone (NNB) to construct tissue-engineered bones, which were subcutaneously implanted into nude mice (6 implants per group) for 4 weeks. Specimens were evaluated through histological and computerized new bone formation analysis. RESULTS: pEGFP-hBMP-4 plasmid was successfully constructed. bMSCs could attach and proliferate on the surface on NNB. In vivo experiment showed that new bone formation in pEGFP-hBMP-4 group was higher than those of the control groups. CONCLUSIONS: Tissue-engineered bone using hBMP-4 gene modified bMSCs might be an ideal alternative for the repair of bone.