Advantages of transformer and its application for medical image segmentation: a survey.

PURPOSE: Convolution operator-based neural networks have shown great success in medical image segmentation over the past decade. The U-shaped network with a codec structure is one of the most widely used models. Transformer, a technology used in natural language processing, can capture long-distance dependencies and has been applied in Vision Transformer to achieve state-of-the-art performance on image classification tasks. Recently, researchers have extended transformer to medical image segmentation tasks, resulting in good models. METHODS: This review comprises publications selected through a Web of Science search. We focused on papers published since 2018 that applied the transformer architecture to medical image segmentation. We conducted a systematic analysis of these studies and summarized the results. RESULTS: To better comprehend the benefits of convolutional neural networks and transformers, the construction of the codec and transformer modules is first explained. Second, the medical image segmentation model based on transformer is summarized. The typically used assessment markers for medical image segmentation tasks are then listed. Finally, a large number of medical segmentation datasets are described. CONCLUSION: Even if there is a pure transformer model without any convolution operator, the sample size of medical picture segmentation still restricts the growth of the transformer, even though it can be relieved by a pretraining model. More often than not, researchers are still designing models using transformer and convolution operators.

Protonation-mediated DNA tile self-assembly with nuclease resistance characteristic for signal-amplified detection of microRNAs.

DNA nanotechnology, developing rapidly in recent years, has unprecedented superiorities in biological application-oriented research including high programmability, convenient functionalization, reconfigurable structure, and intrinsic biocompatibility. However, the susceptibility to nucleases in the physiological environment has been an obstacle to applying DNA nanostructures in biological science research. In this study, a new DNA self-assembly strategy, mediated by double-protonated small molecules instead of classical metal ions, is developed to enhance the nuclease resistance of DNA nanostructures while retaining their integrality and functionality, and the relative application has been launched in the detection of microRNAs (miRNAs). Faced with low-abundance miRNAs, we integrate hybrid chain reaction (HCR) with DNA self-assembly in the presence of double-protonated small molecules to construct a chemiluminescence detection platform with nuclease resistance, which utilizes the significant difference of molecular weight between DNA arrays and false-positive products to effectively separate of reaction products and remove the detection background. This strategy attaches importance to the nucleic acid stability during the assay process via improving nuclease resistance while rendering the detection results for miRNAs more authentic and reliable, opening our eyes to more possibilities for the multiple applications of customized DNA nanostructures in biology, including bioassay, bioimaging, drug delivery, and cell modulation.

A specific and low background nucleic acids sensing strategy based on rolling circle amplification coupled with a magnetic DNA machine.

We propose a universal fluorescence method for detection of nucleic acids based on rolling circle amplification (RCA) combined with a magnetic DNA machine and using dengue virus nucleic acids as an example target. RCA specifically amplifies the target and yields a large number of initiators employing heat-labile double-stranded DNase. The magnetic DNA machine produces a fluorescence signal and eliminates background noise. This method achieved a wide linear range, promising recovery and ultrahigh recognition specificity for one-base mismatches, and indicates the potential application of this sensing strategy in the clinical diagnosis of nucleic acids of pathogens.

Size-Controlled DNA Tile Self-Assembly Nanostructures Through Caveolae-Mediated Endocytosis for Signal-Amplified Imaging of MicroRNAs in Living Cells.

Signal-amplified imaging of microRNAs (miRNAs) is a promising strategy at the single-cell level because liquid biopsy fails to reflect real-time dynamic miRNA levels. However, the internalization pathways for available conventional vectors predominantly involve endo-lysosomes, showing nonideal cytoplasmic delivery efficiency. In this study, size-controlled 9-tile nanoarrays are designed and constructed by integrating catalytic hairpin assembly (CHA) with DNA tile self-assembly technology to achieve caveolae-mediated endocytosis for the amplified imaging of miRNAs in a complex intracellular environment. Compared with classical CHA, the 9-tile nanoarrays possess high sensitivity and specificity for miRNAs, achieve excellent internalization efficiency by caveolar endocytosis, bypassing lysosomal traps, and exhibit more powerful signal-amplified imaging of intracellular miRNAs. Because of their excellent safety, physiological stability, and highly efficient cytoplasmic delivery, the 9-tile nanoarrays can realize real-time amplified monitoring of miRNAs in various tumor and identical cells of different periods, and imaging effects are consistent with the actual expression levels of miRNAs, ultimately demonstrating their feasibility and capacity. This strategy provides a high-potential delivery pathway for cell imaging and targeted delivery, simultaneously offering a meaningful reference for the application of DNA tile self-assembly technology in relevant fundamental research and medical diagnostics.