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

General General

The prototype device for non-invasive diagnosis of arteriovenous fistula condition using machine learning methods.

In Scientific reports ; h5-index 158.0

Pattern recognition and automatic decision support methods provide significant advantages in the area of health protection. The aim of this work is to develop a low-cost tool for monitoring arteriovenous fistula (AVF) with the use of phono-angiography method. This article presents a developed and diagnostic device that implements classification algorithms to identify 38 patients with end stage renal disease, chronically hemodialysed using an AVF, at risk of vascular access stenosis. We report on the design, fabrication, and preliminary testing of a prototype device for non-invasive diagnosis which is very important for hemodialysed patients. The system includes three sub-modules: AVF signal acquisition, information processing and classification and a unit for presenting results. This is a non-invasive and inexpensive procedure for evaluating the sound pattern of bruit produced by AVF. With a special kind of head which has a greater sensitivity than conventional stethoscope, a sound signal from fistula was recorded. The proces of signal acquisition was performed by a dedicated software, written specifically for the purpose of our study. From the obtained phono-angiogram, 23 features were isolated for vectors used in a decision-making algorithm, including 6 features based on the waveform of time domain, and 17 features based on the frequency spectrum. Final definition of the feature vector composition was obtained by using several selection methods: the feature-class correlation, forward search, Principal Component Analysis and Joined-Pairs method. The supervised machine learning technique was then applied to develop the best classification model.

Grochowina Marcin, Leniowska Lucyna, Gala-Błądzińska Agnieszka

2020-Oct-02

General General

Question-driven summarization of answers to consumer health questions.

In Scientific data

Automatic summarization of natural language is a widely studied area in computer science, one that is broadly applicable to anyone who needs to understand large quantities of information. In the medical domain, automatic summarization has the potential to make health information more accessible to people without medical expertise. However, to evaluate the quality of summaries generated by summarization algorithms, researchers first require gold standard, human generated summaries. Unfortunately there is no available data for the purpose of assessing summaries that help consumers of health information answer their questions. To address this issue, we present the MEDIQA-Answer Summarization dataset, the first dataset designed for question-driven, consumer-focused summarization. It contains 156 health questions asked by consumers, answers to these questions, and manually generated summaries of these answers. The dataset's unique structure allows it to be used for at least eight different types of summarization evaluations. We also benchmark the performance of baseline and state-of-the-art deep learning approaches on the dataset, demonstrating how it can be used to evaluate automatically generated summaries.

Savery Max, Abacha Asma Ben, Gayen Soumya, Demner-Fushman Dina

2020-Oct-02

General General

VR and machine learning: novel pathways in surgical hands-on training.

In Current opinion in urology

PURPOSE OF REVIEW : Surgical training has dramatically changed over the last decade. It has become not only the way to prepare surgeons for their everyday work, but also a way to certify their skills thus increasing patient safety. This article reviews advances in the use of machine learning and artificial intelligence applied to virtual reality based surgical training over the last 5 years.

RECENT FINDINGS : Eight articles have been published which met the inclusion criteria. This included six articles about the use of machine learning and artificial intelligence for assessment purposes and two articles about the possibility of teaching applications, including one review and one original research article. All the research articles pointed out the importance of machine learning and artificial intelligence for the stratification of trainees, based on their performance on basic tasks or procedures simulated in a virtual reality environment.

SUMMARY : Machine learning and artificial intelligence are designed to analyse data and use them to take decisions that typically require human intelligence. Evidence in literature is still scarce about this technology applied to virtual reality and existing manuscripts are mainly focused on its potential to stratify surgical performance and provide synthetic feedbacks about it. In consideration of the exponential growth of computer calculation capabilities, it is possible to expect a parallel increase of research about this topic within the next few years.

Veneziano Domenico, Cacciamani Giovanni, Rivas Juan Gomez, Marino Nicola, Somani Bhaskar K

2020-Nov

Cardiology Cardiology

Big data and new information technology: what cardiologists need to know.

In Revista espanola de cardiologia (English ed.)

Technological progress in medicine is constantly garnering pace, requiring that physicians constantly update their knowledge. The new wave of technologies breaking through into clinical practice includes the following: a) mHealth, which allows constant monitoring of biological parameters, anytime, anyplace, of hundreds of patients at the same time; b) artificial intelligence, which, powered by new deep learning techniques, are starting to beat human experts at their own game: diagnosis by imaging or electrocardiography; c) 3-dimensional printing, which may lead to patient-specific prostheses; d) systems medicine, which has arisen from big data, and which will open the way to personalized medicine by bringing together genetic, epigenetic, environmental, clinical and social data into complex integral mathematical models to design highly personalized therapies. This state-of-the-art review aims to summarize in a single document the most recent and most important technological trends that are being applied to cardiology, and to provide and overall view that will allow readers to discern at a glance the direction of cardiology in the next few years.

Baladrón Carlos, Gómez de Diego José Juan, Amat-Santos Ignacio J

2020-Sep-29

Artificial intelligence, Big data, Inteligencia artificial, Medicina de sistemas, Systems medicine, mHealth

Radiology Radiology

Effectiveness of COVID-19 diagnosis and management tools: A review.

In Radiography (London, England : 1995)

OBJECTIVE : To review the available literature concerning the effectiveness of the COVID-19 diagnostic tools.

BACKGROUND : With the absence of specific treatment/vaccines for the coronavirus COVID-19, the most appropriate approach to control this infection is to quarantine people and isolate symptomatic people and suspected or infected cases. Although real-time reverse transcription-polymerase chain reaction (RT-PCR) assay is considered the first tool to make a definitive diagnosis of COVID-19 disease, the high false negative rate, low sensitivity, limited supplies and strict requirements for laboratory settings might delay accurate diagnosis. Computed tomography (CT) has been reported as an important tool to identify and investigate suspected patients with COVID-19 disease at early stage.

KEY FINDINGS : RT-PCR shows low sensitivity (60-71%) in diagnosing patients with COVID-19 infection compared to the CT chest. Several studies reported that chest CT scans show typical imaging features in all patients with COVID-19. This high sensitivity and initial presentation in CT chest can be helpful in rectifying false negative results obtained from RT-PCR. As COVID-19 has similar manifestations to other pneumonia diseases, artificial intelligence (AI) might help radiologists to differentiate COVID-19 from other pneumonia diseases.

CONCLUSION : Although CT scan is a powerful tool in COVID-19 diagnosis, it is not sufficient to detect COVID-19 alone due to the low specificity (25%), and challenges that radiologists might face in differentiating COVID-19 from other viral pneumonia on chest CT scans. AI might help radiologists to differentiate COVID-19 from other pneumonia diseases.

IMPLICATION FOR PRACTICE : Both RT-PCR and CT tests together would increase sensitivity and improve quarantine efficacy, an impact neither could achieve alone.

Alsharif W, Qurashi A

2020-Sep-21

Artificial intelligence, CT scan, Consolidation, Crazy-paving, Ground-glass opacification, RT-PCR

General General

Practicing precision medicine with intelligently integrative clinical and multi-omics data analysis.

In Human genomics

Precision medicine aims to empower clinicians to predict the most appropriate course of action for patients with complex diseases like cancer, diabetes, cardiomyopathy, and COVID-19. With a progressive interpretation of the clinical, molecular, and genomic factors at play in diseases, more effective and personalized medical treatments are anticipated for many disorders. Understanding patient's metabolomics and genetic make-up in conjunction with clinical data will significantly lead to determining predisposition, diagnostic, prognostic, and predictive biomarkers and paths ultimately providing optimal and personalized care for diverse, and targeted chronic and acute diseases. In clinical settings, we need to timely model clinical and multi-omics data to find statistical patterns across millions of features to identify underlying biologic pathways, modifiable risk factors, and actionable information that support early detection and prevention of complex disorders, and development of new therapies for better patient care. It is important to calculate quantitative phenotype measurements, evaluate variants in unique genes and interpret using ACMG guidelines, find frequency of pathogenic and likely pathogenic variants without disease indicators, and observe autosomal recessive carriers with a phenotype manifestation in metabolome. Next, ensuring security to reconcile noise, we need to build and train machine-learning prognostic models to meaningfully process multisource heterogeneous data to identify high-risk rare variants and make medically relevant predictions. The goal, today, is to facilitate implementation of mainstream precision medicine to improve the traditional symptom-driven practice of medicine, and allow earlier interventions using predictive diagnostics and tailoring better-personalized treatments. We strongly recommend automated implementation of cutting-edge technologies, utilizing machine learning (ML) and artificial intelligence (AI) approaches for the multimodal data aggregation, multifactor examination, development of knowledgebase of clinical predictors for decision support, and best strategies for dealing with relevant ethical issues.

Ahmed Zeeshan

2020-Oct-02

Artificial intelligence, Clinics, Genomics, Integrative analysis, Machine learning, Metabolomics, Precision medicine