-
BMC Bioinformatics Oct 2023Recent advancements in computing power and state-of-the-art algorithms have helped in more accessible and accurate diagnosis of numerous diseases. In addition, the...
BACKGROUND
Recent advancements in computing power and state-of-the-art algorithms have helped in more accessible and accurate diagnosis of numerous diseases. In addition, the development of de novo areas in imaging science, such as radiomics and radiogenomics, have been adding more to personalize healthcare to stratify patients better. These techniques associate imaging phenotypes with the related disease genes. Various imaging modalities have been used for years to diagnose breast cancer. Nonetheless, digital breast tomosynthesis (DBT), a state-of-the-art technique, has produced promising results comparatively. DBT, a 3D mammography, is replacing conventional 2D mammography rapidly. This technological advancement is key to AI algorithms for accurately interpreting medical images.
OBJECTIVE AND METHODS
This paper presents a comprehensive review of deep learning (DL), radiomics and radiogenomics in breast image analysis. This review focuses on DBT, its extracted synthetic mammography (SM), and full-field digital mammography (FFDM). Furthermore, this survey provides systematic knowledge about DL, radiomics, and radiogenomics for beginners and advanced-level researchers.
RESULTS
A total of 500 articles were identified, with 30 studies included as the set criteria. Parallel benchmarking of radiomics, radiogenomics, and DL models applied to the DBT images could allow clinicians and researchers alike to have greater awareness as they consider clinical deployment or development of new models. This review provides a comprehensive guide to understanding the current state of early breast cancer detection using DBT images.
CONCLUSION
Using this survey, investigators with various backgrounds can easily seek interdisciplinary science and new DL, radiomics, and radiogenomics directions towards DBT.
Topics: Humans; Female; Deep Learning; Radiographic Image Enhancement; Breast; Breast Neoplasms; Mammography
PubMed: 37884877
DOI: 10.1186/s12859-023-05515-6 -
The American Journal of Pathology Oct 2021The need for huge data sets represents a bottleneck for the application of artificial intelligence. Substantially fewer annotated target lesions than normal tissues for... (Review)
Review
The need for huge data sets represents a bottleneck for the application of artificial intelligence. Substantially fewer annotated target lesions than normal tissues for comparison present an additional problem in the field of pathology. Organic brains overcome these limitations by utilizing large numbers of specialized neural nets arranged in both linear and parallel fashion, with each solving a restricted classification problem. They rely on local Hebbian error corrections as compared to the nonlocal back-propagation used in most artificial neural nets, and leverage reinforcement. For these reasons, even toddlers are able to classify objects after only a few examples. Rather than provide an overview of current AI research in pathology, this review focuses on general strategies for overcoming the data bottleneck. These include transfer learning, zero-shot learning, Siamese networks, one-class models, generative networks, and reinforcement learning. Neither an extensive mathematic background nor advanced programing skills are needed to make these subjects accessible to pathologists. However, some familiarity with the basic principles of deep learning, briefly reviewed here, is expected to be useful in understanding both the current limitations of machine learning and determining ways to address them.
Topics: Algorithms; Data Curation; Humans; Machine Learning; Neural Networks, Computer; Quantum Theory
PubMed: 34129843
DOI: 10.1016/j.ajpath.2021.05.023 -
Journal of the American Society of... Aug 2022The risk of cardiovascular events rises after AKI. Leukocytes promote atherosclerotic plaque growth and instability. We established a model of enhanced remote...
BACKGROUND
The risk of cardiovascular events rises after AKI. Leukocytes promote atherosclerotic plaque growth and instability. We established a model of enhanced remote atherosclerosis after renal ischemia-reperfusion (IR) injury and investigated the underlying inflammatory mechanisms.
METHODS
Atherosclerotic lesions and inflammation were investigated in native and bone marrow-transplanted LDL receptor-deficient ( ) mice after unilateral renal IR injury using histology, flow cytometry, and gene expression analysis.
RESULTS
Aortic root atherosclerotic lesions were significantly larger after renal IR injury than in controls. A gene expression screen revealed enrichment for chemokines and their cognate receptors in aortas of IR-injured mice in early atherosclerosis, and of T cell-associated genes in advanced disease. Confocal microscopy revealed increased aortic macrophage proximity to T cells. Differential aortic inflammatory gene regulation in IR-injured mice largely paralleled the pattern in the injured kidney. Single-cell analysis identified renal cell types that produced soluble mediators upregulated in the atherosclerotic aorta. The analysis revealed a marked early increase in , which CCR2 myeloid cells mainly expressed. CCR2 mediated myeloid cell homing to the post-ischemic kidney in a cell-individual manner. Reconstitution with bone marrow dampened renal post-ischemic inflammation, reduced aortic and inflammatory macrophage marker CD11c, and abrogated excess aortic atherosclerotic plaque formation after renal IR.
CONCLUSIONS
Our data introduce an experimental model of remote proatherogenic effects of renal IR and delineate myeloid CCR2 signaling as a mechanistic requirement. Monocytes should be considered as mobile mediators when addressing systemic vascular sequelae of kidney injury.
Topics: Mice; Animals; Plaque, Atherosclerotic; Atherosclerosis; Monocytes; Inflammation; Ischemia; Reperfusion Injury; Acute Kidney Injury; Mice, Inbred C57BL; Receptors, CCR2; Mice, Knockout
PubMed: 35537780
DOI: 10.1681/ASN.2022010048 -
Journal of Integrative Neuroscience May 2023Lesions of the central nervous system (CNS) can present with numerous and overlapping radiographical and clinical features that make diagnosis difficult based... (Review)
Review
Lesions of the central nervous system (CNS) can present with numerous and overlapping radiographical and clinical features that make diagnosis difficult based exclusively on history, physical examination, and traditional imaging modalities. Given that there are significant differences in optimal treatment protocols for these various CNS lesions, rapid and non-invasive diagnosis could lead to improved patient care. Recently, various advanced magnetic resonance imaging (MRI) techniques showed promising methods to differentiate between various tumors and lesions that conventional MRI cannot define by comparing their physiologic characteristics, such as vascularity, permeability, oxygenation, and metabolism. These advanced MRI techniques include dynamic susceptibility contrast MRI (DSC), diffusion-weighted imaging (DWI), dynamic contrast-enhanced (DCE) MRI, Golden-Angle Radial Sparse Parallel imaging (GRASP), Blood oxygen level-dependent functional MRI (BOLD fMRI), and arterial spin labeling (ASL) MRI. In this article, a narrative review is used to discuss the current trends in advanced MRI techniques and potential future applications in identifying difficult-to-distinguish CNS lesions. Advanced MRI techniques were found to be promising non-invasive modalities to differentiate between paraganglioma, schwannoma, and meningioma. They are also considered promising methods to differentiate gliomas from lymphoma, post-radiation changes, pseudoprogression, demyelination, and metastasis. Advanced MRI techniques allow clinicians to take advantage of intrinsic biological differences in CNS lesions to better identify the etiology of these lesions, potentially leading to more effective patient care and a decrease in unnecessary invasive procedures. More clinical studies with larger sample sizes should be encouraged to assess the significance of each advanced MRI technique and the specificity and sensitivity of each radiologic parameter.
Topics: Humans; Brain Neoplasms; Magnetic Resonance Imaging; Central Nervous System Neoplasms; Glioma; Meningeal Neoplasms
PubMed: 37258452
DOI: 10.31083/j.jin2203073 -
Journal of Experimental & Clinical... Jun 2023Osteosarcoma (OS) is the most common primary bone tumor in children and adolescent. Surgery and multidrug chemotherapy are the standard of treatment achieving 60-70% of...
BACKGROUND
Osteosarcoma (OS) is the most common primary bone tumor in children and adolescent. Surgery and multidrug chemotherapy are the standard of treatment achieving 60-70% of event-free survival for localized disease at diagnosis. However, for metastatic disease, the prognosis is dismal. Exploiting immune system activation in the setting of such unfavorable mesenchymal tumors represents a new therapeutic challenge.
METHODS
In immune competent OS mouse models bearing two contralateral lesions, we tested the efficacy of intralesional administration of a TLR9 agonist against the treated and not treated contralateral lesion evaluating abscopal effect. Multiparametric flow cytometry was used to evaluate changes of the tumor immune microenviroment. Experiments in immune-deficient mice allowed the investigation of the role of adaptive T cells in TLR9 agonist effects, while T cell receptor sequencing was used to assess the expansion of specific T cell clones.
RESULTS
TLR9 agonist strongly impaired the growth of locally-treated tumors and its therapeutic effect also extended to the contralateral, untreated lesion. Multiparametric flow cytometry showed conspicuous changes in the immune landscape of the OS immune microenvironment upon TLR9 engagement, involving a reduction in M2-like macrophages, paralleled by increased infiltration of dendritic cells and activated CD8 T cells in both lesions. Remarkably, CD8 T cells were needed for the induction of the abscopal effect, whereas they were not strictly necessary for halting the growth of the treated lesion. T cell receptor (TCR) sequencing of tumor infiltrating CD8 T cells showed the expansion of specific TCR clones in the treated tumors and, remarkably, their selected representation in the contralateral untreated lesions, providing the first evidence of the rewiring of tumor-associated T cell clonal architectures.
CONCLUSIONS
Overall these data indicate that the TLR9 agonist acts as an in situ anti-tumor vaccine, activating an innate immune response sufficient to suppress local tumor growth while inducing a systemic adaptive immunity with selective expansion of CD8 T cell clones, which are needed for the abscopal effect.
Topics: Animals; Mice; Toll-Like Receptor 9; CD8-Positive T-Lymphocytes; Adaptive Immunity; Osteosarcoma; Bone Neoplasms; Tumor Microenvironment
PubMed: 37365634
DOI: 10.1186/s13046-023-02731-z -
Experimental Neurology May 2023Growing preclinical and clinical evidence highlights neurosteroid pathway imbalances in Parkinson's Disease (PD) and L-DOPA-induced dyskinesias (LIDs). We recently...
Growing preclinical and clinical evidence highlights neurosteroid pathway imbalances in Parkinson's Disease (PD) and L-DOPA-induced dyskinesias (LIDs). We recently reported that 5α-reductase (5AR) inhibitors dampen dyskinesias in parkinsonian rats; however, unraveling which specific neurosteroid mediates this effect is critical to optimize a targeted therapy. Among the 5AR-related neurosteroids, striatal pregnenolone has been shown to be increased in response to 5AR blockade and decreased after 6-OHDA lesions in the rat PD model. Moreover, this neurosteroid rescued psychotic-like phenotypes by exerting marked antidopaminergic activity. In light of this evidence, we investigated whether pregnenolone might dampen the appearance of LIDs in parkinsonian drug-naïve rats. We tested 3 escalating doses of pregnenolone (6, 18, 36 mg/kg) in 6-OHDA-lesioned male rats and compared the behavioral, neurochemical, and molecular outcomes with those induced by the 5AR inhibitor dutasteride, as positive control. The results showed that pregnenolone dose-dependently countered LIDs without affecting L-DOPA-induced motor improvements. Post-mortem analyses revealed that pregnenolone significantly prevented the increase of validated striatal markers of dyskinesias, such as phospho-Thr-34 DARPP-32 and phospho-ERK, as well as D-D receptor co-immunoprecipitation in a fashion similar to dutasteride. Moreover, the antidyskinetic effect of pregnenolone was paralleled by reduced striatal levels of BDNF, a well-established factor associated with the development of LIDs. In support of a direct pregnenolone effect, LC/MS-MS analyses revealed that striatal pregnenolone levels strikingly increased after the exogenous administration, with no significant alterations in downstream metabolites. All these data suggest pregnenolone as a key player in the antidyskinetic properties of 5AR inhibitors and highlight this neurosteroid as an interesting novel tool to target LIDs in PD.
Topics: Male; Rats; Animals; Levodopa; Parkinson Disease; Dutasteride; Oxidopamine; Neurosteroids; Rats, Sprague-Dawley; Dyskinesia, Drug-Induced; Corpus Striatum; Antiparkinson Agents; Disease Models, Animal
PubMed: 36878398
DOI: 10.1016/j.expneurol.2023.114370 -
IEEE Access : Practical Innovations,... 2023Artificial Intelligence (AI)-based medical computer vision algorithm training and evaluations depend on annotations and labeling. However, variability between expert...
Artificial Intelligence (AI)-based medical computer vision algorithm training and evaluations depend on annotations and labeling. However, variability between expert annotators introduces noise in training data that can adversely impact the performance of AI algorithms. This study aims to assess, illustrate and interpret the inter-annotator agreement among multiple expert annotators when segmenting the same lesion(s)/abnormalities on medical images. We propose the use of three metrics for the qualitative and quantitative assessment of inter-annotator agreement: 1) use of a common agreement heatmap and a ranking agreement heatmap; 2) use of the extended Cohen's kappa and Fleiss' kappa coefficients for a quantitative evaluation and interpretation of inter-annotator reliability; and 3) use of the Simultaneous Truth and Performance Level Estimation (STAPLE) algorithm, as a parallel step, to generate ground truth for training AI models and compute Intersection over Union (IoU), sensitivity, and specificity to assess the inter-annotator reliability and variability. Experiments are performed on two datasets, namely cervical colposcopy images from 30 patients and chest X-ray images from 336 tuberculosis (TB) patients, to demonstrate the consistency of inter-annotator reliability assessment and the importance of combining different metrics to avoid bias assessment.
PubMed: 37008654
DOI: 10.1109/access.2023.3249759 -
Clinical and Experimental Dental... Aug 2021Globally, has been an increase in the use of silver fluoride products to arrest carious lesions and a variety of products are available. (Randomized Controlled Trial)
Randomized Controlled Trial
BACKGROUND
Globally, has been an increase in the use of silver fluoride products to arrest carious lesions and a variety of products are available.
OBJECTIVES
To examine differences in caries arrest and lesion colour of primary tooth carious lesions.
MATERIAL AND METHODS
A four-armed, parallel-design cluster-randomised controlled trial which investigated four protocols for caries arrest at 6m and 12m. Children in Group 1 and Group 2 received Rivastar Silver Diammine Fluoride (SDF), and children in Group 3 and Group 4 received a stabilised aqueous silver fluoride solution (AgF). Children in Group 2 and Group 4 received an additional application of KI immediately after the fluoride. Differences in caries arrest and lesion appearance were examined at 6m and 12m using two level logistic regression modelling.
RESULTS
The arrest rate varied by group membership; group 1 and group 3 had higher arrest rates (77.3% and 75.3% respectively) than group 2 and group 4 (65.4% and 51.2% respectively). The use of KI was also associated with lower odds of arrest (12m OR 0.25; CI 0.19, 0.34) and higher odds of avoiding black discolouration (12m OR 6.08; 2.36, 15.67).
CONCLUSIONS
Globally, has been an increase in the use of silver fluoride products to arrest carious lesions and a variety of products are available. This study demonstrated that both AgF and SDF can effectively arrest carious lesions on primary teeth. The use of KI is associated with poorer caries control but better aesthetic outcomes.
Topics: Cariostatic Agents; Child; Dental Caries; Dental Caries Susceptibility; Fluorides; Humans; Potassium Iodide; Silver Compounds; Tooth, Deciduous
PubMed: 33370847
DOI: 10.1002/cre2.367 -
Frontiers in Neuroscience 2021In recent years, an increasing number of people have myopia in China, especially the younger generation. Common myopia may develop into high myopia. High myopia causes...
In recent years, an increasing number of people have myopia in China, especially the younger generation. Common myopia may develop into high myopia. High myopia causes visual impairment and blindness. Parapapillary atrophy (PPA) is a typical retinal pathology related to high myopia, which is also a basic clue for diagnosing high myopia. Therefore, accurate segmentation of the PPA is essential for high myopia diagnosis and treatment. In this study, we propose an optimized Unet (OT-Unet) to solve this important task. OT-Unet uses one of the pre-trained models: Visual Geometry Group (VGG), ResNet, and Res2Net, as a backbone and is combined with edge attention, parallel partial decoder, and reverse attention modules to improve the segmentation accuracy. In general, using the pre-trained models can improve the accuracy with fewer samples. The edge attention module extracts contour information, the parallel partial decoder module combines the multi-scale features, and the reverse attention module integrates high- and low-level features. We also propose an augmented loss function to increase the weight of complex pixels to enable the network to segment more complex lesion areas. Based on a dataset containing 360 images (Including 26 pictures provided by PALM), the proposed OT-Unet achieves a high AUC (Area Under Curve) of 0.9235, indicating a significant improvement over the original Unet (0.7917).
PubMed: 34720868
DOI: 10.3389/fnins.2021.758887