One company is going to great lengths to build it up, but it will be years before it returns to the minimum level.Doses of cholera vaccine are being given to patients as fast as they are produced and the global stockpile has run completely dry, as deadly outbreaks of the disease continue to spread.This does not shock anyone in the field of emergency epidemic response because the vaccine stockpile has been precariously low for years.The surprise — the good news, which is in itself surprising since ‘cholera’ and ‘good news’ are rarely used together — is that three new vaccine makers are setting up production lines and joining the effort to replenish the stockpile.And a fourth company, the only one that currently makes the vaccine, which is given orally, has been working at a pace that experts describe as “heroic” to expand its production.Yet even with all this, the total global supply of the vaccine that will become available this year will be, at best, a quarter of what is needed.At the end of February, countries had already reported 79,300 cases and 1,100 deaths from cholera this year. Since there is no uniform system for counting cases, this is most likely a gross underestimate.We are having trouble retrieving the article content.Please enable JavaScript in your browser settings.Thank you for your patience while we verify access. If you are in Reader mode please exit and log into your Times account, or subscribe for all of The Times.Thank you for your patience while we verify access.Already a subscriber? Log in.Want all of The Times? Subscribe.

A new study showed that a non-invasive imaging test can help identify patients with coronary artery blockage or narrowing who need a revascularization procedure. The findings were published as a Special Report in Radiology: Cardiothoracic Imaging, a journal of the Radiological Society of North America (RSNA).
Doctors use coronary CT angiography (CTA) to diagnose narrowed or blocked arteries in the heart. A CTA exam receives a score from mild (0-1) to moderate (2-3) to severe (4-5). Patients with scores above 3 typically require medical treatments and can potentially benefit from stents or surgeries (revascularization) to restore blood flow to the heart.
“CTA tells you the degree to which a vessel is blocked,” said Mangun Kaur Randhawa, M.D., a post-doctoral research fellow in the Department of Radiology at Massachusetts General Hospital (MGH) in Boston. “But the degree of blockage doesn’t always reliably predict the amount of blood flow in the vessel.”
Doctors have traditionally relied on an invasive procedure known as invasive coronary angiography to image vessels and more recently have added other invasive tests like fractional flow reserve (FFR) to identify and assess significant blockages in the vessels. CT-FFR is a relatively new alternative that non-invasively models a patient’s coronary blood flow using CTA images of the heart, AI algorithms and/or computational fluid dynamics.
To assess the impact of the selective use of CT-FFR on clinical outcomes, Dr. Randhawa’s research team conducted a retrospective study of patients who underwent coronary CTA at MGH between August 2020 and August 2021.
During the study period, 3,098 patients underwent coronary CTA. Of these, 113 coronary bypass grafting patients were excluded. Of the remaining 2,985 patients, 292 (9.7%) were referred for CT-FFR analysis, and eight of these exams were excluded, leaving a final study group of 284.
As expected, most referrals to CT-FFR were patients with scores of 3 or above. CT-FFR was requested in the majority (73.5 %) of patients with a score of 3 (moderate narrowing/blockage).

“In patients with moderate narrowing or blockage of the arteries, there can be ambiguity about who would benefit from invasive testing and revascularization procedures,” Dr. Randhawa said. “CT-FFR helps us identify and select those patients who are most likely to benefit.”
Out of the 284 patients, 160 (56.3%) had a negative CT-FFR result of > 0.80, 88 patients (30.9%) had a clearly positive (abnormal) result of ≤ 0.75, and the remaining 36 patients (12.6%) had a borderline result between 0.76-0.80.
Patients with significant narrowing/blockages on coronary CTA who underwent CT-FFR had lower rates of invasive coronary angiography (25.5% vs. 74.5%) and subsequent percutaneous coronary intervention (21.1% vs. 78.9%) than patients who were not referred for a CT-FFR.
“CT-FFR helps us identify patients who would most benefit from undergoing invasive procedures and to defer stenting or surgical treatment in patients who likely won’t,” said senior author Brian B. Ghoshhajra, M.D., M.B.A., associate chair for operations and academic chief of cardiovascular imaging at MGH. “CT-FFR makes the CT ‘better’, but we found that the benefits were highest when used selectively.”
Dr. Ghoshhajra added that their CT-FFR analysis was successful in the large majority of patients, regardless of challenging factors such as elevated or irregular heart rates and obesity.
“When you objectively measure coronary artery flow with CT-FFR, you induce fewer patients to be further investigated and treated, because you tend to treat not just what the eyeball sees, but what the physiology supports,” he said.
The researchers said the study results demonstrate the utility of CT-FFR in clinical practice, when used selectively, highlighting its potential to reduce the frequency of invasive procedures in patients with significant coronary artery narrowing or blockages without compromising safety.

Treatment for rheumatoid arthritis (RA) has come a long way in recent years. In many cases, a battery of medications can now successfully stymy the inflammatory cells that cause swelling and pain when they infiltrate tissues around the joints.
Yet for some reason, about 20% of patients with painful, visibly swollen joints consistently get no relief from multiple rounds of even the strongest of these anti-inflammatory drugs.
Surgical interventions intended to remove inflamed tissue have revealed why: “In some cases, their joints aren’t actually inflamed,” says co-senior author Dana Orange, an associate professor of clinical investigation in Rockefeller’s Laboratory of Molecular Neuro-oncology. “With these patients, if you press on the joint, it feels mushy and thick to the touch, but it’s not caused by the infiltrating immune cells. They have excessive tissue growth, but without inflammation. So why are they experiencing pain?”
In a new paper in Science Translational Medicine, she and her colleagues suggest an explanation. These patients have a suite of 815 genes that activate abnormal growth of sensory neurons in tissues that cushion the affected joints.
“These 815 genes are rewiring the sensory nerves, which explains why anti-inflammatory drugs don’t work to alleviate pain for these patients,” says Orange. The findings may lead to new treatments for these outliers.
A puzzling disconnection
Rheumatoid arthritis is a tricky chronic disease. Its symptoms — stiffness, tenderness, swelling, limited motion, and pain — slowly emerge in the hands, wrists, feet, and other joints. It occurs symmetrically (not just in one hand but in both, for instance) and sporadically, with irregular flare-ups. Extreme fatigue and depression are also common.

Most cases of RA are caused by products of immune cells such as cytokines, bradykinins, or prostanoids invading the synovium — a soft tissue lining the joints — where they bind to damage-sensing pain receptors. Drugs that target immune mediators have made RA a far more tolerable condition for most, but those suffering from the disconnection between inflammation and ache haven’t benefitted.
Doctors often prescribe these patients drug after anti-inflammatory drug in an ultimately fruitless attempt to give relief. As a result, “we are subjecting some patients to a lot of medications that cause immunosuppression and yet have little chance of making their symptoms better,” Orange says.
She and her colleagues sought answers in the genes expressed in the joint tissue samples of these patients.
Genetic culprits
The researchers looked at tissue samples and self-reported pain reports from 39 patients with RA who had pain but little inflammation. They also developed a machine-learning analysis that they coined graph-based gene expression module identification (GbGMI).
GbGMI tests every possible combination of genes in a dataset to determine the optimal set of genes that together associate with a targeted clinical feature — in this case, pain.

Using RNA sequencing, the researchers found that of the 15,000 genes expressed in the tissue samples, about 2,200 had increased expression in the 39 patients. Using GbGMI, they identified 815 genes that together associated with patient reports of pain.
“This is a challenging problem, because we have a large number of genes but a limited number of patients,” says co-senior author Fei Wang, professor of population health sciences and founding director of the Institute of Artificial Intelligence for Digital Health at Weill Cornell Medicine. “The graph-based approach we used effectively explored the collective associations between a gene set and patient-reported pain.”
Single cell sequencing analysis found that of the four types of fibroblasts in synovial tissue, CD55+ fibroblasts exhibited the highest expression of pain-associated genes. Located in the outer synovial lining, CD55+ cells secrete synovial fluid, allowing for frictionless joint movement. They also expressed the NTN4 gene, which codes for a protein called Netrin-4. Proteins in the netrin family guide axon growth paths and promote new vascular growth.
Surprising pain pathways
These genes, it turned out, were enriched in pathways that are important for neuron axon growth, the researchers discovered. The keys to sensation, sensory neurons receive and transmit information to the central nervous system. Axons are the tendrils that branch out from them into tissues.
“That led us to hypothesize that perhaps the fibroblasts are producing things that alter the growth of sensory nerves,” Orange says.
But what role was the protein playing in the sensation of pain?
To find out, they grew neurons in vitro and then doused them with Netrin-4, which sparked the sprouting and branching of CGRP+ (gene-related peptide) pain receptors. It’s the first time that Netrin-4 has been shown to alter the growth of pain-sensitive neurons, she notes.
Imaging of RA synovial tissue also revealed an overabundance of blood vessels, which feed and nurture new cells. These vessels were encased by CGRP+ sensory nerve fibers and were growing towards the lining fibroblasts in areas of excessive tissue growth, or hyperplasia. This process likely leads to the squishy swelling that many rheumatologists and surgeons have mistaken for inflammation.
Better drugs
In the future, the researchers aim to home in on other products that fibroblasts may be producing that can affect the growth of pain-sensitive neurons. They’ll also delve into the other types of sensory nerves that might be affected.
“We studied one type, but there are about a dozen. We don’t know if all nerves are affected equally. And we don’t want to block all sensation. Sensory nerves are important for knowing that you should avoid certain movements and the position of your joint in space, for instance,” Orange says.
“We want to drill down on those details so that hopefully we can come up with other treatments for patients who don’t have a lot of inflammation. Right now, they’re taking medications that can cost $70,000 a year but have no chance of working. We must do a better job of getting the right drug to the right patient.”

Childhood maltreatment can continue to have an impact long into adulthood because of how it effects an individual’s risk of poor physical health and traumatic experiences many years later, a new study has found.
Individuals who experienced maltreatment in childhood — such as emotional, physical and sexual abuse, or emotional and physical neglect — are more likely to develop mental illness throughout their entire life, but it is not yet well understood why this risk persists many decades after maltreatment first took place.
In a study published in Proceedings of the National Academy of Sciences, scientists from the University of Cambridge and Leiden University found that adult brains continue to be affected by childhood maltreatment in adulthood because these experiences make individuals more likely to experience obesity, inflammation and traumatic events, all of which are risk factors for poor health and wellbeing, which in turn also affect brain structure and therefore brain health.
The researchers examined MRI brain scans from approximately 21,000 adult participants aged 40 to 70 years in UK Biobank, as well as information on body mass index (an indicator of metabolic health), CRP (a blood marker of inflammation) and experiences of childhood maltreatment and adult trauma.
Sofia Orellana, a PhD student at the Department of Psychiatry and Darwin College, University of Cambridge, said: “We’ve known for some time that people who experience abuse or neglect as a child can continue to experience mental health problems long into adulthood and that their experiences can also cause long term problems for the brain, the immune system and the metabolic system, which ultimately controls the health of your heart or your propensity to diabetes for instance. What hasn’t been clear is how all these effects interact or reinforce each other.”
Using a type of statistical modelling that allowed them to determine how these interactions work, the researchers confirmed that experiencing childhood maltreatment made individuals more likely to have an increased body mass index (or obesity) and experience greater rates of trauma in adulthood. Individuals with a history of maltreatment tended to show signs of dysfunction in their immune systems, and the researchers showed that this dysfunction is the product of obesity and repeated exposure to traumatic events.
Next, the researchers expanded their models to include MRI measures of the adult’s brains and were able to show that widespread increases and decreases in brain thickness and volume associated with greater body mass index, inflammation and trauma were attributable to childhood maltreatment having made these factors more likely in the first place. These changes in brain structure likely mean that some form of physical damage is occurring to brain cells, affecting how they work and function.
Although there is more to do to understand how these effects operate at a cellular level in the brain, the researchers believe that their findings advance our understanding of how adverse events in childhood can contribute to life-long increased risk of brain and mind health disorders.
Professor Ed Bullmore from the Department of Psychiatry and an Honorary Fellow at Downing College, Cambridge, said: “Now that we have a better understanding of why childhood maltreatment has long term effects, we can potentially look for biomarkers — biological red flags — that indicate whether an individual is at increased risk of continuing problems. This could help us target early on those who most need help, and hopefully aid them in breaking this chain of ill health.”
The research was supported by MQ: Transforming Mental Health, the Royal Society, Medical Research Council, National Institute for Health and Care Research (NIHR) Cambridge Biomedical Research Centre, the NIHR Applied Research Collaboration East of England, Girton College and Darwin College.

Childhood verbal abuse by adults costs society an estimated $300 billion (£239 billion) a year globally, show findings presented at the first international conference on childhood verbal abuse, hosted by UCL, Words Matter and the World Health Organization (WHO).
The Words Matter: Impact and Prevention of Childhood Verbal Abuse conference marks the first time that experts from around the world have come together to focus attention on the lifetime damage of childhood verbal abuse and the need to develop solutions.
Childhood verbal abuse involves behaviours that can be detrimental to a child’s wellbeing, such as belittling, shouting and threatening language.
The new study, led by Professor Xiangming Fang (China Agricultural University and Georgia State University) used data from the US Centers for Disease Control and Prevention’s Violence Against Children Surveys, in four countries: Cambodia (1212 participants), Kenya (1099 participants), Colombia (1415 participants) and Moldova (906 participants), to analyse the effects of childhood verbal abuse on selected health outcomes, including mental distress, self-harm, drug use and problem drinking.
The study then estimated the Disability-Adjusted Life Years (DALY) lost (the total amount of healthy life years lost due to people dying prematurely or living with a disability caused by a common disease or health problem in the community) due to health outcomes attributed to childhood verbal abuse to estimate its economic burden.
These DALY losses were then converted into monetary value — assuming that one DALY was equal to the country’s per-capita Gross Domestic Product (GDP).
The mean economic burden of childhood verbal abuse across the four countries was found to be 0.34% of GDP. When this figure was applied to global GDP, it equated to approximately $300 billion every year.

Meanwhile, the DALY losses for outcomes attributed to childhood verbal abuse were significantly great than corresponding estimates for breast cancer and liver cancer in the four countries studied, and similar to the Disability-Adjusted Life Years lost to hypertensive heart disease.
Conference Chair, Professor Peter Fonagy (UCL Psychology & Language Sciences), said: “Verbal abuse of children by adults is all too common, but is one of the most significant modifiable causes of life-long mental health disorders.
“Tackling it gives us a powerful lever to prevent mental health disorders and their enormous cost to both the UK and global economy.
“I am delighted that with the Words Matter charity, we have an organisation finally focusing on this problem. Bringing greater awareness to childhood verbal abuse has the potential to dramatically reduce the economic and psychological burden of psychiatric disorders.”
Previous research from experts at UCL and Wingate University* found that childhood verbal abuse can be as harmful as other forms of abuse and have significant adverse impacts on children’s mental and physical health and development — leading to anxiety, depression, eating disorders, self-harm, substance abuse and even suicide.
Professor Xiangming Fang said: “The economic burden of childhood verbal abuse by adults that we have quantified clearly highlights the shocking hidden cost of the damage it causes to children throughout their lifetime. However, this is likely a considerable underestimate given the impact of childhood verbal abuse on several outcomes including healthcare utilisation costs and legal system expenses, which were not included in the analysis due to data unavailability.

“There is clearly a significant opportunity for economic growth by ending childhood verbal abuse, and by revealing these figures, we hope this form of childhood maltreatment will be given the attention it deserves. Vital now is undertaking more research and devoting funds and resources to preventing it, so the cost to society can be reduced.”
Jessica Bondy, Founder of Words Matter, said: “For too long, childhood verbal abuse by adults has gone under the radar, yet it is all around us. We hope this conference helps put the issue firmly on the map and galvanises action. It is possible to bring an end to childhood verbal abuse with greater awareness, understanding and collaboration across the globe to devise solutions. We must act now, given the lifelong impact on children’s mental and physical health and wellbeing and the monumental cost to society. Let’s build children up, not knock them down, and create a better future for children.”
Tim Loughton MP, Chair of the All-Party Parliamentary Group for Children and former Children’s Minister said: “By convening this conference, the work of Words Matter and its expert advisers is filling a gap in understanding the harms that can be inflicted on children from the way adults communicate with them and the routes for prevention. Whilst we are all too familiar with the damage done to children as a result of physical violence, verbal abuse is more insidious and pervasive, impacting so many whose mental health has already been deeply affected due to the Covid pandemic.
“We all have a duty of care to treat children with respect and that includes the words and language we use with them. This conference highlights how much words really do matter and how if used poorly, they can have lasting implications for children and our economy.”
Former Health Minister and Chief Executive of the NHS Confederation and currently Opposition Whip, Lord Philip Hunt of Kings Heath said: “All children deserve to grow up happy and healthy, but millions are suffering verbal abuse by adults which has for far too long been hidden in the shadows. Thanks to the work of Words Matter, we now know that this is not only impacting children’s mental health and development but is also, as this new study shows, having a huge cost on society as a whole. We all want children to develop armed with the tools to lead confident and productive adult lives and the words they hear from adults are so important in building self-esteem and confidence. By shining a light on this abuse, lives can and will be changed.”
Study limitations
The $300 billion is likely a considerable underestimate given the impact of childhood verbal abuse on several outcomes, including healthcare utilisation costs and legal system expenses, which were not included in the analysis, due to data unavailability.
Additionally, costs associated with chronic illnesses such as diabetes, heart disease, and cancer were not taken into account.
Multiple risk factors can contribute to the occurrence of any disease. When attributing a disease to these factors the maximum attribution is capped at 100%. However, if multiple risk factors are involved and overlap, the sum of DALY estimates for all risk factors may exceed 100%. Failure to fully account for these inter-correlations could potentially result in overestimation of the figures instead.
The field lacks reliable longitudinal data to assess the long-term repercussions of childhood verbal abuse.
The absence of high-quality cohort studies that adopt a lifetime perspective in economic data estimation may lead to a substantial underestimation of the economic impact of childhood verbal abuse.
Prioritising the collection of longitudinal data on the consequences of childhood verbal abuse should be a primary focus of future efforts.

Although lung cancer is traditionally thought of as a “smoker’s disease,” a surprising 15-20% of newly diagnosed lung cancers occur in people who have never smoked, many of whom are in their 40s or 50s.
Doctors say this concerning rise in non-smoking lung cancer cases is likely linked to long-term, high exposures of radon gas. This colorless, odorless gas is emitted from the breakdown of radioactive material naturally occurring underground that then seeps through building foundations. The gas can linger and accumulate in people’s homes and lungs silently unless they know to test for it.
Although the U.S. Environmental Protection Agency (EPA) recommends regular radon testing and corrective measures to lower exposure levels in homes, a new consumer survey conducted on behalf of The Ohio State University Comprehensive Cancer Center — Arthur G. James Cancer Hospital and Richard J. Solove Research Institute (OSUCCC — James) showed that a stunning 75% of Americans have not had their homes tested for radon, and over half (55%) are not concerned about radon exposure in their homes, community or schools.
“Anyone with lungs can develop lung cancer, and as a community we should be aware and concerned about radon exposure because it’s thought to be one of the leading causes of lung cancer in never-smokers — and there is something we can do reduce our risk,” said David Carbone, MD, PhD, a thoracic medical oncologist and director of the OSUCCC — James Thoracic Oncology Center. “There are relatively simple tests to measure radon in the home and actions to reduce radon exposure.”
This includes installing outside the home a radon remediation system that sucks air from the basement, where radon gas typically lingers. Increasing air flow by opening windows and using fans/venting in your home, and sealing cracks in the floors, walls and foundation is also important.
Lung cancer rising in young non-smokers
The No. 1 risk factor for lung cancer is long-term cigarette smoking; however, rates of lung cancer among non-smokers continue to rise. The symptoms of the disease are the same regardless of whether the person has smoked: generally not feeling well or feeling tired all the time, frequent cough, chest pain, wheezing, shortness of breath or coughing up blood. These symptoms happen with other illnesses too, but Carbone notes anyone — regardless of age — who has a lingering symptom that doesn’t resolve despite initial treatment should insist on having it checked out.

Lung cancer screening is currently available only to people at the highest risk for the disease — that means people aged 50 to 80 who also have a 20 pack-year history (one pack of cigarettes per day for 20 years, are current smokers or someone who has quit within the past 15 years.
If detected in its earliest stages, the cure rate for lung cancer can be 90-95%. The bulk of cases, however, are not detected until the disease has spread throughout the lung or to other parts of the body, when treatments aren’t as effective. It is important that anyone deemed at risk for lung cancer get timely screening, and that people who might be at increased risk due to secondhand smoke, radon or occupational exposures (like firefighting) talk to their doctors about testing.
“Your health and the health of your family are the most important things you have. Really push to get your concerns addressed if your symptoms aren’t resolving, even if you don’t fit the typical ‘picture’ of lung cancer. It could truly save your life,” said Carbone.
Requiring radon testing in homes, schools and workplaces
Carbone noted that having high levels of radon exposure at school or work is just as much a health hazard as having high-level exposure in your basement.
He says he strongly supports potential legislation to require radon testing at schools, at places of business and during home sales to help reduce community risk. The effects of radon on your lungs is cumulative and can be delayed by decades.
“So your children playing in your basement or going to school today, exposed to unknown levels of radon, could be at risk for developing lung cancer 10, 20, 30 years from now,” Carbone said. “And because the gas is totally colorless and odorless, you would have no idea you were being exposed unless you knew the importance of proactively testing.”
Survey methodology and results
This survey was conducted by SSRS on its Opinion Panel Omnibus platform. The SSRS Opinion Panel Omnibus is a national, twice-per-month, probability-based survey. Data collection was conducted from February 2- February 4, 2024, among a sample of 1,006 respondents. The survey was conducted via web (n=976) and telephone (n=30) and administered in English. The margin of error for total respondents is +/- 3.5 percentage points at the 95% confidence level. All SSRS Opinion Panel Omnibus data are weighted to represent the target population of U.S. adults ages 18 or older.

In biomedicine, segmentation involves annotating pixels from an important structure in a medical image, like an organ or cell. Artificial intelligence models can help clinicians by highlighting pixels that may show signs of a certain disease or anomaly.
However, these models typically only provide one answer, while the problem of medical image segmentation is often far from black and white. Five expert human annotators might provide five different segmentations, perhaps disagreeing on the existence or extent of the borders of a nodule in a lung CT image.
“Having options can help in decision-making. Even just seeing that there is uncertainty in a medical image can influence someone’s decisions, so it is important to take this uncertainty into account,” says Marianne Rakic, an MIT computer science PhD candidate.
Rakic is lead author of a paper with others at MIT, the Broad Institute of MIT and Harvard, and Massachusetts General Hospital that introduces a new AI tool that can capture the uncertainty in a medical image.
Known as Tyche (named for the Greek divinity of chance), the system provides multiple plausible segmentations that each highlight slightly different areas of a medical image. A user can specify how many options Tyche outputs and select the most appropriate one for their purpose.
Importantly, Tyche can tackle new segmentation tasks without needing to be retrained. Training is a data-intensive process that involves showing a model many examples and requires extensive machine-learning experience.
Because it doesn’t need retraining, Tyche could be easier for clinicians and biomedical researchers to use than some other methods. It could be applied “out of the box” for a variety of tasks, from identifying lesions in a lung X-ray to pinpointing anomalies in a brain MRI.

Ultimately, this system could improve diagnoses or aid in biomedical research by calling attention to potentially crucial information that other AI tools might miss.
“Ambiguity has been understudied. If your model completely misses a nodule that three experts say is there and two experts say is not, that is probably something you should pay attention to,” adds senior author Adrian Dalca, an assistant professor at Harvard Medical School and MGH, and a research scientist in the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL).
Their co-authors include Hallee Wong, a graduate student in electrical engineering and computer science; Jose Javier Gonzalez Ortiz PhD ’23; Beth Cimini, associate director for bioimage analysis at the Broad Institute; and John Guttag, the Dugald C. Jackson Professor of Computer Science and Electrical Engineering. Rakic will present Tyche at the IEEE Conference on Computer Vision and Pattern Recognition, where Tyche has been selected as a highlight.
Addressing ambiguity
AI systems for medical image segmentation typically use neural networks. Loosely based on the human brain, neural networks are machine-learning models comprising many interconnected layers of nodes, or neurons, that process data.
After speaking with collaborators at the Broad Institute and MGH who use these systems, the researchers realized two major issues limit their effectiveness. The models cannot capture uncertainty and they must be retrained for even a slightly different segmentation task.

Some methods try to overcome one pitfall, but tackling both problems with a single solution has proven especially tricky, Rakic says.
“If you want to take ambiguity into account, you often have to use an extremely complicated model. With the method we propose, our goal is to make it easy to use with a relatively small model so that it can make predictions quickly,” she says.
The researchers built Tyche by modifying a straightforward neural network architecture.
A user first feeds Tyche a few examples that show the segmentation task. For instance, examples could include several images of lesions in a heart MRI that have been segmented by different human experts so the model can learn the task and see that there is ambiguity.
The researchers found that just 16 example images, called a “context set,” is enough for the model to make good predictions, but there is no limit to the number of examples one can use. The context set enables Tyche to solve new tasks without retraining.
For Tyche to capture uncertainty, the researchers modified the neural network so it outputs multiple predictions based on one medical image input and the context set. They adjusted the network’s layers so that, as data move from layer to layer, the candidate segmentations produced at each step can “talk” to each other and the examples in the context set.
In this way, the model can ensure that candidate segmentations are all a bit different, but still solve the task.
“It is like rolling dice. If your model can roll a two, three, or four, but doesn’t know you have a two and a four already, then either one might appear again,” she says.
They also modified the training process so it is rewarded by maximizing the quality of its best prediction.
If the user asked for five predictions, at the end they can see all five medical image segmentations Tyche produced, even though one might be better than the others.
The researchers also developed a version of Tyche that can be used with an existing, pretrained model for medical image segmentation. In this case, Tyche enables the model to output multiple candidates by making slight transformations to images.
Better, faster predictions
When the researchers tested Tyche with datasets of annotated medical images, they found that its predictions captured the diversity of human annotators, and that its best predictions were better than any from the baseline models. Tyche also performed faster than most models.
“Outputting multiple candidates and ensuring they are different from one another really gives you an edge,” Rakic says.
The researchers also saw that Tyche could outperform more complex models that have been trained using a large, specialized dataset.
For future work, they plan to try using a more flexible context set, perhaps including text or multiple types of images. In addition, they want to explore methods that could improve Tyche’s worst predictions and enhance the system so it can recommend the best segmentation candidates.
This research is funded, in part, by the National Institutes of Health, the Eric and Wendy Schmidt Center at the Broad Institute of MIT and Harvard, and Quanta Computer.

Scientists at Duke-NUS Medical School have uncovered why some pancreatic and colorectal cancers[1] fail to respond to Wnt inhibitors, a promising new class of cancer drugs currently under development for these cancers. Their discovery, published in Science Advances, not only offers a new cancer therapy target but also a potential screening tool to identify those patients who will not benefit from this new therapy once it becomes available.
Many gastrointestinal cancers grow uncontrollably when a mutation sets a key biological pathway that governs cell growth, called Wnt, on hyperdrive. The Wnt pathway is hijacked in this way in more than 80 per cent of colorectal cancers and some pancreatic cancers, driving rampant tumour growth. For this group of patients, drugs that block the Wnt pathway, known as Wnt inhibitors, hold great promise and they have been at the centre of intense scientific study, including at Duke-NUS[2].
“Although Wnt inhibitors have shown some promise in certain patients, our study reveals intrinsic resistance in others,” said Dr Zhong Zheng, who led the study as a postdoctoral fellow at Duke-NUS’ Cancer & Stem Cell Biology Programme. “Understanding the mechanisms behind this resistance is crucial for personalised treatments for patients when the drugs don’t slow tumour growth at all.”
Focusing on colorectal and pancreatic cancers with a hyperactive Wnt pathway, Dr Zhong, together with Professor David Virshup who leads the Programme at Duke-NUS, used their Wnt inhibiting drug ETC-159, whose efficacy had been established in preclinical models, to assess the cancer cells’ responsiveness.
By analysing genetics data on both responsive and non-responsive tumours, they discovered that a second mutation in another gene, known as FBXW7, makes cancer cells stubbornly resistant to Wnt-blocking drugs.
FBXW7 mutations occur in about 15 per cent of colorectal cancers. “The FBXW7 mutations change the personality of the cancer,” explained Dr Zhong. “They no longer ‘care’ about the Wnt pathway and so the drugs no longer can do their work.”
Testing tumours for the FBXW7 genetic mutations could spare many patients from receiving ineffective treatment, making this not only a potential biomarker but also as a target for a new type of cancer treatment.

“Predicting drug resistance is critical for precision oncology,” said senior author Prof Virshup. “This work reveals how cancers can evade dependencies on Wnt signalling and serves as a solid foundation for further development.”
“We can now try to target those backup pathways activated by the FBXW7 mutation to overcome the drug resistance,” added Dr Zhong, pointing to new treatment possibilities.
The findings add to previous work by the scientists on how pancreatic cancers become resistant to treatment. Together, these discoveries increase our understanding of the ways cancers find alternate routes to grow and survive.
With more precise therapeutic targets to lock onto, these findings bring the promise of personalised therapies one step closer to reality. In addition to the discovery of FBXW7, the team found these Wnt inhibitor-resistant tumours to be susceptible to an experimental drug known as dinaciclib. Their next step is to investigate the potential of dinaciclib alone and in combination with other agents in treating these cancers.
“Our ultimate goal is to help patients with fully resistant tumours by targeting the alternate cancer pathways unleashed by FBXW7 mutations,” said Prof Virshup. “We hope to translate our findings into more tailored and potent treatment strategies.”
“This research exemplifies the highly translational nature of the basic scientific research conducted at Duke-NUS. Cancers are notoriously diverse, and it is important that we can understand and map that diversity, so that we can offer truly personalised treatment that is effective for the individual and not leave patients to undergo unnecessary therapies that will not work for them,” said Professor Patrick Tan, Senior Vice-Dean for Research at Duke-NUS. “This study is another important step on our journey to make every cancer a treatable disease and the team’s exemplifies our resolve to deliver more effective therapies to patients.”
[1] Colorectal cancers are the second most common type of cancer diagnosed in Singapore for both men and women. Pancreatic cancer is the 10th most common cause of cancer in men in Singapore. The World Health Organisation estimated that in 2020 alone, more than 1.9 million new cases of colorectal cancers were diagnosed worldwide.
[2] Duke-NUS, together with A*STAR, developed a made-in-Singapore Wnt inhibitor called ETC-159 which is currently in early phase clinical trials.

Optimizing electrochemical reactions is essential for the transition to renewable energies. In electrochemical reactions, electric currents and potential differences are used to binding and induce reactions. Electrochemistry is a pre-requisite for hydrogen production, and for batterie technology, and thus for sustainable chemistry. Although there has been a lot of technological development in this area in recent years, there is still room for improvement and a long way towards large scale industrial applications. Scientists from the Cluster of Excellence RESOLV at the Ruhr University Bochum and École normale supérieure in Paris discovered two new aspects to control and thus optimize electrochemical reactions at electrified interfaces. They describe their results in the Journal of the American Chemical Society published online on April, 10, 2024.
Surface sensitive spectroscopy
In order to understand the complex behavior at electrified interfaces, the team examined a critical parameter, called the acid dissociation constant (pKa) of molecules at electrified metal/water interfaces. Whereas in bulk solutions, this value is well known, it has been speculated that this parameter, which is essential for acid/base chemistry can be quite different in the vicinity of electrodes. However, measuring pKa values under electrochemical conditions is experimentally challenging. To address this, the group of Havenith have combined advanced surface specific spectroscopic techniques, notably Surface-Enhanced Raman Spectroscopy (SERS), with theoretical modelling. The results vary with the applied voltage: Acid-base chemistry at electrified interfaces, is clearly different from chemistry in the bulk solution.
Hydrophobic layer and strong electric fields
Their findings highlight two key mechanisms governing acid-base reactions at electrified interfaces: The influence of local hydrophobicity and the impact of strong local electric fields. By analyzing the protonation/deprotonation of glycine molecules, the researchers observed a hydrophobic water/water interface close to the metal surface, leading to a destabilization of zwitterionic forms of glycine. When increasing the applied potential the effect is amplified.
Their results showcase the changes of local solvation properties at metal/water interfaces, presenting new avenues for fine-tuning reactivity in electrochemistry. These insights offer new opportunities for optimizing electrochemical processes and designing novel strategies for catalysis as both factors can be tuned in a controlled way.

A generative artificial intelligence (AI) model that can analyze the narrative accounts of women who have undergone recent childbirth has shown the ability to accurately screen for post-traumatic stress disorder (CB-PTSD), a study by Massachusetts General Hospital (MGH), a founding member of the Mass General Brigham healthcare system has found.
By exploring the capabilities and shortcomings of several models from OpenAI, including ChatGPT, the researchers identified a version that offers rich insights into maternal mental health following traumatic childbirth.
The model can fit seamlessly into routine obstetric care and could potentially be harnessed to assess other mental health disorders. The results of the study were published in Scientific Reports.
“Evaluation of PTSD related to traumatic birth currently relies on extensive clinician evaluation, which fails to meet the urgent need for a rapid, low-cost assessment strategy,” says Sharon Dekel, PhD, director of MGH’s Postpartum Traumatic Stress Disorders Research Program, and senior author of the study.
“The use of brief patient narratives of childbirth analyzed by AI’s text-based computational methods could become an efficient, low-cost, and patient-friendly strategy for detecting CB-PTSD after a traumatic birth and with more research this tool may potentially aid in identifying women at risk for CB-PTSD before the condition fully develops.”
For an estimated eight million women a year globally, childbirth that is traumatic and/or medically complicated is expected to trigger post-traumatic stress disorder, a condition historically has been associated with military combat or severe sexual assault.
In recent years, childbirth has become acknowledged as a significant PTSD trigger which, if left untreated, can impair the health of both the mother and child and result in significant societal costs.

In previous studies, Dekel’s lab found evidence that brief psychological interventions delivered soon after traumatic childbirth can reduce maternal childbirth-related PTSD symptoms.
In their latest study, Dekel in collaboration with first author Alon Bartal, PhD, of Bar-Ilan University in Israel, investigated the effectiveness of artificial intelligence and related machine learning (ML) analysis strategies to detect CB-PTSD.
Specifically, they evaluated the performance of different large language models (LLMs) and variations of ChatGPT and their ability to extract novel insights from text-based data sets derived from the brief narrative descriptions by postpartum women of their childbirth experience.
As part of their work, the team collected short narrative accounts from 1,295 women who had recently given birth.
The study focused on an OpenAI model known as text-embeddings-ada-002, which converted narrative data from the personal accounts of women with and without probable CB-PTSD to a numerical format that was then analyzed by a trained machine learning algorithm developed by the team.
Researchers showed this model had superior performance in identifying postpartum traumatic stress compared to other ChatGPT and large language models, which are typically trained on huge volumes of data allowing them to understand, analyze and interpret natural language.

“The reliance of the ML model using childbirth narrative input from the Open AI model as its exclusive data source presents an efficient mechanism for data collection during the vulnerable postpartum period, demonstrating 85 percent sensitivity and 75 percent specificity in identifying CB-PTSD cases,” notes Dekel.
“Moreover, the model we developed could potentially improve accessibility to CB-PTSD screening and diagnosis by fitting seamlessly into routine obstetric care and providing a foundation for commercial product development and mainstream adoption.”
Dekel, whose research program is dedicated to exploring women’s mental health following traumatic childbirth, underscores the clinical benefits of using a pre-trained large language model to assess potential PTSD in new mothers.
“Early intervention is essential to prevent the progression of this disorder to chronic stages, which can seriously complicate treatment,” the MGH investigator points out.
“Our unique approach could introduce an innovative and cost-effective screening strategy for identifying high-risk women and facilitating timely treatment. It may also holds promise for assessing other mental health disorders, and consequently improving patient outcomes.”
The emergence of artificial intelligence tools in health has been groundbreaking and has the potential to positively reshape the continuum of care. Mass General Brigham, as one of the nation’s top integrated academic health systems and largest innovation enterprises, is leading the way in conducting rigorous research on new and emerging technologies to inform the responsible incorporation of AI into care delivery, workforce support, and administrative processes.?
Dekel is a psychologist at MGH, and assistant professor of Psychology at Harvard Medical School. Bartal is an assistant professor of Information Systems at Bar-Ilan University in Israel. Co-authors in the Dekel Laboratory include Kathleen Jagodnik, PhD, a Harvard research fellow, and Sabrina Chan, a clinical research coordinator.
Dekel was supported by funds from the NIH (Eunice Kennedy Shriver National Institute of Child Health and Human Development, grants R01HD108619, R21HD109546, and R21HD100817).