WIREs Mechanisms of Disease最新文献

The increasing availability of large collections of electronic health record (EHR) data and unprecedented technical advances in deep learning (DL) have sparked a surge of research interest in developing DL based clinical decision support systems for diagnosis, prognosis, and treatment. Despite the recognition of the value of deep learning in healthcare, impediments to further adoption in real healthcare settings remain due to the black-box nature of DL. Therefore, there is an emerging need for interpretable DL, which allows end users to evaluate the model decision making to know whether to accept or reject predictions and recommendations before an action is taken. In this review, we focus on the interpretability of the DL models in healthcare. We start by introducing the methods for interpretability in depth and comprehensively as a methodological reference for future researchers or clinical practitioners in this field. Besides the methods' details, we also include a discussion of advantages and disadvantages of these methods and which scenarios each of them is suitable for, so that interested readers can know how to compare and choose among them for use. Moreover, we discuss how these methods, originally developed for solving general-domain problems, have been adapted and applied to healthcare problems and how they can help physicians better understand these data-driven technologies. Overall, we hope this survey can help researchers and practitioners in both artificial intelligence and clinical fields understand what methods we have for enhancing the interpretability of their DL models and choose the optimal one accordingly. This article is categorized under: Cancer > Computational Models.

Human papillomaviruses (HPVs) are the etiological agent of a significant, and increasing, fraction of head and neck squamous cell carcinomas (HNSCC)-a heterogenous group of malignancies in the head and neck region. HPV infection accounts for approximately 25% of all cases, with the remainder typically caused by smoking and excessive alcohol consumption. These distinct etiologies lead to profound clinical and immunological differences between HPV-positive (HPV⁺ ) and HPV-negative (HPV^- ) HNSCC, likely related to the expression of exogenous viral antigens in the HPV⁺ subtype. Specifically, HPV⁺ HNSCC patients generally exhibit better treatment response compared to those with HPV^- disease, leading to a more favorable prognosis, with lower recurrence rate, and longer overall survival time. Importantly, a plethora of studies have illustrated that the tumor immune microenvironment (TIME) of HPV⁺ HNSCC has a strikingly distinct immune composition to that of its HPV^- counterpart. The HPV⁺ TIME is characterized as being immunologically "hot," with more immune infiltration, higher levels of T-cell activation, and higher levels of immunoregulation compared to the more immunologically "cold" HPV^- TIME. In general, cancers with an immune "hot" TIME exhibit better treatment response and superior clinical outcomes in comparison to their immune "cold" counterparts. Indeed, this phenomenon has also been observed in HPV⁺ HNSCC patients, highlighting the critical role of the TIME in influencing prognosis, and further validating the use of cancer therapies that capitalize on the mobilization and/or modulation of the TIME. This article is categorized under: Cancer > Molecular and Cellular Physiology Infectious Diseases > Molecular and Cellular Physiology.

Alzheimer's disease (AD) is a debilitating neurodegenerative disorder affecting over five million people globally and has no established cure. Current AD-related treatments only alleviate cognitive and behavioral symptoms and do not address disease onset or progression, underlining the unmet need to create an effective, innovative AD therapeutic. Extracellular vesicles (EVs) have emerged as a new class of nanotherapeutics. These secreted, lipid-bound cellular signaling carriers show promise for potential clinical applications for neurodegenerative diseases like AD. Additionally, analyzing contents and characteristics of patient-derived EVs may address the unmet need for earlier AD diagnostic techniques, informing physicians of altered genetic expression or cellular communications specific to healthy and diseased physiological states. There are numerous recent advances in regenerative medicine using EVs and include bioengineering perspectives to modify EVs, target glial cells in neurodegenerative diseases like AD, and potentially use EVs to diagnose and treat AD earlier. This article is categorized under: Neurological Diseases > Biomedical Engineering Neurological Diseases > Molecular and Cellular Physiology Neurological Diseases > Stem Cells and Development.

The lateral hypothalamus is critical for the control of ingestive behavior and spontaneous physical activity (SPA), as lesion or stimulation of this region alters these behaviors. Evidence points to lateral hypothalamic orexin neurons as modulators of feeding and SPA. These neurons affect a broad range of systems, and project to multiple brain regions such as the dorsal raphe nucleus, which contains serotoninergic neurons (DRN) important to energy homeostasis. Physical activity is comprised of intentional exercise and SPA. These are opposite ends of a continuum of physical activity intensity and structure. Non-goal-oriented behaviors, such as fidgeting, standing, and ambulating, constitute SPA in humans, and reflect a propensity for activity separate from intentional activity, such as high-intensity voluntary exercise. In animals, SPA is activity not influenced by rewards such as food or a running wheel. Spontaneous physical activity in humans and animals burns calories and could theoretically be manipulated pharmacologically to expend calories and protect against obesity. The DRN neurons receive orexin inputs, and project heavily onto cortical and subcortical areas involved in movement, feeding and energy expenditure (EE). This review discusses the function of hypothalamic orexin in energy-homeostasis, the interaction with DRN serotonin neurons, and the role of this orexin-serotonin axis in regulating food intake, SPA, and EE. In addition, we discuss possible brain areas involved in orexin-serotonin cross-talk; the role of serotonin receptors, transporters and uptake-inhibitors in the pathogenesis and treatment of obesity; animal models of obesity with impaired serotonin-function; single-nucleotide polymorphisms in the serotonin system and obesity; and future directions in the orexin-serotonin field. This article is categorized under: Metabolic Diseases > Molecular and Cellular Physiology.

Animal models are useful to study the molecular, cellular, and morphogenetic mechanisms underlying normal and pathological development. Cell-based study models have emerged as an alternative approach to study many aspects of human embryonic development and disease. The neural crest (NC) is a transient, multipotent, and migratory embryonic cell population that generates a diverse group of cell types that arises during vertebrate development. The abnormal formation or development of the NC results in neurocristopathies (NCPs), which are characterized by a broad spectrum of functional and morphological alterations. The impaired molecular mechanisms that give rise to these multiphenotypic diseases are not entirely clear yet. This fact, added to the high incidence of these disorders in the newborn population, has led to the development of systematic approaches for their understanding. In this article, we have systematically reviewed the ways in which experimentation with different animal and cell model systems has improved our knowledge of NCPs, and how these advances might contribute to the development of better diagnostic and therapeutic tools for the treatment of these pathologies. This article is categorized under: Congenital Diseases > Genetics/Genomics/Epigenetics Congenital Diseases > Stem Cells and Development Congenital Diseases > Molecular and Cellular Physiology Neurological Diseases > Genetics/Genomics/Epigenetics.