Latest news with #McGillUniversity
Medscape
a day ago
- Health
- Medscape
AI Algorithm Predicts Transfusion Need in Acute GI Bleeds
SAN DIEGO — A novel generative artificial intelligence (AI) framework known as trajectory flow matching (TFM) can predict the need for red blood cell transfusion and mortality risk in intensive care unit (ICU) patients with acute gastrointestinal (GI) bleeding, researchers reported at Digestive Disease Week (DDW) 2025. Acute GI bleeding is the most common cause of digestive disease–related hospitalization, with an estimated 500,000 hospital admissions annually. It's known that predicting the need for red blood cell transfusion in the first 24 hours may improve resuscitation and decrease both morbidity and mortality. However, an existing clinical score known as the Rockall Score does not perform well for predicting mortality, Xi (Nicole) Zhang, an MD-PhD student at McGill University, Montreal, Quebec, Canada, told attendees at DDW. With an area under the curve of 0.65-0.75, better prediction is needed, Zhang said, whose coresearchers included Dennis Shung, MD, MHS, PhD, assistant professor of medicine and director of Applied Artificial Intelligence at Yale University School of Medicine, New Haven, Connecticut. 'We'd like to predict multiple outcomes in addition to mortality,' said Zhang, who is also a student at the Mila-Quebec Artificial Intelligence Institute. As a result, the researchers turned to the TFM approach, applying it to ICU patients with acute GI bleeding to predict both the need for transfusion and in-hospital mortality risk. The all-cause mortality rate is up to 11%, according to a 2020 study by James Y. W. Lau, MD, and colleagues. The rebleeding rate of nonvariceal upper GI bleeds is up to 10.4%. Zhang said the rebleeding rate for variceal upper gastrointestinal bleeding is up to 65%. The AI method the researchers used outperformed a standard deep learning model at predicting the need for transfusion and estimating mortality risk. Defining the AI Framework 'Probabilistic flow matching is a class of generative artificial intelligence that learns how a simple distribution becomes a more complex distribution with ordinary differential equations,' Zhang told Medscape Medical News. 'For example, if you had a few lines and shapes you could learn how it could become a detailed portrait of a face. In our case, we start with a few blood pressure and heart rate measurements and learn the pattern of blood pressures and heart rates over time, particularly if they reflect clinical deterioration with hemodynamic instability.' Another way to think about the underlying algorithm, Zhang said, is to think about a river with boats where the river flow determines where the boats end up. 'We are trying to direct the boat to the correct dock by adjusting the flow of water in the canal. In this case we are mapping the distribution with the first few data points to the distribution with the entire patient trajectory.' The information gained, she said, could be helpful in timing endoscopic evaluation or allocating red blood cell products for emergent transfusion. Study Details The researchers evaluated a cohort of 2602 patients admitted to the ICU, identified from the publicly available MIMIC-III database. They divided the patients into a training set of 2342 patients and an internal validation set of 260 patients. Input variables were severe liver disease comorbidity, administration of vasopressor medications, mean arterial blood pressure, and heart rate over the first 24 hours. Excluded was hemoglobin, since the point was to test the trajectory of hemodynamic parameters independent of hemoglobin thresholds used to guide red blood cell transfusion. The outcome measures were administration of packed red blood cell transfusion within 24 hours and all-cause hospital mortality. The TFM was more accurate than a standard deep learning model in predicting red blood cell transfusion, with an accuracy of 93.6% vs 43.2%; P ≤ .001. It was also more accurate at predicting all-cause in-hospital mortality, with an accuracy of 89.5% vs 42.5%, P = .01. The researchers concluded that the TFM approach was able to predict the hemodynamic trajectories of patients with acute GI bleeding defined as deviation and outperformed the baseline from the measured mean arterial pressure and heart rate. Expert Perspective 'This is an exciting proof-of-concept study that shows generative AI methods may be applied to complex datasets in order to improve on our current predictive models and improve patient care,' said Jeremy Glissen Brown, MD, MSc, an assistant professor of medicine and a practicing gastroenterologist at Duke University who has published research on the use of AI in clinical practice. He reviewed the study for Medscape Medical News but was not involved in the research. 'Future work will likely look into the implementation of a version of this model on real-time data.' He added: 'We are at an exciting inflection point in predictive models within GI and clinical medicine. Predictive models based on deep learning and generative AI hold the promise of improving how we predict and treat disease states, but the excitement being generated with studies such as this needs to be balanced with the trade-offs inherent to the current paradigm of deep learning and generative models compared to more traditional regression-based models. These include many of the same 'black box' explainability questions that have risen in the age of convolutional neural networks as well as some method-specific questions due to the continuous and implicit nature of TFM.' Elaborating on that, Glissen Brown said: 'TFM, like many deep learning techniques, raises concerns about explainability that we've long seen with convolutional neural networks — the 'black box' problem, where it's difficult to interpret exactly how and why the model arrives at a particular decision. But TFM also introduces unique challenges due to its continuous and implicit formulation. Since it often learns flows without explicitly defining intermediate representations or steps, it can be harder to trace the logic or pathways it uses to connect inputs to outputs. This makes standard interpretability tools less effective and calls for new techniques tailored to these continuous architectures.' 'This approach could have a real clinical impact,' said Robert Hirten, MD, associate professor of medicine and artificial intelligence, Icahn School of Medicine at Mount Sinai, New York City, who also reviewed the study. 'Accurately predicting transfusion needs and mortality risk in real time could support earlier, more targeted interventions for high-risk patients. While these findings still need to be validated in prospective studies, it could enhance ICU decision-making and resource allocation.' 'For the practicing gastroenterologist, we envision this system could help them figure out when to perform endoscopy in a patient admitted with acute gastrointestinal bleeding in the ICU at very high risk of exsanguination,' Zhang told Medscape Medical News. The approach, the researchers said, will be useful in identifying unique patient characteristics, make possible the identification of high-risk patients and lead to more personalized medicine. Hirten, Zhang, and Shung had no disclosures. Glissen Brown reported consulting relationships with Medtronic, OdinVision, Doximity, and Olympus.
CBC
2 days ago
- General
- CBC
Dinos, meet drones: How new technology could reshape the fossil record
A new study is challenging a long-standing method of dating dinosaur fossils in Alberta's Dinosaur Provincial Park — using drone technology. Previously, one of the methods paleontologists have used to date fossils in the UNESCO World Heritage Site is by measuring how high or low skeletons were found above a distinct boundary where two major rock layers meet. That boundary serves as a time stamp that fossils are dated in relation to. But this method gives only a rough age estimate, according to Alexandre Demers-Potvin, the study's lead author and PhD student at McGill University's Redpath Museum. He and his team used drones to capture over 1,000 high-resolution images of a section of the park and recreated it as a 3D model. The findings, published in the journal Palaeontologia Electronica, show that the boundary used to date fossils in the park actually fluctuates in elevation by as much as 12 metres in relatively short distances. That means the reference point itself varies and could be throwing off the estimated ages of fossils measured against it. The drone method, however, brings a new level of precision to fossil dating in the park. "This is easily one of the studies of which I'm proudest," said Demers-Potvin. "It feels great because this is the kind of work that takes years to complete," he said, citing the collaborative effort between researchers and students who contributed to the study. He said drone-assisted 3D modelling "might be a promising way to better understand which dinosaur fossils are actually older than others in that part of Alberta." "If you're able to take a step back by looking at a larger area from the air, it's easier to notice those small differences." Taking to the sky In 2018, Demers-Potvin began exploring a key fossil site in the park called "Bonebed 190," alongside a crew of McGill's vertebrate paleontology field course researchers. This particular section proved to have a rich biodiversity and high preservation quality of fossils, which sparked a long-term project to further research the area. Demers-Potvin and his team introduced drones to the mix in 2021, taking pictures from above, while a ground team placed down markers using a GPS in order to better geolocate the final 3D rendering. Using a technique called structure-from-motion photogrammetry, they stitched the photos together to construct a detailed 3D model of the terrain. "The quality of the images of that model enabled us to essentially split all those hills into overlying rock layers … which means that we essentially split this whole section [into] different time zones, and this is something that hasn't been done very often for that kind of rock outcrop before," Demers-Potvin explained. The researchers say this application of 3D modelling may be a more reliable tool to date fossils, correcting previous estimates and filling in the gaps that exist in the current timeline. Drones help piece together more accurate fossil record Emily Bamforth, a paleontologist and curator at the Philip J. Currie Dinosaur Museum in northwestern Alberta, said drone use is becoming increasingly common in her field. She is not affiliated with the study but said McGill's research is important to understanding the fossil record with accuracy and placing fossils in both a spatial and temporal context. "In the last two decades, there has … been a shift in dinosaur paleontology away from isolated specimens and towards understanding the broader environments and ecosystems in which these animals lived and how they changed over time," she said. "In this, the stratigraphic context in which a fossil is found is critical, to the point that where a fossil is found is as important as the fossil itself." Although Dinosaur Provincial Park is a well-researched site, Demers-Potvin hopes to continue refining what they already know with drone technology — potentially gaining more insight into the biodiversity of an ancient world. "I think we're getting away from that older method [of dating], and I think now we're just filling the gaps between one data point and another data point and that entire stack of sedimentary rock layers that you can find in the badlands," Demers-Potvin said. "This is only the first step as part of a much bigger project where we hope to cover the entire park."
The Guardian
6 days ago
- Health
- The Guardian
How to turn aquafaba into a delectable vegan chocolate mousse – recipe
I love the simplicity of today's dish. Just two ingredients – chocolate and aquafaba – come together to create a waste-saving treat. Bean water has a mild savoury taste, but the dominant flavour here is chocolate, which, without the addition of extra fat (usually in the form of cream), becomes incredibly intense, amplifying its sheer chocolatiness. Aquafaba has magical properties that mean it behaves like egg whites when whisked, creating an airy foam that can be used for everything from mayonnaise to mousse. According to Ada McVean in an article asking What is Aquafaba? for McGill University in Montreal, the water-soluble proteins and sugars in the beans leak into the water during cooking, giving it a similar composition to egg whites; it also contain saponins, which help it foam so well. The other week, I was on a panel about the magic of food, and one of the first questions we were asked was how to balance taste, health and convenience, to assist people to cook healthily and diversify their diets. Well, this dish is a perfect example: it's not only healthy, but it is also incredibly delicious, fun and easy to make. Dark chocolate is low in sugar and rich in minerals such as iron, magnesium and zinc, and has been linked to various health benefits, including reducing the risk of heart disease. Personally, I enjoy a row of super-dark chocolate every afternoon as a little pick-me-up, and it always makes me smile to think that it's good for me. Chocolate melts at body temperature, which is possibly one reason it feels so satisfying on the tongue. However, it is also heat-sensitive: if it is overheated, it will split and turn oily, so always melt chocolate gently, ideally in a heatproof bowl set over a larger pot of hot but not boiling water. 120g aquafaba (ie, from a 400g can of chickpeas)¼ tsp cream of tartar or 1 tsp lemon juice (optional, and only if the aquafaba is struggling to form firm peaks)100g dark chocolate, plus a little extra for grating on top In a clean, grease-free bowl, use an electric whisk to beat the aquafaba to stiff peaks, then keep whisking for at least five more minutes, to ensure the aquafaba is firmly set (unlike egg whites, it is hard to over-whisk). If the aquafaba struggles to form stiff peaks, add the cream of tartar or lemon juice, although that shouldn't be necessary. Gently melt the chocolate in a heatproof bowl set over a pot of hot but not boiling water, stirring occasionally until smooth. Take off the heat and leave to cool slightly – this prevents it seizing up when mixed with the whipped aquafaba. Gently fold a third of the whipped aquafaba into the chocolate to loosen it, then carefully fold in the rest, keeping as much air in the mix as possible. Spoon into four small serving glasses or ramekins and chill for at least eight hours, or overnight. Before serving, top with grated chocolate, if you like. The mousse will keep in the fridge for three to five days.
Medscape
6 days ago
- Health
- Medscape
HIV Medication Adherence Critical for Viral Suppression
A recent study found that antiretroviral therapy adherence below 90% was associated with significantly lower odds of viral suppression among women living with HIV. METHODOLOGY: Researchers analyzed data from community-based prospective cohort study including women living with HIV across British Columbia, Ontario, and Quebec between 2013 and 2018. Overall, 1187 participants (median age, 42 years) who reported antiretroviral therapy use and specified their regimen were included in the analysis. Participants completed questionnaire surveys at three timepoints, at 18 months interval between 2013 and 2018. Analysis focused on the relationship between adherence to antiretroviral therapy and viral suppression using data from the study period between 2017 and 2018. TAKEAWAY: The use of integrase strand transfer inhibitors increased from 13.6% in 2013 to 30.6% in 2018, whereas the utilization of other antiretroviral classes showed declining trends. Among participants, 78.2% reported undetectable viral loads, and viral suppression was achieved by 89.6% of those on regimens consisting of a backbone plus a third agent. Less than 70% adherence (adjusted odds ratio [aOR], 0.06; 95% CI, 0.01-0.27) and 80%-89% adherence (aOR, 0.21; 95% CI, 0.05-0.86) were associated with lower odds of viral suppression compared with ≥ 95% adherence. Researchers found no significant difference in the odds of viral suppression between 90%-94% adherence and ≥ 95% adherence. IN PRACTICE: 'With the chronic nature of HIV and the risk of resistance, supporting patients in overcoming the challenges that accompany long-term medication adherence through cotailored approaches should remain a priority when providing care to women living with HIV,' the authors wrote. SOURCE: The study was led by Alexandra de Pokomandy, Department of Family Medicine, McGill University, Montreal, Quebec, Canada. It was published online on May 2, 2025, in HIV Medicine . LIMITATIONS: The study was limited by a small sample size in groups with lower adherence, which may have reduced the statistical power to detect significant differences. All measures were self-reported, introducing social desirability bias and recall bias. The study also lacked information on the status of mental health treatment and the duration of viral suppression. DISCLOSURES: The study received funding from the Canadian Institutes of Health Research, the Canadian HIV Trials Network, the Ontario HIV Treatment Network, and the Ontario Academic Health Science Centers Alternative Funding Plans Innovation. Two authors reported having financial ties with several pharmaceutical companies.
Yahoo
7 days ago
- Health
- Yahoo
Trees May Be Able to Warn Us When a Volcano Is About to Erupt
The science of predicting volcanic eruptions can genuinely save lives – potentially, a lot of lives – and researchers have shown that tree leaf colors can act as warning signals around a volcano that's about to blow. As volcanoes get more active and closer to an eruption, they push magma up closer to the surface, releasing higher levels of carbon dioxide. That in turn can boost the health of the surrounding trees, making leaves greener. And those changes – specifically in the measurement known as the normalized difference vegetation index (NDVI) – can be spotted by satellites in space. We could be looking at an early warning system for eruptions that doesn't require any local field work or ground sensors, so it could work in remote and difficult-to-access areas. "There are plenty of satellites we can use to do this kind of analysis," says volcanologist Nicole Guinn, from the University of Houston. Guinn was the first author of a recent study looking at carbon dioxide levels around Mount Etna in Italy. The study compared data from sensors around the volcano with satellite imagery, finding a strong relationship between more carbon dioxide and greener trees. Across the course of two years, the team found 16 clear spikes in carbon dioxide and the NDVI, matching magma movements underground. The patterns were even observed farther away from faults in the mountain. That study referenced earlier research from 2019, led by volcanologist Robert Bogue of McGill University, which showed that carbon dioxide emitted by two active volcanoes in Costa Rica had an impact on leaf color in tropical trees in the area. Now Guinn and Bogue, together with other researchers, are working on a project led by NASA and the Smithsonian Institution, analyzing changes in the color of plant life around volcanoes in Panama and Costa Rica. It's part of the collaborative Airborne Validation Unified Experiment: Land to Ocean (AVUELO) mission, which is looking to develop more ways in which we can measure the health of the planet from satellites. Current methods, like NASA's Orbiting Carbon Observatory 2, are only strong enough to pick up major eruptions. "A volcano emitting the modest amounts of carbon dioxide that might presage an eruption isn't going to show up in satellite imagery," says Bogue. "The whole idea is to find something that we could measure instead of carbon dioxide directly, to give us a proxy to detect changes in volcano emissions." There are multiple signals that can be interpreted to predict volcanic eruptions, including the rumble of seismic waves and changes in ground height. With the greening of leaves from carbon dioxide emissions, we now have another signal to measure – even if it won't be suitable for all sites. The AVUELO researchers are also interested in the broader effects of increased carbon dioxide on trees. As our world warms up due to human emissions of carbon dioxide, we could be increasingly reliant on vegetation to regulate this greenhouse gas. "We're interested not only in tree responses to volcanic carbon dioxide as an early warning of eruption, but also in how much the trees are able to take up, as a window into the future of the Earth when all of Earth's trees are exposed to high levels of carbon dioxide." says climate scientist Josh Fisher, from Chapman University in California and part of the AVUELO team. The Mount Etna research was published in Remote Sensing of Environment. 60% of The Ocean Floor Could Harbor 'Rare' Supergiant Crustacean Watch: 1,000-Foot Lava Jets Erupt From Hawaii's Kīlauea Volcano Giant Megalodon's Prey Finally Revealed, And It's Not What We Thought
