From pixels to patients: the evolution and future of deep learning in cancer diagnostics.

Deep learning has revolutionized cancer diagnostics, shifting from pixel-based image analysis to more comprehensive, patient-centric care. This opinion article explores recent advancements in neural network architectures, highlighting their evolution in biomedical research and their impact on medical imaging interpretation and multimodal data integration. We emphasize the need for domain-specific artificial intelligence (AI) systems capable of handling complex clinical tasks, advocating for the development of multimodal large language models that can integrate diverse data sources.

Transformer-based representation learning and multiple-instance learning for cancer diagnosis exclusively from raw sequencing fragments of bisulfite-treated plasma cell-free DNA.

Early cancer diagnosis from bisulfite-treated cell-free DNA (cfDNA) fragments requires tedious data analytical procedures. Here, we present a deep-learning-based approach for early cancer interception and diagnosis (DECIDIA) that can achieve accurate cancer diagnosis exclusively from bisulfite-treated cfDNA sequencing fragments. DECIDIA relies on transformer-based representation learning of DNA fragments and weakly supervised multiple-instance learning for classification.

Heat-dependent properties of methane diffusion in coal: an experimental study and mechanistic analysis.

Coalbed methane thermodynamic extraction, as an emerging ECBM recovery method, can effectively improve gas recovery rates. And clarifying methane diffusion and migration law in coal under thermal stimulation is crucial for the selection of its process parameters. Based on laboratory methane adsorption-release experiments, the evolution law of methane diffusion characteristics with temperature and pressure was studied, and the control mechanism of heat-dependent methane diffusion behavior was explored.

Large language model produces high accurate diagnosis of cancer from end-motif profiles of cell-free DNA.

Instruction-tuned large language models (LLMs) demonstrate exceptional ability to align with human intentions. We present an LLM-based model-instruction-tuned LLM for assessment of cancer (iLLMAC)-that can detect cancer using cell-free deoxyribonucleic acid (cfDNA) end-motif profiles. Developed on plasma cfDNA sequencing data from 1135 cancer patients and 1106 controls across three datasets, iLLMAC achieved area under the receiver operating curve (AUROC) of 0.

Development of a deep learning model for cancer diagnosis by inspecting cell-free DNA end-motifs.

Accurate discrimination between patients with and without cancer from cfDNA is crucial for early cancer diagnosis. Herein, we develop and validate a deep-learning-based model entitled end-motif inspection via transformer (EMIT) for discriminating individuals with and without cancer by learning feature representations from cfDNA end-motifs. EMIT is a self-supervised learning approach that models rankings of cfDNA end-motifs.

Prediction of tumor origin in cancers of unknown primary origin with cytology-based deep learning.

Cancer of unknown primary (CUP) site poses diagnostic challenges due to its elusive nature. Many cases of CUP manifest as pleural and peritoneal serous effusions. Leveraging cytological images from 57,220 cases at four tertiary hospitals, we developed a deep-learning method for tumor origin differentiation using cytological histology (TORCH) that can identify malignancy and predict tumor origin in both hydrothorax and ascites.

Macrophage-coated tumor cluster aggravates hepatoma invasion and immunotherapy resistance via generating local immune deprivation.

Immune checkpoint inhibitors (ICIs) represent a promising treatment for hepatocellular carcinoma (HCC) due to their capacity for abundant lymphocyte infiltration. However, some patients with HCC respond poorly to ICI therapy due to the presence of various immunosuppressive factors in the tumor microenvironment. Our research reveals that a macrophage-coated tumor cluster (MCTC) signifies a unique spatial structural organization in HCC correlating with diminished recurrence-free survival and overall survival in a total of 572 HCC cases from 3 internal cohorts and 2 independent external validation cohorts.

Language model enables end-to-end accurate detection of cancer from cell-free DNA.

We present a language model Affordable Cancer Interception and Diagnostics (ACID) that can achieve high classification performance in the diagnosis of cancer exclusively from using raw cfDNA sequencing reads. We formulate ACID as an autoregressive language model. ACID is pretrained with language sentences that are obtained from concatenation of raw sequencing reads and diagnostic labels.

Gas migration characteristics in vehicle tank under different ventilation conditions.

In recent years, due to the frequent occurrence of accidents in confined space operations, horizontal ammonia tank trucks with higher accident frequencies were selected for numerical simulation research through comparative analysis. The ammonia concentration variation characteristics of horizontal ammonia tank cars were simulated under four conditions: natural ventilation with 0° incoming air, natural ventilation with 45° incoming air, mechanical ventilation with extraction, and mechanical ventilation with compression. The results indicate that natural ventilation requires 48 h to reduce the ammonia concentration to a safe range for operation, while mechanical ventilation reduces the ammonia concentration to infinity and approaches zero within 30 min according to regulations, making the working environment safer; Set up monitoring points inside the tank to monitor the gas disturbance inside the tank at different wind speeds.

Research on psychophysiological characteristics of construction workers during consciously unsafe behaviors.

Workers' unsafe behavior is a primary cause leading to falling accidents on construction sites. This study aimed to explore how to utilize psychophysiological characteristics to predict consciously unsafe behaviors of construction workers. In this paper, a psychological questionnaire was compiled to measure risky psychology, and wireless wearable physiological recorders were employed to real-timely measure the physiological signals of subjects.

Single-Component Color-Tunable Smart Organic Emitters with Simultaneous Multistage Stimuli-Responsiveness and Multimode Emissions.

Achieving color-tunable emission in single-component organic emitters with multistage stimuli-responsiveness is of vital significance for intelligent optoelectronic applications, but remains enormously challenging. Herein, we present an unprecedented example of a color-tunable single-component smart organic emitter (DDOP) that simultaneously exhibits multistage stimuli-responsiveness and multimode emissions. DDOP based on a highly twisted amide-bridged donor-acceptor-donor structure has been found to facilitate intersystem crossing, form multimode emissions, and generate multiple emissive species with multistage stimuli-responsiveness.

Comparative analysis of domestic and foreign coal mine safety supervision modes based on knowledge map.

In order to grasp the research status and hot frontier of coal mine safety supervision mode in the world in the past 40 years, this paper takes the relevant literature in the field of "coal mine safety" and "supervision" included in the core collection of Web of Science (WOS) and the core journals of CNKI as the data source; based on the methods of statistical analysis and bibliometrics, the visualization analysis software CiteSpace is used to draw the map of scientific knowledge. Through the visualization analysis of the main research institutions, countries, and authors in this field, the main research forces and the distribution of researchers in this field are described. Through the visualization analysis of key words and research clustering, the research hotspots and future development trends in this field are described.

Generative pretraining from large-scale transcriptomes for single-cell deciphering.

Exponential accumulation of single-cell transcriptomes poses great challenge for efficient assimilation. Here, we present an approach entitled generative pretraining from transcriptomes () for learning feature representation of transcriptomes. is conceptually simple in that it autoregressive models the ranking of a gene in the context of its preceding neighbors.

Evolutionary Model and Experimental Validation of Gas-Bearing Coal Permeability under Negative Pressure Conditions.

Coal bed methane drainage is the main approach to lower risks of coal seam while raising the efficiency in natural resource utilization. The negative pressure used for extraction in coal mines is largely determined empirically due to a lack of experimental research on how coal permeability changes under the combined influence of effective stress and negative pressure. This results in low gas extraction efficiency and concentration.

Integrated analysis of racial disparities in genomic architecture identifies a trans-ancestry prognostic subtype in bladder cancer.

The incidence of bladder cancer and patient survival vary greatly among different populations, but the influence of the associated molecular features and evolutionary processes on its clinical treatment and prognostication remains unknown. Here, we analyze the genomic architectures of 505 bladder cancer patients from Asian/Black/White populations. We identify a previously unknown association between AHNAK mutations and activity of the APOBEC-a mutational signature, the activity of which varied substantially across populations.

A single-cell analysis reveals tumor heterogeneity and immune environment of acral melanoma.

Acral melanoma is a dismal subtype of melanoma occurring in glabrous acral skin, and has a higher incidence in East Asians. We perform single-cell RNA sequencing for 63,394 cells obtained from 5 acral and 3 cutaneous melanoma samples to investigate tumor heterogeneity and immune environment. We define 5 orthogonal functional cell clusters that are involved in TGF-beta signaling, Type I interferon, Wnt signaling, Cell cycle, and Cholesterol efflux signaling.

Intrinsically Stretchable Electroluminescent Elastomers with Self-Confinement Effect for Highly Efficient Non-Blended Stretchable OLEDs.

Ultra-flexible stretchable organic light-emitting diodes (OLEDs) are emerging as a basic component of flexible electronics and human-machine interfaces. However, the brightness and efficiency of stretchable OLEDs remain still far inferior to their rigid counterparts, owing to the scarcity of satisfactory stretchable electroluminescent materials. Herein, we explore a general concept based on the self-confinement effect to dramatically improve the stretchability of elastomers, without affecting electroluminescent properties.

The integrated landscape of eRNA in gastric cancer reveals distinct immune subtypes with prognostic and therapeutic relevance.

The comprehensive regulation effect of eRNA on tumor immune cell infiltration and the outcome remains obscure. We comprehensively identify the eRNA-mediated immune infiltration patterns of gastric cancer (GC) samples. We creatively proposed a random forest machine-learning (ML) algorithm to map eRNA to mRNA expression patterns.

Scalable batch-correction approach for integrating large-scale single-cell transcriptomes.

Integration of accumulative large-scale single-cell transcriptomes requires scalable batch-correction approaches. Here we propose Fugue, a simple and efficient batch-correction method that is scalable for integrating super large-scale single-cell transcriptomes from diverse sources. The core idea of the method is to encode batch information as trainable parameters and add it to single-cell expression profile; subsequently, a contrastive learning approach is used to learn feature representation of the additive expression profile.