Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

General General

A Deep Learning Approach for Human Action Recognition Using Skeletal Information.

In Advances in experimental medicine and biology

In this paper we present an approach toward human action detection for activities of daily living (ADLs) that uses a convolutional neural network (CNN). The network is trained on discrete Fourier transform (DFT) images that result from raw sensor readings, i.e., each human action is ultimately described by an image. More specifically, we work using 3D skeletal positions of human joints, which originate from processing of raw RGB sequences enhanced by depth information. The motion of each joint may be described by a combination of three 1D signals, representing its coefficients into a 3D Euclidean space. All such signals from a set of human joints are concatenated to form an image, which is then transformed by DFT and is used for training and evaluation of a CNN. We evaluate our approach using a publicly available challenging dataset of human actions that may involve one or more body parts simultaneously and for two sets of actions which resemble common ADLs.

Mathe Eirini, Maniatis Apostolos, Spyrou Evaggelos, Mylonas Phivos

2020

Activities of daily living, Convolutional neural networks, Human action recognition

General General

Multivariate Data Analysis and Machine Learning for Prediction of MCI-to-AD Conversion.

In Advances in experimental medicine and biology

There has always been a need for discovering efficient and dependable Alzheimer's disease (AD) diagnostic biomarkers. Like the majority of diseases, the earlier the diagnosis, the most effective the treatment. (Semi)-automated structural magnetic resonance imaging (MRI) processing approaches are very popular in AD research. Mild cognitive impairment (MCI) is considered to be a stage between normal cognitive ageing and dementia. MCI can often be the prodromal stage of AD. Around 10-15% of MCI patients convert to AD per year. In this study, we used three supervised machine learning (ML) techniques to differentiate MCI converters (MCIc) from MCI non-converters (MCInc) and predict their conversion rates from baseline MRI data (cortical thickness (CTH) and hippocampal volume (HCV)). A total of 803 participants from the ADNI cohort were included in this study (188 AD, 107 MCIc, 257 MCInc and 156 healthy controls (HC)). We studied the classification abilities of three different WEKA classifiers (support vector machine (SVM), decision trees (J48) and Naive Bayes (NB)). We built six different classification models, three models based on CTH and three based on HCV (CTH-SVM, CTH-J48, CTH-NB, HCV-SVM, HCV-J48 and HCV-NB). For the classification experiments, we obtained up to 71% sensitivity and up to 56% specificity. The prediction of conversion showed accuracy for up to 84%. The value of certain multivariate models derived from the classification experiments has exhibited robust and effective results in MCIc identification. However, there was a limitation in this study since we could not compare the CTH with the HCV models seeing as the data used originated from different subjects. As future direction, we propose the creation of a model that would combine various features with data originating from the same subjects, thus being a far more reliable and accurate prognostic tool.

Skolariki Konstantina, Terrera Graciella Muniz, Danso Samuel

2020

Alzheimer’s biomarkers, Decision trees, MCI-to-AD progression, Machine learning, Naive Bayes, Support vector machine

General General

Neuroeducation and Computer Programming: A Review.

In Advances in experimental medicine and biology

Over the past 5 years, a significant number of studies focused on computer programming and code writing (software development, code comprehension, program debugging, code optimization, developer training), using the capabilities of brain imaging techniques and of biomarkers. With the use of the aforementioned techniques, researchers have explored the role of programming experience and knowledge, the relation between coding and writing, and the possibilities of improving program debugging with machine learning techniques. In this paper, a review of existing literature and discussion of research issues that should be examined in the future are explored. Research may link the neuroscientific field with training issues in programming, so as to contribute to the learning process.

Giannopoulou Panagiota, Papalaskari Mary-Angela, Doukakis Spyridon

2020

Computer programming education, Mobile applications, Neuroeducation, Training

Surgery Surgery

Antibody Clustering Using a Machine Learning Pipeline that Fuses Genetic, Structural, and Physicochemical Properties.

In Advances in experimental medicine and biology

Antibody V domain clustering is of paramount importance to a repertoire of immunology-related areas. Although several approaches have been proposed for antibody clustering, still no consensus has been reached. Numerous attempts use information from genes, protein sequences, 3D structures, and 3D surfaces in an effort to elucidate unknown action mechanisms directly related to their function and to either link them directly to diseases or drive the discovery of new medicines, such as antibody drug conjugates (ADC). Herein, we describe a new V domain antibody clustering method based on the comparison of the interaction sites between each antibody and its antigen. A more specific clustering analysis of the antibody's V domain was provided using deep learning and data mining techniques. The multidimensional information was extracted from the structural resolved antibodies when they were captured to interact with other proteins. The available 3D structures of protein antigen-antibody (Ag-Ab) interfaces contain information about how antibody V domains recognize antigens as well as about which amino acids are involved in the recognition. As such, the antibody surface holds information about antigens' folding that reside with the Ab-Ag interface residues and how they interact. In order to gain insight into the nature of such interactions, we propose a new simple philosophy to transform the conserved framework (fragment regions, complementarity-determining regions) of antibody V domain in a binary form using structural features of antibody-antigen interactions, toward identifying new antibody signatures in V domain binding activity. Finally, an advanced three-level hybrid classification scheme has been set for clustering antibodies in subgroups, which can combine the information from the protein sequences, the three-dimensional structures, and specific "key patterns" of recognized interactions. The clusters provide multilevel information about antibodies and antibody-antigen complexes.

Papageorgiou Louis, Maroulis Dimitris, Chrousos George P, Eliopoulos Elias, Vlachakis Dimitrios

2020

Antibodies, Antibody drug conjugates, Antibody-antigen complexes, Classification scheme, Clustering, Immunology

Pathology Pathology

Whole Slide Imaging (WSI) in Pathology: Current Perspectives and Future Directions.

In Journal of digital imaging

Whole slide imaging (WSI), ever since its first introduction about two decades ago, has been validated for a number of applications in the field of pathology. The recent approval of US FDA to a WSI system for use in primary surgical pathology diagnosis has opened avenues for wider acceptance and application of this technology in routine practice. The ongoing technological advances in digital scanners, image visualization methods, and the integration of artificial intelligence-derived algorithms with these systems provide opportunities of its newer applications. Its benefits are innumerable such as ease of access through internet, avoidance of physical storage space, and no risk of deterioration of staining quality or breakage of slides to name a few. Various barriers such as the high cost, technical glitches, and professional hesitation to adopt a new technology have hindered its use in pathology. This review article summarizes the technical aspects of WSI, its applications in diagnostic pathology, training, and research along with future perspectives. It highlights the benefits, limitations, and challenges delaying the use of this technology in routine practice. The review is targeted at students, residents, and budding pathologists to better acquaint them with the key aspects of state-of-the-art technology and enable them to implement WSI judiciously.

Kumar Neeta, Gupta Ruchika, Gupta Sanjay

2020-May-28

Automated image analysis, Cytopathology, Diagnosis, Education, Regulation, Telepathology, Validation, Whole slide imaging

Radiology Radiology

An Open-Source, Vender Agnostic Hardware and Software Pipeline for Integration of Artificial Intelligence in Radiology Workflow.

In Journal of digital imaging

Although machine learning (ML) has made significant improvements in radiology, few algorithms have been integrated into clinical radiology workflow. Complex radiology IT environments and Picture Archiving and Communication System (PACS) pose unique challenges in creating a practical ML schema. However, clinical integration and testing are critical to ensuring the safety and accuracy of ML algorithms. This study aims to propose, develop, and demonstrate a simple, efficient, and understandable hardware and software system for integrating ML models into the standard radiology workflow and PACS that can serve as a framework for testing ML algorithms. A Digital Imaging and Communications in Medicine/Graphics Processing Unit (DICOM/GPU) server and software pipeline was established at a metropolitan county hospital intranet to demonstrate clinical integration of ML algorithms in radiology. A clinical ML integration schema, agnostic to the hospital IT system and specific ML models/frameworks, was implemented and tested with a breast density classification algorithm and prospectively evaluated for time delays using 100 digital 2D mammograms. An open-source clinical ML integration schema was successfully implemented and demonstrated. This schema allows for simple uploading of custom ML models. With the proposed setup, the ML pipeline took an average of 26.52 s per second to process a batch of 100 studies. The most significant processing time delays were noted in model load and study stability times. The code is made available at "http://bit.ly/2Z121hX". We demonstrated the feasibility to deploy and utilize ML models in radiology without disrupting existing radiology workflow.

Sohn Jae Ho, Chillakuru Yeshwant Reddy, Lee Stanley, Lee Amie Y, Kelil Tatiana, Hess Christopher Paul, Seo Youngho, Vu Thienkhai, Joe Bonnie N

2020-May-28

Artificial intelligence, Informatics, Machine learning, PACS, Quality improvement