New artificial intelligence technique dramatically improves the quality of medical imaging

A radiologist’s ability to make accurate diagnoses from high-quality diagnostic imaging studies directly impacts patient outcome. However, acquiring sufficient data to generate the best quality imaging comes at a cost — increased radiation dose for computed tomography (CT) and positron emission tomography (PET) or uncomfortably long scan times for magnetic resonance imaging (MRI). Now researchers with the Athinoula A. Martinos Center for Biomedical Imaging at Massachusetts General Hospital (MGH) have addressed this challenge with a new technique based on artificial intelligence and machine learning, enabling clinicians to acquire higher quality images without having to collect additional data. They describe the technique — dubbed AUTOMAP (automated transform by manifold approximation) — in a paper published today in the journal Nature.

“An essential part of the clinical imaging pipeline is image reconstruction, which transforms the raw data coming off the scanner into images for radiologists to evaluate,” says Bo Zhu, PhD, a research fellow in the MGH Martinos Center and first author of the Nature paper. “The conventional approach to image reconstruction uses a chain of handcrafted signal processing modules that require expert manual parameter tuning and often are unable to handle imperfections of the raw data, such as noise. We introduce a new paradigm in which the correct image reconstruction algorithm is automatically determined by deep learning artificial intelligence.

“With AUTOMAP, we’ve taught imaging systems to ‘see’ the way humans learn to see after birth, not through directly programming the brain but by promoting neural connections to adapt organically through repeated training on real-world examples,” Zhu explains. “This approach allows our imaging systems to automatically find the best computational strategies to produce clear, accurate images in a wide variety of imaging scenarios.”

The technique represents an important leap forward for biomedical imaging. In developing it, the researchers took advantage of the many strides made in recent years both in the neural network models used for artificial intelligence and in the graphical processing units (GPUs) that drive the operations, since image reconstruction — particularly in the context of AUTOMAP — requires an immense amount of computation, especially during the training of the algorithms. Another important factor was the availability of large datasets (“big data”), which are needed to train large neural network models such as AUTOMAP. Because it capitalizes on these and other advances, Zhu says, the technique would not have been possible five years ago or maybe even one year ago.

AUTOMAP offers a number of potential benefits for clinical care, even beyond producing high-quality images in less time with MRI or with lower doses with X-ray, CT and PET. Because of its processing speed, the technique could help in making real-time decisions about imaging protocols while the patient is in the scanner.

“Since AUTOMAP is implemented as a feedforward neural network, the speed of image reconstruction is almost instantaneous — just tens of milliseconds,” says senior author Matt Rosen, PhD, director of the Low-field MRI and Hyperpolarized Media Laboratory and co-director of the Center for Machine Learning at the MGH Martinos Center. “Some types of scans currently require time-consuming computational processing to reconstruct the images. In those cases, immediate feedback is not available during initial imaging, and a repeat study may be required to better identify a suspected abnormality. AUTOMAP would provide instant image reconstruction to inform the decision-making process during scanning and could prevent the need for additional visits.”

Notably, the technique could also aid in advancing other artificial intelligence and machine learning applications. Much of the current excitement surrounding machine learning in clinical imaging is focused on computer-aided diagnostics. Because these systems rely on high-quality images for accurate diagnostic evaluations, AUTOMAP could play a role in advancing them for future clinical use.

“Our AI approach is showing remarkable improvements in accuracy and noise reduction and thus can advance a wide range of applications,” Rosen says. “We’re incredibly excited to have the opportunity to roll this out into the clinical space where AUTOMAP can work together with inexpensive GPU-accelerated computers to improve clinical imaging and outcomes.”

Story Source:

Materials provided by Massachusetts General Hospital. Note: Content may be edited for style and length.

Share Button

Drought-induced changes in forest composition amplify effects of climate change

The face of American forests is changing, thanks to climate change-induced shifts in rainfall and temperature that are causing shifts in the abundance of numerous tree species, according to a new paper by University of Florida researchers.

The result means some forests in the eastern U.S. are already starting to look different, but more important, it means the ability of those forests to soak up carbon is being altered as well, which could in turn bring about further climate change.

“Although climate change has been less dramatic in the eastern U.S. compared to some other regions, such as Alaska and the southwestern U.S., we were interested to see if there were signals in forest inventory data that might indicate climate-induced changes in eastern U.S. forests,” said Jeremy Lichstein, senior author and a UF assistant professor of biology. “The changes we documented are easily masked by other disturbances, which is probably why no one had previously documented them. Without a long-term dataset with millions of trees, we probably could not have detected these changes.”

The study appears today in the journal Nature.

Lichstein and his team based their findings on systematic forest inventories of trees in the eastern U.S. from the 1980s to the 2000s. The team looked specifically at forest biomass, tree species composition, and climate variability. The researchers found that decades of changes in water deficit have reduced forest biomass, causing an influx of trees that are more tolerant to drought but slower growing. This shift results in significant changes in forest species composition with their accompanying ecological effects and, moreover, affects the capacity of forest biomass (the mass of living trees) to store carbon. Healthy forests play a key role in global ecosystems as they contain much of the terrestrial biodiversity on the planet and act as a net sink for capturing atmospheric carbon. As climate change affects the forests, so do the forests affect climate change.

Water stress can be caused by rising temperatures, decreases in rainfall, or a combination of the two. To study changes in soil moisture, the researchers used the Palmer drought severity index to examine average water availability and loss over the study period.

Forests are affected by other human activities such as farming or logging, and many are in a stage of ecological succession with lower biomass compared to mature forests. This history of disturbance made the researchers’ analysis challenging. To solve this, researchers compared forests on the basis of their age. “We compared forests in the 1980s of a given age (for example, an 80-year-old forest) to forests of the same age in the 2000s,” Lichstein said. “In areas where the climate got wetter, our analysis showed increases in biomass over the two decades, whereas in the areas that got drier, there were decreases in biomass. When we look at the eastern U.S. as a whole, there was an overall trend towards a drier climate from the 1980s to the 2000s, and therefore the overall effect of climate over the two decades was to reduce forest biomass.”

Drought-tolerant tree species tend to allocate more carbon to fine roots and less to their leaves and woody parts that would sequester more carbon. Lichstein said that although they expected an increase in drought-tolerant tree abundance would prevent biomass losses triggered by water deficits, the opposite appears to be true. “Functional shifts amplified the effects of climate by making forest biomass more responsive to drying or wetting,” he said. “In hindsight, this makes sense, because drought-tolerant species tend to be slow growing. So, if drought causes a shift towards more drought-tolerant species, biomass will decline compared to forests dominated by fast-growing, drought-intolerant species.”

Overall, the study shows that forest biomass and tree species composition and their combined impact on carbon storage are affected by climatic variability on a sensitive and short timeline — just a few decades. “It is premature to say whether or not the amplification effect that we documented is a widespread phenomenon,” Lichstein said. “We hope that our findings will stimulate further research into relationships between species composition, ecosystem function, and climate variability.”

Story Source:

Materials provided by University of Florida. Note: Content may be edited for style and length.

Share Button

Many back pain patients ‘getting wrong care’

Experts are asking the medical profession to stop offering ineffective and potentially harmful treatments.

Share Button

NHS pay: Unions agree deal for 1.3 million staff

Unions back a deal that gives NHS staff a minimum pay rise of more than 6% over three years.

Share Button

Health24.com | Is coconut water actually good for you?

Coconut water is literally everywhere, from the smoothie bowls on your Instagram feed to the aisles of the nearest supermarket. It’s almost as ubiquitous as water. But does it deserve to be?

“It’s trendy,” says Lauren Richer, a registered dietician. “There’s the idea that coconuts are God’s given gift to the Earth, everything from the meat to the water to the oil.”

People swear by coconut water as a workout recovery drink, but there’s also some evidence showing that coconut water helps lower blood sugar and cholesterol, and that it might even play a role in dental health.

Moment of truth: Is coconut water good for you?

Read more: Can coconut oil really help you lose weight?

Coconut water nutrition

First off, here’s the nutritional breakdown for one cup of unsweetened coconut water:

Kilojoules: 184
Fat: 0g
Protein: 0.5g
Carbohydrate: 10.4g
Sugar: 9.6g

Some points in coconut water’s favour: It has some important nutrients like potassium (404mg per cup, almost as much as what you’ll find in a banana), vitamin C (24mg, around 30% of your recommended daily value) and magnesium (15mg, roughly 5% of your daily recommended intake).

The downside: That sugar content. Even though this is unsweetened water and these are natural sugars, 9.6g is more than a third of your daily maximum amount of sugar. Yikes.

Read more: Could coconut actually be the key to weight loss?

The kilojoules are also somewhat problematic, according to Richer. Sure, 184kJ seems pretty small but Richer says if you’re going to drink it in place of water (which has, you know, zero kilojoules), you should minimise your kilojoule intake throughout the day to compensate.

That’s especially important given that coconut water is often sold in packages bigger than a one-cup serving. For example, a 330ml carton of Vita Coco is billed as a serving, but is greater than a cup – and contains 246kJ and 15g of sugar.

Once you’re sipping, it might be hard to stop and save that portion for another day.

Read more: A must-try coconut flour crust pizza

Coconut water benefits

Coconut water has been billed as a workout recovery drink (like nature’s Powerade) since it’s high in electrolytes minerals we sweat out during exercise that include sodium, chloride and potassium.

But Richer says for the average person, water is still a better choice to drink post-sweat session than coconut water – unless you’re an Olympic athlete who’s literally sweating buckets for several hours a day. (Although if you’re stuck between a sports drink or coconut water, Richer says go for the latter.)

For post-workout fuel, she recommends a glass of water and a potassium-rich banana instead.

Read more: Can cinnamon water really help you lose weight?

What to look for on the label

Say you want to treat yourself or are actually training for the Olympics (in which case – go you!), Richer advises reading the ingredients to determine where exactly the kilojoules and sugar are coming from.

You should make sure you’re consuming 100% coconut water – so everything is coming from the coconut itself. Some brands are sweetened with coconut palm sugar or high-fructose corn syrup – both of which can affect your blood sugar levels, says Richer.

So… is coconut water good for you?

If you eat a balanced diet, chances are there’s some wiggle room for coconut water in moderation. But again, Richer says that you should be careful about going overboard.

This article was originally published on www.womenshealthmag.com

Image credit: iStock

NEXT ON HEALTH24X

Share Button

IVF egg donor use rises sharply, HFEA figures show

Experts say it may be down to a greater awareness of egg donation and more donors coming forward.

Share Button

Water bacteria probe extended to Queen Elizabeth Hospital

Four patients at the Royal Hospital for Children in Glasgow are now being treated for infections.

Share Button

State-by-state causes of infant mortality in the US

Sudden unexpected death of infants (SUDI) was the most common cause of infant mortality among children born full term in the U.S. according to estimates from a state-by-state study published this week in PLOS Medicine. The study, conducted by Neha Bairoliya of Harvard University and Gunther Fink of the Swiss Tropical and Public Health Institute, found that very low rates of SUDI — which includes sudden infant death syndrome (SIDS), accidental or abuse-related death, and sudden death from ill-defined causes — have been achieved in a few states while rates remain high in most others. The results suggest potential for a major reduction in infant mortality through interventions to reduce SUDI risk.

In their study, Bairoliya and Fink used linked birth and death records to identify the state- and cause-specific burden of infant mortality among full-term infants (born at 37-42 weeks of gestation). A total of 10,175,481 infants born full-term in the US between January 1, 2010, and December 31, 2012, were analyzed. Full-term infant mortality rate (FTIMR) was 2.2 per 1,000 live births overall, and ranged between 1.29 (Connecticut, 95% confidence interval 1.08, 1.53) and 3.77 (Mississippi, 95% CI 3.39, 4.19) at the state level. SUDI accounted for 43% of FTIM; congenital malformations and perinatal conditions accounted for 31% and 11.3% of FTIM, respectively. The largest mortality differentials between states with good and states with poor FTIMR were found for SUDI, with particularly large risk differentials for deaths due to sudden infant death syndrome (SIDS) (odds ratio [OR] 2.52, 95% CI 1.86, 3.42) and suffocation (OR 4.40, 95% CI 3.71, 5.21).

The extent to which these state differentials are due to antenatal care standards, access to health services, or postnatal practices could not be determined. However, after adjustment for differences in maternal education, race, and maternal health, substantial state-level variation in in SUDI remained (OR 1.70, 95% CI 1.48, 1.94). Bairoliya and Fink project a reduction in full-term infant mortality of 4,003 deaths (95% CI 2,284, 5,587) annually in the U.S. if all states were to achieve the mortality levels of the best-performing state in each cause-of-death category. The authors state, “[g]iven the high mortality burden due to SIDS and suffocation, policy efforts to promote compliance with recommended sleeping arrangements could be an effective first step in this direction.”

Story Source:

Materials provided by PLOS. Note: Content may be edited for style and length.

Share Button

Scientists discern new antibiotics resistance mechanism to peptide antibiotics

Non-ribosomal peptide antibiotics, including polymyxin, vancomycin, and teixobactin, most of which contain D-amino acids, are highly effective against multidrug-resistant bacteria. However, overusing antibiotics while ignoring the risk of resistance arising has inexorably led to widespread emergence of resistant bacteria. Elucidating the little known mechanisms of resistance to peptide antibiotics is critical when implementing peptide antibiotics and would improve effectively improve efficiency.

In a recent study, a group of scientists from the Hong Kong University of Science and Technology reveals both the widespread distribution and broad-spectrum resistance potential of D-stereospecific peptidases, providing a potential early indicator of antibiotic resistance to non-ribosomal peptide antibiotics.

Their findings were published in the journal Nature Chemical Biology on Feb 26, 2018.

“We applied an approach to 5,585 complete bacterial genomes spanning the entire domain of bacteria,” said Pei-Yuan Qian, chair professor of the Division of Life Science, HKUST, and lead author of the paper. “With subsequent chemical and enzymatic analyses, we demonstrated a mechanism of resistance toward non-ribosomal peptide antibiotics that is based on hydrolytic cleavage by D-stereo specific peptidases.”

The team identified a family of D-stereospecific resistance peptidases (DRPs) that are phylogenetically widely distributed in nature. The finding DRPs was found to be involved in combating widely-distributed antibiotics containing D-aa for the survival of their host, which was experimentally validated by a combination of CRISPR/Cas9 based gene editing, chemical and enzymatic analyses.

“Given the potential of DRPs for broad-spectrum resistance and their potential to target clinically important antibiotics containing D-aa, these widely distributed resistance genes are likely to be particularly dangerous if they are transferred to opportunistic pathogens,” said Prof. Qian. “The finding DRPs in nature constitute only the tip of the iceberg, which will lead to intense research on the use and development of peptide antibiotics to combat antibiotic resistance.’

Story Source:

Materials provided by Hong Kong University of Science and Technology. Note: Content may be edited for style and length.

Share Button

TRAPPIST-1 planets provide clues to the nature of habitable worlds

TRAPPIST-1 is an ultra-cool red dwarf star that is slightly larger, but much more massive, than the planet Jupiter, located about 40 light-years from the Sun in the constellation Aquarius.

Among planetary systems, TRAPPIST-1 is of particular interest because seven planets have been detected orbiting this star, a larger number of planets than have been than detected in any other exoplanetary system. In addition, all of the TRAPPIST-1 planets are Earth-sized and terrestrial, making them an ideal focus of study for planet formation and potential habitability.

ASU scientists Cayman Unterborn, Steven Desch, and Alejandro Lorenzo of the School of Earth and Space Exploration, with Natalie Hinkel of Vanderbilt University, have been studying these planets for habitability, specifically related to water composition. Their findings have been recently published in Nature Astronomy.

Water on the TRAPPIST-1 Planets

The TRAPPIST-1 planets are curiously light. From their measured mass and volume, all of this system’s planets are less dense than rock. On many other, similarly low-density worlds, it is thought that this less-dense component consists of atmospheric gasses.

“But the TRAPPIST-1 planets are too small in mass to hold onto enough gas to make up the density deficit,” explains geoscientist Unterborn. “Even if they were able to hold onto the gas, the amount needed to make up the density deficit would make the planet much puffier than we see.”

So scientists studying this planetary system have determined that the low-density component must be something else that is abundant: water. This has been predicted before, and possibly even seen on larger planets like GJ1214b, so the interdisciplinary ASU-Vanderbilt team, composed of geoscientists and astrophysicists, set out to determine just how much water could be present on these Earth-sized planets and how and where the planets may have formed.

Calculating water amounts on TRAPPIST-1 planets

To determine the composition of the TRAPPIST-1 planets, the team used a unique software package, developed by Unterborn and Lorenzo, that uses state-of-the-art mineral physics calculators. The software, called ExoPlex, allowed the team to combine all of the available information about the TRAPPIST-1 system, including the chemical makeup of the star, rather than being limited to just the mass and radius of individual planets.

Much of the data used by the team to determine composition was collected from a dataset called the Hypatia Catalog, developed by contributing author Hinkel. This catalog merges data on the stellar abundances of stars near to our Sun, from over 150 literature sources, into a massive repository.

What they found through their analyses was that the relatively “dry” inner planets (labeled “b” and “c” on this image) were consistent with having less than 15 percent water by mass (for comparison, Earth is 0.02 percent water by mass). The outer planets (labeled “f” and “g” on this image) were consistent with having more than 50 percent water by mass. This equates to the water of hundreds of Earth-oceans. The masses of the TRAPPIST-1 planets continue to be refined, so these proportions must be considered estimates for now, but the general trends seem clear.

“What we are seeing for the first time are Earth-sized planets that have a lot of water or ice on them,” says ASU astrophysicist and contributing author, Steven Desch.

But the researchers also found that the ice-rich TRAPPIST-1 planets are much closer to their host star than the ice line. The “ice line” in any solar system, including TRAPPIST-1’s, is the distance from the star beyond which water exists as ice and can be accreted into a planet; inside the ice line water exists as vapor and will not be accreted. Through their analyses, the team determined that the TRAPPIST-1 planets must have formed much farther from their star, beyond the ice line, and migrated in to their current orbits close to the host star.

There are many clues that planets in this system and others have undergone substantial inward migration, but this study is the first to use composition to bolster the case for migration. What’s more, knowing which planets formed inside and outside of the ice line allowed the team to quantify for the first time how much migration took place.

Because stars like TRAPPIST-1 are brightest right after they form and gradually dim thereafter, the ice line tends to move in over time, like the boundary between dry ground and snow-covered ground around a dying campfire on a snowy night. The exact distances the planets migrated inward depends on when they formed. “The earlier the planets formed,” says Desch, “the further away from the star they needed to have formed to have so much ice.” But for reasonable assumptions about how long planets take to form, the TRAPPIST-1 planets must have migrated inward from at least twice as far away as they are now.

Too much of a good thing

Interestingly, while we think of water as vital for life, the TRAPPIST-1 planets may have too much water to support life.

“We typically think having liquid water on a planet as a way to start life, since life, as we know it on Earth, is composed mostly of water and requires it to live,” explains Hinkel. “However, a planet that is a water world, or one that doesn’t have any surface above the water, does not have the important geochemical or elemental cycles that are absolutely necessary for life.”

Ultimately, this means that while M-dwarf stars, like TRAPPIST-1, are the most common stars in the universe (and while it’s likely that there are planets orbiting these stars), the huge amount of water they are likely to have makes them unfavorable for life to exist, especially enough life to create a detectable signal in the atmosphere that can be observed. “It’s a classic scenario of ‘too much of a good thing,'” says Hinkel.

So, while we’re unlikely to find evidence of life on the TRAPPIST-1 planets, through this research we may gain a better understanding of how icy planets form and what kinds of stars and planets we should be looking for in our continued search for life.

Share Button