An accessible deep learning tool for voxel-wise classification of brain malignancies from perfusion MRI.

Noninvasive differential diagnosis of brain tumors is currently based on the assessment of magnetic resonance imaging (MRI) coupled with dynamic susceptibility contrast (DSC). However, a definitive diagnosis often requires neurosurgical interventions that compromise patients' quality of life. We apply deep learning on DSC images from histology-confirmed patients with glioblastoma, metastasis, or lymphoma. The convolutional neural network trained on ∼50,000 voxels from 40 patients provides intratumor probability maps that yield clinical-grade diagnosis. Performance is tested in 400 additional cases and an external validation cohort of 128 patients. The tool reaches a three-way accuracy of 0.78, superior to the conventional MRI metrics cerebral blood volume (0.55) and percentage of signal recovery (0.59), showing high value as a support diagnostic tool. Our open-access software, Diagnosis In Susceptibility Contrast Enhancing Regions for Neuro-oncology (DISCERN), demonstrates its potential in aiding medical decisions for brain tumor diagnosis using standard-of-care MRI.

Radiomics in CT and MR imaging of the liver and pancreas: tools with potential for clinical application.

Radiomics allows the extraction of quantitative imaging features from clinical magnetic resonance imaging (MRI) and computerized tomography (CT) studies. The advantages of radiomics have primarily been exploited in oncological applications, including better characterization and staging of oncological lesions and prediction of patient outcomes and treatment response. The potential introduction of radiomics in the clinical setting requires the establishment of a standardized radiomics pipeline and a quality assurance program. Radiomics and texture analysis of the liver have improved the differentiation of hypervascular lesions such as adenomas, focal nodular hyperplasia, and hepatocellular carcinoma (HCC) during the arterial phase, and in the pretreatment determination of HCC prognostic factors (e.g., tumor grade, microvascular invasion, Ki-67 proliferation index). Radiomics of pancreatic CT and MR images has enhanced pancreatic ductal adenocarcinoma detection and its differentiation from pancreatic neuroendocrine tumors, mass-forming chronic pancreatitis, or autoimmune pancreatitis. Radiomics can further help to better characterize incidental pancreatic cystic lesions, accurately discriminating benign from malignant intrapancreatic mucinous neoplasms. Nonetheless, despite their encouraging results and exciting potential, these tools have yet to be implemented in the clinical setting. This non-systematic review will describe the essential steps in the implementation of the radiomics and feature extraction workflow from liver and pancreas CT and MRI studies for their potential clinical application. A succinct overview of reported radiomics applications in the liver and pancreas and the challenges and limitations of their implementation in the clinical setting is also discussed, concluding with a brief exploration of the future perspectives of radiomics in the gastroenterology field.

Clinical impact of artificial intelligence-based solutions on imaging of the pancreas and liver.

Artificial intelligence (AI) has experienced substantial progress over the last ten years in many fields of application, including healthcare. In hepatology and pancreatology, major attention to date has been paid to its application to the assisted or even automated interpretation of radiological images, where AI can generate accurate and reproducible imaging diagnosis, reducing the physicians' workload. AI can provide automatic or semi-automatic segmentation and registration of the liver and pancreatic glands and lesions. Furthermore, using radiomics, AI can introduce new quantitative information which is not visible to the human eye to radiological reports. AI has been applied in the detection and characterization of focal lesions and diffuse diseases of the liver and pancreas, such as neoplasms, chronic hepatic disease, or acute or chronic pancreatitis, among others. These solutions have been applied to different imaging techniques commonly used to diagnose liver and pancreatic diseases, such as ultrasound, endoscopic ultrasonography, computerized tomography (CT), magnetic resonance imaging, and positron emission tomography/CT. However, AI is also applied in this context to many other relevant steps involved in a comprehensive clinical scenario to manage a gastroenterological patient. AI can also be applied to choose the most convenient test prescription, to improve image quality or accelerate its acquisition, and to predict patient prognosis and treatment response. In this review, we summarize the current evidence on the application of AI to hepatic and pancreatic radiology, not only in regard to the interpretation of images, but also to all the steps involved in the radiological workflow in a broader sense. Lastly, we discuss the challenges and future directions of the clinical application of AI methods.

Role of artificial intelligence in multidisciplinary imaging diagnosis of gastrointestinal diseases.

The use of artificial intelligence-based tools is regarded as a promising approach to increase clinical efficiency in diagnostic imaging, improve the interpretability of results, and support decision-making for the detection and prevention of diseases. Radiology, endoscopy and pathology images are suitable for deep-learning analysis, potentially changing the way care is delivered in gastroenterology. The aim of this review is to examine the key aspects of different neural network architectures used for the evaluation of gastrointestinal conditions, by discussing how different models behave in critical tasks, such as lesion detection or characterization (i.e. the distinction between benign and malignant lesions of the esophagus, the stomach and the colon). To this end, we provide an overview on recent achievements and future prospects in deep learning methods applied to the analysis of radiology, endoscopy and histologic whole-slide images of the gastrointestinal tract.

Targeting the CRD F-face of Human Galectin-3 and Allosterically Modulating Glycan Binding by Angiostatic PTX008 and a Structurally Optimized Derivative.

Calix[4]arene PTX008 is an angiostatic agent that inhibits tumor growth in mice by binding to galectin-1, a ß-galactoside-binding lectin. To assess the affinity profile of PTX008 for galectins, we used 15 N,1 H HSQC NMR spectroscopy to show that PTX008 also binds to galectin-3 (Gal-3), albeit more weakly. We identified the contact site for PTX008 on the F-face of the Gal-3 carbohydrate recognition domain. STD NMR revealed that the hydrophobic phenyl ring crown of the calixarene is the binding epitope. With this information, we performed molecular modeling of the complex to assist in improving the rather low affinity of PTX008 for Gal-3. By removing the N-dimethyl alkyl chain amide groups, we produced PTX013 whose reduced alkyl chain length and polar character led to an approximately eightfold stronger binding than PTX008. PTX013 also binds Gal-1 more strongly than PTX008, whereas neither interacts strongly, if at all, with Gal-7. In addition, PTX013, like PTX008, is an allosteric inhibitor of galectin binding to the canonical ligand lactose. This study broadens the scope for galectin targeting by calixarene-based compounds and opens the perspective for selective galectin blocking.

Structure and N-acetylglucosamine binding of the distal domain of mouse adenovirus 2 fibre.

Murine adenovirus 2 (MAdV-2) infects cells of the mouse gastrointestinal tract. Like human adenoviruses, it is a member of the genus Mastadenovirus, family Adenoviridae. The MAdV-2 genome has a single fibre gene that expresses a 787 residue-long protein. Through analogy to other adenovirus fibre proteins, it is expected that the carboxy-terminal virus-distal head domain of the fibre is responsible for binding to the host cell, although the natural receptor is unknown. The putative head domain has little sequence identity to adenovirus fibres of known structure. In this report, we present high-resolution crystal structures of the carboxy-terminal part of the MAdV-2 fibre. The structures reveal a domain with the typical adenovirus fibre head topology and a domain containing two triple ß-spiral repeats of the shaft domain. Through glycan microarray profiling, saturation transfer difference nuclear magnetic resonance spectroscopy, isothermal titration calorimetry and site-directed mutagenesis, we show that the fibre specifically binds to the monosaccharide N-acetylglucosamine (GlcNAc). The crystal structure of the complex reveals that GlcNAc binds between the AB and CD loops at the top of each of the three monomers of the MAdV-2 fibre head. However, infection competition assays show that soluble GlcNAc monosaccharide and natural GlcNAc-containing polymers do not inhibit infection by MAdV-2. Furthermore, site-directed mutation of the GlcNAc-binding residues does not prevent the inhibition of infection by soluble fibre protein. On the other hand, we show that the MAdV-2 fibre protein binds GlcNAc-containing mucin glycans, which suggests that the MAdV-2 fibre protein may play a role in viral mucin penetration in the mouse gut.

Structure and Sialyllactose Binding of the Carboxy-Terminal Head Domain of the Fibre from a Siadenovirus, Turkey Adenovirus 3.

The virulent form of turkey adenovirus 3 (TAdV-3), also known as turkey hemorrhagic enteritis virus (THEV), is an economically important poultry pathogen, while the avirulent form is used as a vaccine. TAdV-3 belongs to the genus Siadenovirus. The carboxy-terminal region of its fibre does not have significant sequence similarity to any other adenovirus fibre heads of known structure. Two amino acid sequence differences between virulent and avirulent TAdV-3 map on the fibre head: where virulent TAdV-3 contains Ile354 and Thr376, avirulent TAdV-3 contains Met354 and Met376. We determined the crystal structures of the trimeric virulent and avirulent TAdV-3 fibre head domains at 2.2 Å resolution. Each monomer contains a beta-sandwich, which, surprisingly, resembles reovirus fibre head more than other adenovirus fibres, although the ABCJ-GHID topology is conserved in all. A beta-hairpin insertion in the C-strand of each trimer subunit embraces its neighbouring monomer. The avirulent and virulent TAdV-3 fibre heads are identical apart from the exact orientation of the beta-hairpin insertion. In vitro, sialyllactose was identified as a ligand by glycan microarray analysis, nuclear magnetic resonance spectroscopy, and crystallography. Its dissociation constant was measured to be in the mM range by isothermal titration calorimetry. The ligand binds to the side of the fibre head, involving amino acids Glu392, Thr419, Val420, Lys421, Asn422, and Gly423 binding to the sialic acid group. It binds slightly more strongly to the avirulent form. We propose that, in vivo, the TAdV-3 fibre may bind a sialic acid-containing cell surface component.

Serine versus threonine glycosylation with α-O-GalNAc: unexpected selectivity in their molecular recognition with lectins.

The molecular recognition of several glycopeptides bearing Tn antigen (α-O-GalNAc-Ser or α-O-GalNAc-Thr) in their structure by three lectins with affinity for this determinant has been analysed. The work yields remarkable results in terms of epitope recognition, showing that the underlying amino acid of Tn (serine or threonine) plays a key role in the molecular recognition. In fact, while Soybean agglutinin and Vicia villosa agglutinin lectins prefer Tn-threonine, Helix pomatia agglutinin shows a higher affinity for the glycopeptides carrying Tn-serine. The different conformational behaviour of the two Tn biological entities, the residues of the studied glycopeptides in the close proximity to the Tn antigen and the topology of the binding site of the lectins are at the origin of these differences.

Peptides derived from human galectin-3 N-terminal tail interact with its carbohydrate recognition domain in a phosphorylation-dependent manner.

Galectin-3 (Gal-3) is a multi-functional effector protein that functions in the cytoplasm and the nucleus, as well as extracellularly following non-classical secretion. Structurally, Gal-3 is unique among galectins with its carbohydrate recognition domain (CRD) attached to a rather long N-terminal tail composed mostly of collagen-like repeats (nine in the human protein) and terminating in a short non-collagenous terminal peptide sequence unique in this lectin family and not yet fully explored. Although several Ser and Tyr sites within the N-terminal tail can be phosphorylated, the physiological significance of this post-translational modification remains unclear. Here, we used a series of synthetic (phospho)peptides derived from the tail to assess phosphorylation-mediated interactions with (15)N-labeled Gal-3 CRD. HSQC-derived chemical shift perturbations revealed selective interactions at the backface of the CRD that were attenuated by phosphorylation of Tyr 107 and Tyr 118, while phosphorylation of Ser 6 and Ser 12 was essential. Controls with sequence scrambling underscored inherent specificity. Our studies shed light on how phosphorylation of the N-terminal tail may impact on Gal-3 function and prompt further studies using phosphorylated full-length protein.