Health Life

AI detects COVID-19 on chest X-rays with accuracy and speed

Generated heatmaps appropriately highlighted abnormalities in the lung fields in those images accurately labeled as COVID-19 positive (A-C) in contrast to images which were accurately labeled as negative for COVID-19 (D). Intensity of colors on the heatmap correspond to features of the image that are important for prediction of COVID-19 positivity. Credit: Northwestern University

Northwestern University researchers have developed a new artificial intelligence (A.I.) platform that detects COVID-19 by analyzing X-ray images of the lungs.

Called DeepCOVID-XR, the outperformed a team of specialized thoracic radiologists—spotting COVID-19 in X-rays about 10 times faster and 1-6% more accurately.

The researchers believe physicians could use the A.I. system to rapidly screen patients who are admitted into hospitals for reasons other than COVID-19. Faster, earlier detection of the highly contagious virus could potentially protect health care workers and other patients by triggering the positive patient to isolate sooner.

The study’s authors also believe the algorithm could potentially flag patients for isolation and testing who are not otherwise under investigation for COVID-19.

The study will be published on Nov. 24 in the journal Radiology.

“We are not aiming to replace actual testing,” said Northwestern’s Aggelos Katsaggelos, an A.I. expert and senior author of the study. “X-rays are routine, safe and inexpensive. It would take seconds for our system to screen a patient and determine if that patient needs to be isolated.”

“It could take hours or days to receive results from a COVID-19 test,” said Dr. Ramsey Wehbe, a cardiologist and postdoctoral fellow in A.I. at the Northwestern Medicine Bluhm Cardiovascular Institute. “A.I. doesn’t confirm whether or not someone has the virus. But if we can flag a patient with this algorithm, we could speed up triage before the test results come back.”

Katsaggelos is the Joseph Cummings Professor of Electrical and

Health Life

Areas where the next pandemic could emerge are revealed

Illustrative map of ‘red-alert’ zone. Circles represent approximate location of risk; circle size indicates level of risk. Credit: Michael Walsh, University of Sydney

An international team of human- and animal-health experts has incorporated environmental, social and economic considerations—including air transit centrality—to identify key areas at risk of leading to the next pandemic.

An international team of researchers has taken a holistic approach to reveal for the first time where wildlife-human interfaces intersect with areas of poor human outcomes and highly globalized cities, which could give rise to the next pandemic unless preventative measures are taken.

Areas exhibiting a high degree of human pressure on wildlife also had more than 40 percent of the world’s most connected cities in or adjacent to areas of likely spillover, and 14-20 percent of the world’s most connected cities at risk of such spillovers likely to go undetected because of poor health infrastructure (predominantly in South and South East Asia and Sub-Saharan Africa). As with COVID-19, the impact of such spillovers could be global.

Led by the University of Sydney and with academics spanning the United Kingdom, India and Ethiopia, the open-access paper shows the cities worldwide that are at risk. Last month, an IPBES report highlighted the role biodiversity destruction plays in pandemics and provided recommendations. This Sydney-led research pinpoints the geographical areas that require greatest attention.

The paper, “Whence the next pandemic? The intersecting global geography of the animal-human interface, poor health systems and air transit centrality reveals conduits for high-impact spillover”, has published in the leading Elsevier journal, One Health. City lists for yellow, orange and red alert zones are available in open access.

Lead author Dr. Michael Walsh, who co-leads the One Health Node at Sydney’s Marie Bashir Institute for Infectious Diseases and Biosecurity, said that previously, much has been

Health Life

Caltech’s AI-driven COVID-19 model routinely outperforms competitors

Credit: CC0 Public Domain

A new model for predicting COVID-19’s impact using artificial intelligence (AI) dramatically outperforms other models, so much so that it has attracted the interest of public health officials across the country.

While existing models to predict the spread of a disease already exist, few, if any, incorporate AI, which allows a to make predictions based on observations of what is actually happening—for example, increasing cases among specific populations—as opposed to what the model’s designers think will happen. With the use of AI, it is possible to discover patterns hidden in data that humans alone might not recognize.

“AI is a powerful tool, so it only makes sense to apply it to one of the most urgent problems the world faces,” says Yaser Abu-Mostafa (Ph.D. ’83), professor of electrical engineering and computer science, who led the development of the new CS156 model (so-named for the Caltech computer science class where it got its start).

The researchers evaluate the accuracy of the model by comparing it to the predictions of an ensemble model built by the Centers for Disease Control and Prevention from 45 major models from universities and institutes across the country. Using 1,500 predictions as points of comparison with the CDC ensemble, the researchers found that the CS156 model was more accurate than the ensemble model 58 percent of the time as of November 25. It also routinely outperforms the benchmark projections of the Institute for Health Metrics and Evaluation (IHME).

The CS156 model is an amalgam of several models developed by Abu-Mostafa and his group over the past nine months; the weight that each model has on the overall CS156 model’s output is determined based on its performance over the previous weeks (using statistics reported in The New York Times for comparison).

Abu-Mostafa is

Health Life

Immune strategy based on limited information in the network

Schematic illustration of our limited knowledge immunization strategy. Credit: ©Science China Press

The novel coronavirus outbreak is a global pandemic that has spread to more than 200 countries and territories around the world. Currently, countries and territories are fighting the spread of the disease using social distancing such as quarantine, testing or isolation which can be regarded as immunization.

Given the important role that networks play in disease spreading, much effort has been made to understand and develop efficient methods such as targeted immunization. Previous models have typically assumed full of the network structure and immunized the most central nodes (see left panel of the below figure). However, in real-world scenarios, knowledge and observations of the full social network is usually limited thereby precluding a full assessment of the optimal (most central) nodes to test or quarantine or immunize that will efficiently stop the spread of pandemic.

Here, they present and study a novel and efficient immunization strategy incorporating the realistic case that we have only limited observability of the . One can assume that only n nodes can be observed at a given time and that the most central of these n is immunized. This could represent a case where separate teams are sent to immunize or quarantine individuals. Each team examines n individuals and immunizes or quarantines the most connected of these n nodes.

The authors find, both analytically and via simulations, that as n increases, to even moderate numbers (approximately 10), the percolation threshold increases significantly towards its optimal value for i.e., towards percolation with full information. Larger values of percolation thresholds imply greater efficiency since a lower fraction of nodes can be immunized to stop the epidemic. They develop a general analytical framework for this approach of information on only n nodes, and determine