AI Identifies Novel Predictors of TB in People With HIV
An artificial intelligence (AI) model using routinely collected data predicted subsequent development of active tuberculosis (TB), Swiss researchers reported. The AI model outperformed biological tests for latent TB in identifying HIV-positive patients at high risk of developing TB.
As well as immune function and sociodemographic variables, the AI model retained several biomarkers indicative of patients' well-being and metabolism.
In Switzerland and other countries with good access to antiretroviral therapy, TB is a rare but serious co-infection in people living with HIV, frequently linked with late HIV diagnosis. To prevent progression to active TB disease, people known to have latent TB infection can be offered preventive treatment with isoniazid and/or rifampicin.
But detection of latent TB is challenging, especially in people with HIV. In a previous Swiss analysis, a combined approach using interferon gamma release assays (IGRA) and tuberculin skin tests identified only 30% of people who subsequently developed active TB.
'It was worse than tossing a coin,' Joahnnes Nemeth, MD, an attending physician in the department of infectious diseases and hospital epidemiology at the University of Zürich, Zürich, Switzerland, told Medscape Medical News.
The problem is that the tests rely on immune response, which may be impaired. 'You interrogate the very system that is malfunctioning during HIV infection, so it's not a surprise that the tests perform poorly,' he explained.
This led him and his colleagues to look into alternative ways to identify patients at risk. They leveraged data from the Swiss HIV Cohort Study, which includes around 70% of people receiving HIV care in the country.
Over 23 years' worth of data were analyzed using machine learning, a subset of AI that enables computers to learn patterns from data and make predictions without being explicitly programmed for each task. Their machine learning model employed a random forest — an algorithm which combines the outputs from multiple decision trees.
The model looked at data collected at HIV diagnosis in order to predict active TB disease that developed at least 6 months later. Rather than only considering variables which the researchers thought were potential risk factors, the model reviewed all the variables for which they had sufficient data.
'What I really liked about this machine learning approach is that we threw all the data we collect into the machine and just asked it: Can you do something with that?' Nemeth said. 'I think that really paid off.'
The first iteration of the model included 48 variables and had a sensitivity of 70.1% and a specificity of 81.0%. A streamlined second version retained 20 variables — making it computationally less demanding — while delivering a sensitivity of 57.1% and specificity of 77.8%.
Given that biologic tests had a sensitivity of 30% and specificity of 94%, for Nemeth this 'blows everything of the water.' The model doesn't require additional data collection or have the expense of IGRA.
As might be expected, the 20 retained variables included immunological parameters, hematological markers, and sociodemographic factors, but some were more surprising: along with several variables linked with metabolism (cholesterol, high-density lipoprotein, glucose, and creatinine), body mass index, and mean arterial pressure.
The researchers noted that TB is associated with malnutrition and said that some of these markers may reflect metabolic perturbations and compromised muscle mass in people at risk for TB.
The model was first validated on a portion of the Swiss cohort which it was not trained on, and then on a cohort in Austria. Despite the many parallels between the two cohorts, initially the model performed badly in Austria.
The researchers realized the issue stemmed from different migration patterns between the countries: Most people with TB in Switzerland have moved from sub-Saharan Africa, while in Austria, most come from the former Soviet republics. Only after modifying the ethnicity and region of birth variables did the model begin to work effectively.
'This is a cautionary tale,' said Nemeth. 'You go to a very similar setting with a little difference, and all this stops working. With machine learning models, we really have to be careful and test them vigorously before we rely on them.'
Emily Wong, MD, is an associate professor at the University of Alabama at Birmingham who has used AI to aid interpretation of chest radiography in South Africa, but was not involved in the new study.
The Swiss research 'opens one's eyes to the idea that with very large data sets with lots of clinical variables, you can discern meaningful and predictive patterns that predict whether someone will go on to develop TB,' she told Medscape Medical News.
Nemeth is working on an implementation study in which physicians whose patients have never been tested for TB will be randomly allocated to either receive a reminder to test, or a risk score based on the machine learning model. A key question is whether the latter will be enough to convince physicians to take further action, such as offering preventative therapy.
Wong noted that the potential benefits and risks (including liver toxicity) of preventative therapy need to be weighed up for each patient. But a machine learning model could help clinicians to do this.
'The idea that in the future, based on key demographic and clinical information of a person, and maybe including their chest x-ray or IGRA test, or maybe not, we would have a well-functioning clinical decision making tool that would guide a health care worker to make TB prevention decisions for the patient in front of them is definitely a worthy goal,' she said.
The study was funded by the Swiss National Science Foundation. Nemeth declared receiving honoraria for presentations from Oxford Immunotec and ViiV.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
CNBC
an hour ago
- CNBC
Gilead Sciences CEO on FDA approval of Yeztugo: A milestone moment in the history of HIV
Gilead Sciences chairman and CEO Daniel O'Day joins 'Squawk Box' to discuss the FDA approval of the company's HIV prevention medication, how the twice-yearly antiviral drug Yeztugo works, the importance of preventive medicine, the fight against HIV, and more.
Yahoo
12 hours ago
- Yahoo
Here's what to know about the twice-yearly preventative HIV shot
The Food and Drug Administration has approved a twice-yearly preventive HIV shot that could protect millions from the virus. People at risk for HIV may choose to take PrEP, a medication that helps prevent HIV infection. PrEp has been prescribed as a daily pill or a shot given every two months. This new twice-yearly drug called lenacapavir is now the longest-lasting type of PrEP. Ian Haddock, a Houston man who participated in a study of the drug, told The Associated Press it 'expands the opportunity for prevention.' 'Now I forget that I'm on PrEP because I don't have to carry around a pill bottle,' he said. The shot, made by Gilead Sciences, is injected under the skin of the abdomen. It leaves a small 'depot' of medication that slowly absorbs into the body. 'This really has the possibility of ending HIV transmission,' Greg Millett, public policy director at amfAR, The Foundation for AIDS Research, told the AP. But the upheaval in U.S. healthcare — including cuts to public health agencies and Medicaid — and slashing of American foreign aid to fight HIV are clouding the prospects. Millett said "gaping holes in the system" in the U.S. and globally "are going to make it difficult for us to make sure we not only get lenacapavir into people's bodies but make sure they come back' even as little as twice a year. Gilead's drug already is sold to treat HIV under the brand name Sunlenca. The prevention dose will be sold under a different name, Yeztugo. Gilead didn't immediately announce its price. The drug only prevents HIV transmission – it doesn't block other sexually transmitted diseases. Global efforts at ending the HIV pandemic by 2030 have stalled. There still are more than 30,000 new infections in the U.S. each year and about 1.3 million worldwide. Only about 400,000 Americans already use some form of PrEP, a fraction of those estimated to benefit. A recent study found states with high use of PrEP saw a decrease in HIV infections, while rates continued rising elsewhere. About half of new infections are in women, who often need protection they can use without a partner's knowledge or consent. One rigorous study in South Africa and Uganda compared more than 5,300 sexually active young women and teen girls given twice-yearly lenacapavir or the daily pills. There were no HIV infections in those receiving the shot while about 2% in the comparison group caught HIV from infected sex partners. A second study found the twice-yearly shot nearly as effective in gay men and gender-nonconforming people in the U.S. and in several other countries hard-hit by HIV. Haddock, who leads the Normal Anomaly Initiative, a nonprofit serving Black LGBTQ+ communities, had tried PrEP off and on since 2015 but he jumped at the chance to participate in the lenacapavir study and continues with the twice-yearly shots as part of the research follow-up. 'Men, women, gay, straight – it really just kinds of expands the opportunity for prevention,' he added. Just remembering a clinic visit every six months 'is a powerful tool versus constantly having to talk about, like, condoms, constantly making sure you're taking your pill every day.' 'Everyone in every country who's at risk of HIV needs access to PrEP,' Dr. Gordon Crofoot of Houston, who helped lead the study in men, told the AP. 'We need to get easier access to PrEP that's highly effective like this is.'
Yahoo
15 hours ago
- Yahoo
What would happen if you tried to land on a gas giant?
Our solar system contains three types of planets. Between the four terrestrial planets–Mercury, Venus, Earth, and Mars–and the distant ice giants of Neptune and Uranus, sit two gas giants: Saturn and Jupiter. These planets are mostly composed of hydrogen and helium gas. Researchers now appreciate that gas planets are more complex than first thought. New findings have implications for our understanding of how these planets formed and will help design future missions to potentially visit them. Gas giants originate from one of two processes. The first method is called core accretion, explains Ravit Helled, a professor of theoretical astrophysics at the University of Zürich. This starts with the birth of a new star, when molecular clouds collapse under gravitational pressure. Whorls of gas–called protoplanetary disks–start to spin around these new stars. Within these gas disks will be heavier particles–dust, rock, or any elements heavier than helium. These particles can clump together and then suck in gas from the surrounding disk, forming a giant planet mainly composed of gas. A second method that may form gas giants called disk instability–this is a newer theory that still causes some controversy among planetary theorists. According to this idea, when massive protoplanetary disks cool down, they become unstable and can produce clumps of rock and gas that evolve into gas giants. Importantly, this proposed formation process happens much more quickly than core accretion. Helled says that Saturn and Jupiter likely formed via core accretion, but that disk instability may 'explain very massive planets at large orbits or giant planets around small mass stars.' Regardless of how they form, the structure of gas giants is nothing like that of terrestrial planets like Earth. Jupiter and Saturn don't have a surface in the same way Earth does. Instead, their atmosphere simply gets thinner until there isn't enough density left to call the surrounding air part of the planet anymore. 'There is no location where you can say, okay, this is where the planet stops,' says Helled. A spaceship attempting to 'land' on Jupiter's 'surface' would have to overcome some significant obstacles. Once you enter the cloud of gas that roughly marks the beginning of a giant like Jupiter, temperature and pressure steadily increase as your head toward the planet's core, and gaseous hydrogen and helium morph into liquid form. While our solar system's gas giants are far from the sun, the core of a gas giant is likely to be incredibly hot–Jupiter's is estimated at around 43,000 degrees Fahrenheit. You'd also have to pass through the thick clouds of ammonia found in Jupiter's upper atmosphere. If you make your ship from tough stuff–tougher than any known substance on Earth–that could survive these conditions, it might make it to a gas giant's core. What it would find there in the alien murk is still unclear. 'For decades, it was assumed that there was a defined core,' says Helled. Recent probe missions, like Juno and Cassini, have orbited Jupiter and Saturn, respectively. The information these probes sent back has changed that view. 'We now think that they have what we call fuzzy or diluted cores,' says Helled. This means that there isn't a clear transition point between the upper layers of liquid gas and liquid hydrogen and helium and the planet's core. In truth, Juno and Cassini's data has revolutionized our understanding of these planet's structures. Helled explains that they likely have complex heat and composition gradients. Jupiter is famously wracked with massive storms, like the Great Red Spot, which produces winds up to 425 mph (640 km/h). Some of these shifts can produce dramatic phenomena. Jupiter and Saturn likely have regions in which helium gas separates from hydrogen. Here, the helium becomes a rain of droplets that pour towards the planet's core. These insights can reveal more about our solar system's giants, as well as similar planets outside our solar system. 'Now we realize that some of the simple assumptions that we've made to model these planets are wrong, and we need to modify the models,' says Helled. This story is part of Popular Science's Ask Us Anything series, where we answer your most outlandish, mind-burning questions, from the ordinary to the off-the-wall. Have something you've always wanted to know? Ask us.
