Geisinger develops rECHOmmend model to predict undiagnosed structural heart disease

Published in Circulation, the rECHOmmend study expands on AI-focused research to improve patient outcomes in cardiology

A team of clinicians and scientists from Tempus and Geisinger have found that a new artificial intelligence model can accurately identify patients at increased risk of undiagnosed structural heart disease. Source: Getty Images

Structural heart disease (SHD) is a group of conditions that adversely affect the valves, walls, chambers, or muscles of the heart. SHD is typically a progressive disease that causes a variety of debilitating symptoms or death, making it important to diagnose and treat patients early to prevent these poor outcomes. However, many patients with the disease are undiagnosed.

The Tempus and Geisinger study sought to address this diagnostic gap by developing a novel machine learning model that uses data from a 12-lead electrocardiogram (ECG)—an inexpensive and commonly used test measuring the electrical signals of the heart—to identify patients at high risk for of undiagnosed SHD. Published in Circulation, the rECHOmmend model can predict any one of seven structural heart diseases that are diagnosable by echocardiography (an ultrasound of the heart).

The team of data scientists and medical researchers used 2.2 million ECGs from more than 480,000 patients over 37 years of patient care at Geisinger to train a deep neural network—a specialized type of AI model—to predict who, among patients without a prior history of SHD, would develop a clinically significant disease that could benefit from guideline-directed monitoring or treatment. Overall, the study found that the model achieved excellent performance, exceeding the performance of any previously published model predicting any single disease. The findings show that clinicians using this model could find more diseases with fewer diagnostic studies.

“Structural heart disease carries a high burden of morbidity and mortality, and this model is both actionable and practical for identifying undiagnosed patients in clinical practice,” said Joel Dudley, Ph.D., chief scientific officer at Tempus. “Our two teams are continuing to find new ways of applying AI to predict heart disease before it reaches a severe stage of irreversible debilitation for patients, and the rECHOmmend study builds on that foundational work.”

“Past studies have shown the ability of artificial intelligence to enable single disease screening with echocardiography. The rECHOmmend study builds on those to further improve the feasibility of echocardiography as a screening tool for structural heart disease,” said Alvaro Ulloa Cerna, Ph.D., senior data scientist at Geisinger and a lead author of the study. “This could allow for earlier diagnosis and potentially avoid further disease development and its debilitating symptoms.”

This study expands the AI-based cardiology research the Tempus and Geisinger teams have pursued in recent years demonstrating that AI can predict mortality directly from ECG data even in the large subset of ECGs interpreted by physicians as normal. In 2021, a jointly created AI model that can predict the risk of new atrial fibrillation (AF) and AF-related stroke was published in Circulation and was later granted Breakthrough Device Designation by the U.S. Food & Drug Administration.

Nixu continues developing cloud security for Finnish government’s Valtori

The Finnish Government ICT Centre Valtori started its Cloud program in 2019 to enable around 100 Finnish public sector organizations to securely utilize cloud environments' capabilities. Therefore, Valtori needed to ensure that their cloud security is on par with the legal requirements and national information security criteria PiTuKri. Nixu helped in realizing this goal by delivering a governance model and Cloud Security Posture Management (CSPM) framework for Valtori’s cloud services. Nixu will continue to develop Valtori’s Cloud Security also in the future, as the contract has now been extended until 2025. a4e06517bace3b0c 800x800ar 36086

The challenge for central government actors in using cloud platforms has been the information security aspect of cloud-based services. The security concerns meant that for a long time, the rule of thumb was to save only public information in the cloud. However, the tides turned in 2019 when the Finnish Ministry of Finance published new guidelines for public sector organizations on how to use cloud services. By the end of 2020, the same ministry already considered the cloud an equal or even a preferred alternative to traditional data centers. 

Valtori provides sector-independent ICT services for the central government as well as information and data communications technology services that meet the requirements of high preparedness and security. Valtori serves a client base that comprises around 100 government agencies and departments with tens of thousands of service users. Therefore, Valtori has a large responsibility for providing secure ICT services to its clients. 

Need for a secure governance model for cloud environments

One of the services that Valtori offers to governmental organizations is a governance model for cloud environments. A critical aspect of this service is ensuring that its information security corresponds with legal requirements and the Criteria for Assessing the Information Security of Cloud Services (PiTuKri). PiTuKri is published by the Finnish Transport and Communications Agency's National Cyber Security Center, NCSC. Implementing its criteria improves security in situations where authorities process classified information in the cloud. Consequently, it affects Valtori and all of its clients. 

To offer a secure governance model, Valtori needed to find a service provider that could master the technical execution, that is, define the relevant security controls for measuring security posture, which would also match the PiTuKri criteria. Due to Nixu’s proven Cloud Security expertise and prior experience with similar projects, Valtori chose Nixu as the service provider at the beginning of 2020. In addition, Nixu delivered documented instructions for implementing Cloud Security Posture Management (CSPM) for Amazon AWS and Microsoft Azure cloud environments and guidelines for further development. 

Nixu also produced data protection guidelines for Valtori’s cloud platforms by the PiTuKri criteria. “Valtori had a vision of Privacy by design (Data protection by design and default), meaning that privacy is considered in the project right from the beginning. This is ideal for a privacy specialist, and working with Valtori’s multidisciplinary team to implement this has been smooth and rewarding,” commends Nixu’s Privacy Specialist Tuisku Sarrala.

Successful pioneer work results in real-time visibility and continuous compliance

After two years of collaboration with Nixu, Valtori can now offer its public sector clients an information security service that enables the users to have real-time visibility on the status of their cloud security controls. The controls follow the defined framework and ensure that the cloud platforms continuously comply with the PiTuKri criteria. Around 75% of Valtori’s clients currently use cloud services, and the service package is installed into all AWS and Azure accounts. This makes life easier – and more secure – for the end-users.

“Our ability to offer validated security controls to our clients advances the use of cloud within the public sector because it encourages cautious decision-makers to trust cloud services and start building cloud environments within their organizations. They can focus on their core work and rely on the fact that if their cloud security controls are not up to date, this security component will notify them to make the needed fixes. Our clients can have peace of mind from a compliance point of view,” states Juha Nieminen, Development Manager at Valtori.

The partnership has been constructive and, with all the tackled challenges, also educational for both sides. “Working with Valtori has been smooth from day one. They had a well-thought vision which we started to work towards together in a very collaborative manner. You can see clearly how much emphasis is put on cybersecurity at Valtori, which makes the work meaningful for everyone involved”, praises Sakari Pihlhjerta, Business Unit Lead for Cloud Security at Nixu.

“I appreciate fluency, flexibility, and strong expertise because those elements ensure that the work gets done. Nixu’s team has delivered us that special know-how we have longed for, and I don’t think there are many other companies in Finland we could have executed this project with”, Nieminen concludes. “We were pioneers who, through iteration, worked to accomplish something that had never been done before. The information security solution we’ve created with Nixu has been one of the biggest wins within our Cloud program.”

OU meteorologist's cloud research recognized by National Science Foundation

Most people associate clouds with precipitation like rain or snow, but clouds influence many atmospheric processes. Cloud formation is affected by turbulence, which creates fluctuations in wind speed, wind direction, temperature, humidity, and the concentration of water droplets in the air. However, those interactions are currently difficult to study.low-res_2022-02-14-22-05-35-jp_355e1_2b2e5.jpg

Scott Salesky, an assistant professor of meteorology in the College of Atmospheric and Geographic Sciences at the University of Oklahoma, is leading research that will improve the way clouds are represented in weather and climate models. The five-year project is funded by a $763,930 Faculty Early Career Development (CAREER) Award from the National Science Foundation.

“Clouds have a very important influence on Earth’s climate,” said Salesky. “There’s a lot of focus on the role of greenhouse gases in climate, but clouds are also very important. Clouds can reflect solar radiation, and changes in cloud cover could be as important to climate as increases in greenhouse gases.

“The interesting thing about clouds is that there are a lot of processes that happen at different spatial scales that are all coupled together,” he added. “At very small scales, you have cloud droplets forming and growing, and they can interact with turbulence which can impact large-scale cloud properties, such as clouds’ lifetimes, how much sunlight they reflect (what we call albedo), and also the sizes of the droplets in the clouds, which can determine how quickly it might precipitate.”

This project will use simulations of small-scale interactions between millions of cloud droplets and turbulence and will allow the researchers to track the droplets’ positions, movements and sizes as they interact to form, grow and evaporate. 

“When you run a weather or climate model, none of those small-scale processes are going to be resolved, so we’re learning how to accurately represent this small-scale droplet formation and growth and how it’s impacted by turbulence in large-scale weather and climate models,” Salesky said. “From this very small-scale information, we’re going to better understand interactions between turbulence and what we call the microphysics – droplet formation, growth, and evaporation.”

In the later stages of the project, Salesky plans to use the simulations he develops to improve meteorologists’ basic understanding of interactions between turbulence and clouds and to understand how to better model cloud processes in weather and climate models.

The project also supports an educational component to increase graduate student enrollment in atmospheric sciences from students in other STEM fields and to develop lesson plans that connect engineering and physics concepts with atmospheric science.

“The majority of applicants to the meteorology graduate program have a bachelor’s degree in meteorology,” Salesky said. “We want to engage with students from physics, engineering, and other backgrounds to bring their experience with concepts like fluid dynamics to broaden the expertise in our field.”

Salesky is working with faculty partners at Cameron University in Lawton, the University of Science and Arts of Oklahoma in Chickasha, and the University of Central Oklahoma in Edmond to test lesson plans he will develop that teach physics concepts in a meteorological context. The faculty partners will then provide feedback on the lesson plans so that final versions can be made publicly available through the K20 LEARN Repository at OU’s K20 Center, a statewide research and development center. The collaboration with the Oklahoma universities will also support undergraduate research opportunities for students from the three schools to gain experience in atmospheric science research at OU.

Berrien Moore, dean of the College of Atmospheric and Geographic Sciences at OU, said, “Professor Salesky’s receipt of an NSF Career Award sounds and is highly technical and quite advanced. It is also very important in subject matter and extraordinary achievement for a young scientist.

“Not only will Scott advance important science, but he is also going to increase engagement of undergraduates from physics and other STEM fields in atmospheric science,” he added. “Both his scientific research and his teaching are focused on the future, and for that, we are both proud and thankful.”