OBJECTIVE: Pneumothorax is an acute thoracic disease caused by abnormal air collection between the lungs and chest wall. Recently, artificial intelligence (AI), especially deep learning (DL), has been increasingly employed for automating the diagnostic process of pneumothorax. To address the opaqueness often associated with DL models, explainable artificial intelligence (XAI) methods have been introduced to outline regions related to pneumothorax. However, these explanations sometimes diverge from actual lesion areas, highlighting the need for further improvement.
METHODS: We propose a template-guided approach to incorporate the clinical knowledge of pneumothorax into model explanations generated by XAI methods, thereby enhancing the quality of the explanations. Utilizing one lesion delineation created by radiologists, our approach first generates a template that represents potential areas of pneumothorax occurrence. This template is then superimposed on model explanations to filter out extraneous explanations that fall outside the template\'s boundaries. To validate its efficacy, we carried out a comparative analysis of three XAI methods (Saliency Map, Grad-CAM, and Integrated Gradients) with and without our template guidance when explaining two DL models (VGG-19 and ResNet-50) in two real-world datasets (SIIM-ACR and ChestX-Det).
RESULTS: The proposed approach consistently improved baseline XAI methods across twelve benchmark scenarios built on three XAI methods, two DL models, and two datasets. The average incremental percentages, calculated by the performance improvements over the baseline performance, were 97.8% in Intersection over Union (IoU) and 94.1% in Dice Similarity Coefficient (DSC) when comparing model explanations and ground-truth lesion areas. We further visualized baseline and template-guided model explanations on radiographs to showcase the performance of our approach.
CONCLUSIONS: In the context of pneumothorax diagnoses, we proposed a template-guided approach for improving model explanations. Our approach not only aligns model explanations more closely with clinical insights but also exhibits extensibility to other thoracic diseases. We anticipate that our template guidance will forge a novel approach to elucidating AI models by integrating clinical domain expertise.

BACKGROUND: Current approaches for initial coronary artery disease (CAD) assessment rely on pretest probability (PTP) based on risk factors and presentations, with limited performance. Infrared thermography (IRT), a non-contact technology that detects surface temperature, has shown potential in assessing atherosclerosis-related conditions, particularly when measured from body regions such as faces. We aim to assess the feasibility of using facial IRT temperature information with machine learning for the prediction of CAD.
METHODS: Individuals referred for invasive coronary angiography or coronary CT angiography (CCTA) were enrolled. Facial IRT images captured before confirmatory CAD examinations were used to develop and validate a deep-learning IRT image model for detecting CAD. We compared the performance of the IRT image model with the guideline-recommended PTP model on the area under the curve (AUC). In addition, interpretable IRT tabular features were extracted from IRT images to further validate the predictive value of IRT information.
RESULTS: A total of 460 eligible participants (mean (SD) age, 58.4 (10.4) years; 126 (27.4%) female) were included. The IRT image model demonstrated outstanding performance (AUC 0.804, 95% CI 0.785 to 0.823) compared with the PTP models (AUC 0.713, 95% CI 0.691 to 0.734). A consistent level of superior performance (AUC 0.796, 95% CI 0.782 to 0.811), achieved with comprehensive interpretable IRT features, further validated the predictive value of IRT information. Notably, even with only traditional temperature features, a satisfactory performance (AUC 0.786, 95% CI 0.769 to 0.803) was still upheld.
CONCLUSIONS: In this prospective study, we demonstrated the feasibility of using non-contact facial IRT information for CAD prediction.

Accurate detection of pathogens, particularly distinguishing between Gram-positive and Gram-negative bacteria, could improve disease treatment. Host gene expression can capture the immune system\'s response to infections caused by various pathogens. Here, we present a deep neural network model, bvnGPS2, which incorporates the attention mechanism based on a large-scale integrated host transcriptome dataset to precisely identify Gram-positive and Gram-negative bacterial infections as well as viral infections. We performed analysis of 4,949 blood samples across 40 cohorts from 10 countries using our previously designed omics data integration method, iPAGE, to select discriminant gene pairs and train the bvnGPS2. The performance of the model was evaluated on six independent cohorts comprising 374 samples. Overall, our deep neural network model shows robust capability to accurately identify specific infections, paving the way for precise medicine strategies in infection treatment and potentially also for identifying subtypes of other diseases.

Data categorization is a top concern in medical data to predict and detect illnesses; thus, it is applied in modern healthcare informatics. In modern informatics, machine learning and deep learning models have enjoyed great attention for categorizing medical data and improving illness detection. However, the existing techniques, such as features with high dimensionality, computational complexity, and long-term execution duration, raise fundamental problems. This study presents a novel classification model employing metaheuristic methods to maximize efficient positives on Chronic Kidney Disease diagnosis. The medical data is initially massively pre-processed, where the data is purified with various mechanisms, including missing values resolution, data transformation, and the employment of normalization procedures. The focus of such processes is to leverage the handling of the missing values and prepare the data for deep analysis. We adopt the Binary Grey Wolf Optimization method, a reliable subset selection feature using metaheuristics. This operation is aimed at improving illness prediction accuracy. In the classification step, the model adopts the Extreme Learning Machine with hidden nodes through data optimization to predict the presence of CKD. The complete classifier evaluation employs established measures, including recall, specificity, kappa, F-score, and accuracy, in addition to the feature selection. Data related to the study show that the proposed approach records high levels of accuracy, which is better than the existing models.

BACKGROUND: Our study aims to delineate the miRSNP-microRNA-gene-pathway interactions in the context of hypertrophic scars (HS) and keloids.
METHODS: We performed a computational biology study involving differential expression analysis to identify genes and their mRNAs in HS and keloid tissues compared to normal skin, identifying key hub genes and enriching their functional roles, comprehensively analyzing microRNA-target genes and related signaling pathways through bioinformatics, identifying MiRSNPs, and constructing a pathway-based network to illustrate miRSNP-miRNA-gene-signaling pathway interactions.
RESULTS: Our results revealed a total of 429 hub genes, with a strong enrichment in signaling pathways related to proteoglycans in cancer, focal adhesion, TGF-β, PI3K/Akt, and EGFR tyrosine kinase inhibitor resistance. Particularly noteworthy was the substantial crosstalk between the focal adhesion and PI3K/Akt signaling pathways, making them more susceptible to regulation by microRNAs. We also identified specific miRNAs, including miRNA-1279, miRNA-429, and miRNA-302e, which harbored multiple SNP loci, with miRSNPs rs188493331 and rs78979933 exerting control over a significant number of miRNA target genes. Furthermore, we observed that miRSNP rs188493331 shared a location with microRNA302e, microRNA202a-3p, and microRNA20b-5p, and these three microRNAs collectively targeted the gene LAMA3, which is integral to the focal adhesion signaling pathway.
CONCLUSIONS: The study successfully unveils the complex interactions between miRSNPs, miRNAs, genes, and signaling pathways, shedding light on the genetic factors contributing to HS and keloid formation.

暂无摘要。

Colorectal cancer (CRC) intricacies, involving dysregulated cellular processes and programmed cell death (PCD), are explored in the context of N6-methyladenosine (m6A) RNA modification. Utilizing the TCGA-COADREAD/CRC cohort, 854 m6A-related PCD genes are identified, forming the basis for a robust 10-gene risk model (CDRS) established through LASSO Cox regression. qPCR experiments using CRC cell lines and fresh tissues was performed for validation. The CDRS served as an independent risk factor for CRC and showed significant associations with clinical features, molecular subtypes, and overall survival in multiple datasets. Moreover, CDRS surpasses other predictors, unveiling distinct genomic profiles, pathway activations, and associations with the tumor microenvironment. Notably, CDRS exhibits predictive potential for drug sensitivity, presenting a novel paradigm for CRC risk stratification and personalized treatment avenues.

The utilization of vast amounts of EHR data is crucial to the studies in medical informatics. Physicians are medical participants who directly record clinical data into EHR with their personal expertise, making their roles essential in follow-up data utilization, which current studies have yet to recognize. This paper proposes a physician-centered perspective for EHR data utilization and emphasizes the feasibility and potentiality of digging into physicians\' latent decision patterns in EHR. To support our proposal, we design a physician-centered CDS approach named PhyC and test it on a real-world EHR dataset. Experiments show that PhyC performs significantly better in the auxiliary diagnosis of multiple diseases than globally learned models. Discussions on experimental results suggest that physician-centered data utilization can help to derive more objective CDS models, while more means for utilization need further exploration.

暂无摘要。

Fusion of multimodal medical data provides multifaceted, disease-relevant information for diagnosis or prognosis prediction modeling. Traditional fusion strategies such as feature concatenation often fail to learn hidden complementary and discriminative manifestations from high-dimensional multimodal data. To this end, we proposed a methodology for the integration of multimodality medical data by matching their moments in a latent space, where the hidden, shared information of multimodal data is gradually learned by optimization with multiple feature collinearity and correlation constrains. We first obtained the multimodal hidden representations by learning mappings between the original domain and shared latent space. Within this shared space, we utilized several relational regularizations, including data attribute preservation, feature collinearity and feature-task correlation, to encourage learning of the underlying associations inherent in multimodal data. The fused multimodal latent features were finally fed to a logistic regression classifier for diagnostic prediction. Extensive evaluations on three independent clinical datasets have demonstrated the effectiveness of the proposed method in fusing multimodal data for medical prediction modeling.