Goto

Collaborating Authors

Results


NHS Introduces A New AI-Based Technology That Can Detect Heart Disease At Record Speed And With 40 Percent Higher Accuracy

#artificialintelligence

The NHS is now employing a cutting-edge AI program that can diagnose heart illness in just 20 SECONDS. While the patient is in the scanner, the computer tool, which resembles human ability but with more precision and speed, can analyze cardiac MRI data in 20 seconds. According to the British Heart Foundation (BHF), which has supported research into the technology, this is significantly faster than a doctor physically examining the pictures following an MRI scan, which may take up to 13 minutes. The technology identifies heart structure and function changes with 40% greater accuracy and retrieves 40% more information than a human can. According to the new research, the approach was more accurate at analyzing MRIs than the work of three specialists.


NHS rolls out AI tool which detects heart disease in 20 seconds

#artificialintelligence

The NHS has rolled out a new artificial intelligence (AI) tool which can detect heart disease in just 20 seconds while patients are in an MRI scanner. A British Heart Foundation (BHF) funded study published in the Journal of Cardiovascular Magnetic Resonance concluded the machine analysis had superior precision to three clinicians. It would usually take a doctor 13 minutes or more to manually analyse images after an MRI scan has been performed. The technology is being used on more than 140 patients a week at University College London (UCL) Hospital, Barts Heart Centre at St Bartholomew's Hospital, and Royal Free Hospital. Later this year it will be introduced to a further 40 locations across the UK and globally.


UCL: AI heart disease detector begins NHS roll-out

#artificialintelligence

The first-of-its-kind AI tool, described in a new paper in the Journal of Cardiovascular Magnetic Resonance, analyses heart MRI scans in just 20 seconds whilst the patient is in the scanner. This compares to the 13 minutes or more it would take for a doctor to manually analyse the images after the MRI scan has been performed. Each year, around 120,000 heart MRI scans are performed in the UK. The researchers say that the AI will free-up valuable time of healthcare professionals – saving around 3,000 clinician days every year – so their attention can be directed to seeing more patients on NHS waiting lists, which will ultimately help with the backlog in vital heart care. The AI will also give patients and doctors more confidence in the results so that they can make better decisions about a person's treatment and possible surgeries.


New artificial intelligence tool 'can detect heart disease at record speed'

#artificialintelligence

A new artificial intelligence (AI) tool being used in the NHS can detect heart disease at record speed, experts say. The computer tool, which mimics human ability but with greater precision and at a faster speed, can analyse heart MRI scans in just 20 seconds while the patient is in the scanner. This is much quicker than the 13 minutes or more it would take for a doctor to manually examine the images after an MRI scan, according to the British Heart Foundation (BHF), which has funded research into the tool. The technique also detects changes to the heart structure and function with 40% higher accuracy and extracts more information than a human can, the BHF said. A new study concluded the technique was more precise at analysing MRIs than the work of three specialist doctors.


Maintaining fairness across distribution shift: do we have viable solutions for real-world applications?

arXiv.org Machine Learning

Fairness and robustness are often considered as orthogonal dimensions when evaluating machine learning models. However, recent work has revealed interactions between fairness and robustness, showing that fairness properties are not necessarily maintained under distribution shift. In healthcare settings, this can result in e.g. a model that performs fairly according to a selected metric in "hospital A" showing unfairness when deployed in "hospital B". While a nascent field has emerged to develop provable fair and robust models, it typically relies on strong assumptions about the shift, limiting its impact for real-world applications. In this work, we explore the settings in which recently proposed mitigation strategies are applicable by referring to a causal framing. Using examples of predictive models in dermatology and electronic health records, we show that real-world applications are complex and often invalidate the assumptions of such methods. Our work hence highlights technical, practical, and engineering gaps that prevent the development of robustly fair machine learning models for real-world applications. Finally, we discuss potential remedies at each step of the machine learning pipeline.


Scientists publish a blueprint to apply artificial intelligence to extend human longevity

#artificialintelligence

In the article the authors describe a new field of study converging AI, basic research, and medicine referred to as Longevity Medicine. Another definition for Longevity Medicine is the preventative and restorative medicine enabled by the deep aging clocks and artificial intelligence. The article was authored by Alex Zhavoronkov, the founder and chief longevity officer of Deep Longevity, a computer scientist with a PhD in biophysics, Evelyne Yehudit Bischof, a practicing medical doctor trained in the top European and the US medical schools actively engaged in aging research and gerooncology at the University Hospital Basel in Switzerland, and at Shanghai University of Medicine and Health Sciences, and one of the most prolific scientists and entrepreneurs in artificial intelligence, Kai-Fu Lee. The traditional approach to medicine is to treat diseases. However, scientists estimate (Cutler and Mattson, 2006) that complete elimination of cancer would result in only 2.3 year increase in life expectancy in the US at birth and 1.3 year gain at age 65.


New NHS imaging resource assists AI in Covid fight

#artificialintelligence

Published today in the Open-Access, Open-Data journal GigaScience is the National COVID-19 Chest Imaging Database (NCCID), a centralised database containing chest X-rays, Computed Tomography (CT) and MRI scans from patients across the UK. Utilising the unique position as the world's single largest integrated healthcare system, the benefits of collecting chest imaging data this large are extensive and already being used by doctors and the research community. The database is already supporting the development of Artificial Intelligence (AI)-powered image processing software and diagnostic products and models being used to predict COVID-19 mortality in the UK. And also has the potential to become a long-term resource for teaching radiologists. These efforts provide the potential to enable faster patient assessment in Accident and Emergency, save clinicians time, and increase the safety and consistency of care across the UK. With the GigaScience paper describing how to access this Open Data resource, the NCCID training data is available to users anywhere in the world, including software developers, academics and clinicians, via a rigorous Data Access Request process.


Artificial intelligence successfully predicts protein interactions

#artificialintelligence

DALLAS – Nov. 16, 2021 – UT Southwestern and University of Washington researchers led an international team that used artificial intelligence (AI) and evolutionary analysis to produce 3D models of eukaryotic protein interactions. The study, published in Science, identified more than 100 probable protein complexes for the first time and provided structural models for more than 700 previously uncharacterized ones. Insights into the ways pairs or groups of proteins fit together to carry out cellular processes could lead to a wealth of new drug targets. "Our results represent a significant advance in the new era in structural biology in which computation plays a fundamental role," said Qian Cong, Ph.D., Assistant Professor in the Eugene McDermott Center for Human Growth and Development with a secondary appointment in Biophysics. Dr. Cong led the study with David Baker, Ph.D., Professor of Biochemistry and Dr. Cong's postdoctoral mentor at the University of Washington prior to her recruitment to UT Southwestern.


MURAL: An Unsupervised Random Forest-Based Embedding for Electronic Health Record Data

arXiv.org Artificial Intelligence

A major challenge in embedding or visualizing clinical patient data is the heterogeneity of variable types including continuous lab values, categorical diagnostic codes, as well as missing or incomplete data. In particular, in EHR data, some variables are {\em missing not at random (MNAR)} but deliberately not collected and thus are a source of information. For example, lab tests may be deemed necessary for some patients on the basis of suspected diagnosis, but not for others. Here we present the MURAL forest -- an unsupervised random forest for representing data with disparate variable types (e.g., categorical, continuous, MNAR). MURAL forests consist of a set of decision trees where node-splitting variables are chosen at random, such that the marginal entropy of all other variables is minimized by the split. This allows us to also split on MNAR variables and discrete variables in a way that is consistent with the continuous variables. The end goal is to learn the MURAL embedding of patients using average tree distances between those patients. These distances can be fed to nonlinear dimensionality reduction method like PHATE to derive visualizable embeddings. While such methods are ubiquitous in continuous-valued datasets (like single cell RNA-sequencing) they have not been used extensively in mixed variable data. We showcase the use of our method on one artificial and two clinical datasets. We show that using our approach, we can visualize and classify data more accurately than competing approaches. Finally, we show that MURAL can also be used to compare cohorts of patients via the recently proposed tree-sliced Wasserstein distances.


Artificial intelligence successfully predicts protein interactions

#artificialintelligence

An international team led by researchers at UT Southwestern and the University of Washington predicted the structures using artificial intelligence techniques. UT Southwestern and University of Washington researchers led an international team that used artificial intelligence (AI) and evolutionary analysis to produce 3D models of eukaryotic protein interactions. The study, published in Science, identified more than 100 probable protein complexes for the first time and provided structural models for more than 700 previously uncharacterized ones. Insights into the ways pairs or groups of proteins fit together to carry out cellular processes could lead to a wealth of new drug targets. "Our results represent a significant advance in the new era in structural biology in which computation plays a fundamental role," said Qian Cong, Ph.D., Assistant Professor in the Eugene McDermott Center for Human Growth and Development with a secondary appointment in Biophysics.