Trained on 1.5 million whole-slide images from 100,000 patients, a pathology foundation model is shown to improve performance of specialized models in detection of rare cancers.
This post is also available in: 
עברית (Hebrew)
Experts from the Universities of East Anglia, Sheffield, and Leeds have developed a new groundbreaking AI method that improves the accuracy and efficiency of analyzing MRI heart scans. This innovation could provide a way for faster, more accurate, and non-invasive diagnosis of heart failure and other cardiac conditions, thus saving valuable time and resources for the healthcare sector.
According to Innovation News Network, the research team used data from 814 patients at Sheffield and Leeds Teaching Hospitals to train an AI model, which was then tested using scans and data from 101 patients at Norfolk and Norwich University Hospitals to ensure accuracy.
To tackle this issue, Israeli medical technology startup NanoSynex has developed a rapid personalized diagnostic test, which will enable doctors to prescribe the correct antibiotics at the moment they are needed most.
“There is a clear issue of misuse and overuse of antibiotics, and one of the ways to address this crisis, other than developing new antibiotics, is to better use existing antibiotics, by boosting the development of rapid and reliable diagnostic solutions, which is what we are doing at NanoSynex,” Diane Abensur, CEO and co-founder of NanoSynex, tells NoCamels.
NanoSynex offers laboratories better and faster solutions to determine the best treatment plan by providing rapid and accurate results for Antimicrobial Susceptibility Testing (AST), tests that are used to determine which specific antibiotics a particular bacteria or fungus is sensitive to.
Researchers from the Complexity Science Hub and the University of Veterinary Medicine Vienna have dissected the complex interactions involved in zoonoses, which annually affect over two billion people worldwide. They introduce the concept of a “zoonotic web,” a detailed network representation of the relationships between zoonotic agents, their hosts, vectors, food sources, and the environment.
The treated mice were known as “supermodel grannies” in the lab because of their youthful appearance.
They were healthier, stronger and developed fewer cancers than their unmedicated peers.
The drug is already being tested in people, but whether it would have the same anti-ageing effect is unknown.
Implementing error correction in a quantum computer requires putting together a lot of different things. Of course, you want to start with good physical qubits that have as low a physical error rate that you can achieve. You want to add in an error correction algorithm, like the surface code, color code, q-LDPC, or others that can be implemented in your architecture, and you need a fast real time error decoder that can look at the circuit output and very quickly determine what the error is so it can be corrected. The error decoder portion doesn’t get as much attention in the media as the other things, but it is a very critical portion of the solution. Riverlane is concentrating on providing products for this with a series of solutions they name Deltaflow which consists of both a classical ASIC chip along with software. The Deltaflow solution consists of a powerful error decoding layer for identifying errors and sending back corrective instructions, a universal interface that communicates with the computer;s control system, and a orchestration layer for coordinating activities.
Riverlane has released its Deltaflow Error Correction Stack Roadmap that show yearly updates to the technology to support an increase in the number of QuOps (error free Quantum Operations) by 10X every year. We reported last year on a chip called DD1 that is part of their Deltaflow 1 solution that is capable of supporting 1,000 QuOps using a surface code error correction algorithm. And now, Riverlane is defining solutions that will achieve 10,000 QuOps with Deltaflow 2 later this year, 100,000 QuOps with Deltaflow 3 in 2025, and 1,000,000 QuOps, also called MegaQuops in 2026, with their Deltaflow Mega solution.
One characteristic that Riverlane is emphasizing in these designs is to perform the decoding in real time in order to keep the latencies low. Although it is fine for an academic paper to send the ancilla data off to a classical computer and have it determine the error, it might take milliseconds for the operation to complete. That won’t cut it in a production environment running real jobs. With their Deltaflow chips, these operations can be performed at megahertz rates and Riverlane has implemented techniques such as a streaming, sliding window, and parallized decoding approaches to increase the throughput of the decoder chips as much as possible. In future chips they will be implementing “fast logic” capabilities for Clifford gates using approaches including lattice surgery and transversal CZ gates.
For the first time, scientists have discovered that a small region of our brain shuts down to take microsecond-long naps while we’re awake. What’s more, these same areas ‘flicker’ awake while we’re asleep. These new findings could offer pivotal insights into neurodevelopmental and neurodegenerative diseases, which are linked to sleep dysregulation.
Scientists from Washington University in St. Louis (WashU) and the University of California Santa Cruz (UCSC) made these findings by accident, noticing how brain waves in one tiny area of the brain shut down suddenly for just milliseconds when we’re awake. And in this same region, those brain waves jolt suddenly, for the same amount of time, when we’re asleep.
“With powerful tools and new computational methods, there’s so much to be gained by challenging our most basic assumptions and revisiting the question of ‘what is a state?’” said Keith Hengen, Assistant Professor of Biology at WashU. “Sleep or wake is the single greatest determinant of your behavior, and then everything else falls out from there. So if we don’t understand what sleep and wake actually are, it seems like we’ve missed the boat.”
Two companies are coming together to develop an AI Health Coach that uses the power of artificial intelligence to battle chronic diseases.
Identifying individuals who are at a high risk of age-related morbidities may aid in personalized medicine. Circulating proteins can discriminate disease cases from controls and delineate the risk of incident diagnoses1,2,3,4,5,6,7,8. While singular protein markers offer insight into the mediators of disease5,9,10,11, simultaneously harnessing multiple proteins may improve clinical utility12. Clinically available non-omics scores such as QRISK typically profile the 10-year onset risk of a disease13. Proteomic scores have recently been trained on diabetes, cardiovascular and lifestyle traits as outcomes in 16,894 individuals14. Proteomic and metabolomic scores have also been developed for time-to-event outcomes, including all-cause mortality6,15,16,17,18,19,20,21.
Here, we demonstrate how large-scale proteomic sampling can identify candidate protein targets and facilitate the prediction of leading age-related incident outcomes in mid to later life (see the study design summary in Extended Data Fig. 1). We used 1,468 Olink plasma protein measurements in 47,600 individuals (aged 40–70 years) available as part of the UK Biobank Pharma Proteomics Project (UKB-PPP)22. Cox proportional hazards (PH) models were used to characterize associations between each protein and 24 incident outcomes, ascertained through electronic health data linkage. Next, the dataset was randomly split into training and testing subsets to train proteomic scores (ProteinScores) and assess their utility for modeling either the 5-or 10-year onset of the 19 incident outcomes that had a minimum of 150 cases available. We modeled ProteinScores alongside clinical biomarkers, polygenic risk scores (PRS) and metabolomics measures to investigate how these markers may be used to augment risk stratification.
