In this paper, we give a short introduction to machine learning and survey its applications in radiology. We focused on six categories of applications in radiology: medical image segmentation, registration, computer aided detection and diagnosis, brain function or activity analysis and neurological disease diagnosis from fMR images, content-based image retrieval systems for CT or MRI images, and text analysis of radiology reports using natural language processing (NLP) and natural language understanding (NLU). This survey shows that machine learning plays a key role in many radiology applications. Machine learning identifies complex patterns automatically and helps radiologists make intelligent decisions on radiology data such as conventional radiographs, CT, MRI, and PET images and radiology reports. In many applications, the performance of machine learning-based automatic detection and diagnosis systems has shown to be comparable to that of a well-trained and experienced radiologist. Technology development in machine learning and radiology will benefit from each other in the long run. Key contributions and common characteristics of machine learning techniques in radiology are discussed. We also discuss the problem of translating machine learning applications to the radiology clinical setting, including advantages and potential barriers.
The rapidly increasing global demand for energy combined with the environmental impact of fossil fuels has spurred the search for alternative sources of clean energy. One promising approach is to convert solar energy into hydrogen fuel using photoelectrochemical cells. However, the semiconducting photoelectrodes used in these cells typically have low efficiencies and/or stabilities. Here we show that a silicon-based photocathode with a capping epitaxial oxide layer can provide efficient and stable hydrogen production from water. In particular, a thin epitaxial layer of strontium titanate (SrTiO3) was grown directly on Si(001) by molecular beam epitaxy. Photogenerated electrons can be transported easily through this layer because of the conduction-band alignment and lattice match between single-crystalline SrTiO3 and silicon. The approach was used to create a metal-insulator-semiconductor photocathode that, under a broad-spectrum illumination at 100 mW cm(-2), exhibits a maximum photocurrent density of 35 mA cm(-2) and an open circuit potential of 450 mV; there was no observable decrease in performance after 35 hours of operation in 0.5 M H2SO4. The performance of the photocathode was also found to be highly dependent on the size and spacing of the structured metal catalyst. Therefore, mesh-like Ti/Pt nanostructured catalysts were created using a nanosphere lithography lift-off process and an applied-bias photon-to-current efficiency of 4.9% was achieved.
The mechanism of leukocyte migration through venular walls in vivo is largely unknown. By using immunofluorescence staining and confocal microscopy, the present study demonstrates the existence of regions within the walls of unstimulated murine cremasteric venules where expression of key vascular basement membrane (BM) constituents, laminin 10, collagen IV, and nidogen-2 (but not perlecan) are considerably lower (<60%) than the average expression detected in the same vessel. These sites were closely associated with gaps between pericytes and were preferentially used by migrating neutrophils during their passage through cytokine-stimulated venules. Although neutrophil transmigration did not alter the number/unit area of extracellular matrix protein low expression sites, the size of these regions was enlarged and their protein content was reduced in interleukin-1β–stimulated venules. These effects were entirely dependent on the presence of neutrophils and appeared to involve neutrophil-derived serine proteases. Furthermore, evidence was obtained indicating that transmigrating neutrophils carry laminins on their cell surface in vivo. Collectively, through identification of regions of low extracellular matrix protein localization that define the preferred route for transmigrating neutrophils, we have identified a plausible mechanism by which neutrophils penetrate the vascular BM without causing a gross disruption to its intricate structure.
Automated Lymph Node (LN) detection is an important clinical diagnostic task but very challenging due to the low contrast of surrounding structures in Computed Tomography (CT) and to their varying sizes, poses, shapes and sparsely distributed locations. State-of-the-art studies show the performance range of 52.9% sensitivity at 3.1 false-positives per volume (FP/vol.), or 60.9% at 6.1 FP/vol. for mediastinal LN, by one-shot boosting on 3D HAAR features. In this paper, we first operate a preliminary candidate generation stage, towards ~100% sensitivity at the cost of high FP levels (~40 per patient), to harvest volumes of interest (VOI). Our 2.5D approach consequently decomposes any 3D VOI by resampling 2D reformatted orthogonal views N times, via scale, random translations, and rotations with respect to the VOI centroid coordinates. These random views are then used to train a deep Convolutional Neural Network (CNN) classifier. In testing, the CNN is employed to assign LN probabilities for all N random views that can be simply averaged (as a set) to compute the final classification probability per VOI. We validate the approach on two datasets: 90 CT volumes with 388 mediastinal LNs and 86 patients with 595 abdominal LNs. We achieve sensitivities of 70%/83% at 3 FP/vol. and 84%/90% at 6 FP/vol. in mediastinum and abdomen respectively, which drastically improves over the previous state-of-theart work.
Focal segmental glomerulosclerosis (FSGS) is a frequent and severe glomerular disease characterized by destabilization of podocyte foot processes. We report that transgenic expression of the microRNA miR-193a in mice rapidly induces FSGS with extensive podocyte foot process effacement. Mechanistically, miR-193a inhibits the expression of the Wilms' tumor protein (WT1), a transcription factor and master regulator of podocyte differentiation and homeostasis. Decreased expression levels of WT1 lead to downregulation of its target genes PODXL (podocalyxin) and NPHS1 (nephrin), as well as several other genes crucial for the architecture of podocytes, initiating a catastrophic collapse of the entire podocyte-stabilizing system. We found upregulation of miR-193a in isolated glomeruli from individuals with FSGS compared to normal kidneys or individuals with other glomerular diseases. Thus, upregulation of miR-193a provides a new pathogenic mechanism for FSGS and is a potential therapeutic target.
Solid polymer electrolytes (SPEs) with tunable network structures are prepared by a facile one-pot reaction of polyhedral oligomeric silsesquioxane and poly(ethylene glycol). These SPEs, with high conductivity and high modulus, exhibit superior resistance to lithium dendrite growth even at high current densities. Measurements of lithium metal batteries with a LiFePO4 cathode show excellent cycling stability and rate capability.
Tissue homeostasis and remodeling are processes that involve high turnover of biological macromolecules. Many of the waste molecules that are by-products or degradation intermediates of biological macromolecule turnover enter the circulation and are subsequently cleared by liver sinusoidal endothelial cells (LSEC). Besides the mannose receptor, stabilin-1 and stabilin-2 are the major scavenger receptors expressed by LSEC. To more clearly elucidate the functions of stabilin-1 and -2, we have generated mice lacking stabilin-1, stabilin-2, or both stabilin-1 and -2 (Stab1 -/-Stab2 -/-mice). Mice lacking either stabilin-1 or stabilin-2 were phenotypically normal; however, Stab1 -/-Stab2 -/-mice exhibited premature mortality and developed severe glomerular fibrosis, while the liver showed only mild perisinusoidal fibrosis without dysfunction. Upon kidney transplantation into WT mice, progression of glomerular fibrosis was halted, indicating the presence of profibrotic factors in the circulation of Stab1 -/-Stab2 -/-mice. While plasma levels of known profibrotic cytokines were unaltered, clearance of the TGF-β family member growth differentiation factor 15 (GDF-15) was markedly impaired in Stab1 -/-Stab2 -/-mice but not in either Stab1 -/-or Stab2 -/-mice, indicating that it is a common ligand of both stabilin-1 and stabilin-2. These data lead us to conclude that stabilin-1 and -2 together guarantee proper hepatic clearance of potentially noxious agents in the blood and maintain tissue homeostasis not only in the liver but also distant organs.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
334 Leonard St
Brooklyn, NY 11211
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.