Low-dose CT chest scans could help detect pneumonia in at-risk patients while exposing them to only small amounts of radiation, a new study has found. The research, published in Radiology: Cardiothoracic Imaging, shows that ultra-low-dose scans can effectively detect pneumonia in patients with compromised immune systems, enabling doctors to treat the infection before it becomes life-threatening. According to the researchers, these scans expose patients to just 2% of the radiation dose used in a standard CT scan.

"This study paves the way for safer, AI-driven imaging that reduces radiation exposure while preserving diagnostic accuracy,” lead researcher Dr Maximiliano Klug, a radiologist with the Sheba Medical Center in Ramat Gan, Israel, said in a news release. He added that CT scans are the gold standard for detecting pneumonia but there are concerns regarding the risk posed by repeated exposure to radiation. There is a solution- ultra-low-dose CT scan. However, the problem is that these scans can be grainy and hard to read, researchers said.

Study Gives Solution To This

To overcome that, Klug's team developed an AI program that could help "de-noise" low-dose scans, making them sharper and easier to read. Between September 2020 and December 2022, 54 patients with compromised immune systems who had fevers underwent a pair of chest CT scans -- a normal dose scan and an ultra-low-dose scan. The AI program cleaned up the low-dose scan, and then both sets of images were given to a pair of radiologists for assessment. Radiologists had 100% accuracy in detecting pneumonia and other lung problems with the AI-cleaned low-dose scans, but 91% to 98% accuracy in examining the scans that hadn’t been improved through AI, results show.

"This pilot study identified infection with a fraction of the radiation dose," Klug said. "This approach could drive larger studies and ultimately reshape clinical guidelines, making denoised ultra-low dose CT the new standard for young immunocompromised patients.

How Can You Detect Pneumonia?

Pneumonia is a lung infection that causes the air sacs in the lungs to fill with fluid or pus and can be caused by bacteria, viruses, or fungi. The symptoms can range from milk to severe, which includes:

Coughing with or without cough

Fever

Chills

Trouble breathing

Chest pain, especially when breathing deeply or coughing

Sweating or chills

Rapid heart rate

Loss of appetite

Bluish skin, lips, and nails

Confusion.

How to detect Pneumonia in coughing newborns and toddlers?

Pneumonia can severely affect newborns and young children as their lungs are comparatively more sensitive. As per Dr Goyal, young children can cough for various reasons including seasonal infections and tonsillitis, which is very common in this age group. But if they look visibly irritable and have poor sleep patterns, then parents must reach out to an expert. "I am not saying that parents must visit a hospital but any local paediatrician would be able to detect pneumonia in your kid.

The US Food and Drug Administration (FDA) has approved Merck's Idvynso (doravirine/islatravir), a new, once-daily pill for the treatment of HIV-1 infection in adults.

The two-drug single tablet replaces the current antiretroviral regimen in those who are virologically suppressed (HIV-1 RNA <50 copies per mL).

The single tablet contains 100 mg doravirine and 0.25 mg islatravir. The FDA has approved it for adults on a stable antiretroviral regimen with no history of virologic treatment failure and no known substitutions associated with resistance to doravirine.

“IDVYNSO combines islatravir, a next-generation NRTI with multiple mechanisms of action, including translocation inhibition, with doravirine, an NNRTI with an established efficacy and safety profile,” said Dr. Eliav Barr, senior vice president and chief medical officer, Merck Research Laboratories, in a statement.

How Was The Pill Approved?

The approval is based on data from two randomized, active-controlled, noninferiority trials. In the double-blind Trial 052, participants were randomly assigned to stay on Glilead's Biktarvy (bictegravir/emtricitabine/tenofovir alafenamide; 171 individuals) or switch to Idvynso (342 individuals).

Results showed that 1 per cent of participants in both groups had a viral load of ≥50 copies/mL at 48 weeks.

In the open-label Trial 051, participants were randomly assigned to stay on their oral antiretroviral therapy (ART) regimen (185 individuals) or switch to Idvynso (366 individuals).

Results showed that 1 per cent of participants who were switched to Idvynso had a viral load of ≥50 copies/mL at week 48 versus 5 per cent who continued on ART.

"As the only two-drug, non-integrase strand transfer inhibitor, tenofovir-free regimen, Idvynso expands therapeutic diversity beyond the currently available oral treatment options," Barr said.

"As the health needs of adults living with HIV change over time, Idvynso gives clinicians a new choice for HIV treatment."

What Is IDVYNSO? How Does It Work?

IDVYNSO is a fixed-dose combination of two medicines, doravirine with islatravir.

Doravirine is a non-nucleoside reverse transcriptase inhibitor (NNRTI) that inhibits HIV-1 replication by non-competitive inhibition of HIV-1 reverse transcriptase.

Islatravir is a potent, next-generation nucleoside analog reverse transcriptase inhibitor (NRTI) that blocks HIV-1 replication by multiple mechanisms, including:

• inhibition of reverse transcriptase translocation, resulting in immediate chain termination,
• induction of structural changes in the viral DNA (delayed chain termination).

Global Burden Of HIV

According to the latest data from UNAIDS, 40.8 million people globally were living with HIV in 2024. Of these, 39.4 million were adults (15 years or older) and 1.4 million were children (0–14 years).

While 1.3 million people became newly infected with HIV in 2024, 630,000 died from AIDS-related illnesses.

About 87 per cent of all people living with HIV knew their HIV status, and 5.3 million people did not know that they were living with HIV.

The US eliminated measles in 2000, but since 2025, the highly infectious disease has spread to 45 states. As of early May 2026, the US has recorded 1,814 confirmed measles cases this year across 36 states. This follows a record-high 2,288 cases in 2025.

Now, the US Centers for Disease Control (CDC) has warned about additional measles cases ahead of the travel season.

The CDC urged public health agencies across the US to be prepared for more measles cases in the coming months.

"With continued measles transmission in areas across North America and expected increases in international and domestic travel and large events during spring and summer, additional measles cases are anticipated in the coming months," the agency said.

Travel Guidance By CDC

The guidance advised public health agencies to document and report details of each case of measles, including close contacts and locations visited while a person was infectious.

Active surveillance should be conducted to identify any additional suspected cases and quickly transport specimens for laboratory confirmation, the guidance said.

The agency also encouraged outreach to under-vaccinated communities and suggested using state-based syndromic surveillance systems to detect changes in health care–seeking behavior for fever and rash illnesses or signs of vitamin A toxicity.

Measles is a highly contagious virus that spreads through direct contact or through the air when an infected person coughs or sneezes. It is a vaccine-preventable disease that can cause devastating complications, including blindness, pneumonia, encephalitis, and long-term immune dysfunction.

Is The US At Risk Of Losing Measles Elimination Status?

Also read: Robert F. Kennedy Jr. Denies Link To Measles Outbreak At Senate Hearing

So far, there have been 24 new outbreaks reported in 2026, and 93 per cent of confirmed cases (1,688 of 1,814) are outbreak-associated (415 from outbreaks starting in 2026 and 1,273 from outbreaks that started in 2025).

In 2025, 48 outbreaks were reported, and 90 per cent of confirmed cases (2,065 of 2,288) were outbreak-associated.

According to researchers at Boston Children's Hospital, the US has missed four of the seven criteria for measles elimination status, which could put the country off track. More are at risk, they said.

In a Correspondence published in The Lancet, the researchers explained the missed indicators of measles elimination status in the US.

• Low number of cases with a cutoff of less than one case per 10 million people. As of early 2026, the US had about 93 cases per 10 million people, exceeding this limit.

• Most of the measles cases should come from abroad rather than from internal spread. Since the start of 2025, only 6–7 per cent of the measles cases have come from abroad, meaning most cases derive from within the US.

• A limited number of outbreaks (a cutoff of approximately four) with no more than about six cases each. Last year in the US, 48 outbreaks resulted in more than 2,000 cases. And in early 2026, at least 19 outbreaks have already resulted in more than 1,600 cases.

• A level of transmission less than one, meaning one infected person only spreads measles to fewer than one other person on average. The US exceeded this rate more than 75 per cent of the time since early 2025.

Indicators At Risk

Read More: India Concerned Over Measles Outbreak, Action Underway: Dr N K Arora| Exclusive

• The country achieves four weeks with all infections deriving from outside the US. Since the first infection in January 2025, the US hasn't gone four weeks without infections, with 90 per cent of cases acquired here.

• Herd immunity through vaccination. An estimated 95 per cent of people need to receive two doses of the measles vaccine to achieve herd immunity, typically given as part of the MMR (measles, mumps, and rubella) vaccine. However, the US average vaccination rate of kindergartners in the 2024–2025 school year was 92 per cent, meaning this indicator is at high risk.

• Measles cases don't share a common viral strain. From ongoing genetic analyses, preliminary data show that the majority of cases share the same viral strain and are likely part of the same transmission chain.

The researchers believe these findings make a strong case for vaccinating children to protect them from a young age.

"Viral infections aren't all benign, and a measles infection, even when cleared, can result in lifelong problems," said Maimuna Majumder, from Boston Children's.

"Babies less than a year old are among those at greatest risk for severe complications, and the full impact on children exposed during the current outbreak may only show up years later," Majumder added.

Pancreatic cancer is one of the most painful forms of tumour and is also likely to become the second-leading cause of cancer-related deaths in the US by 2030. This could be because 85 per cent of cancer cases in the US are not diagnosed until the disease has spread. Pancreatic cancer is a disease that is usually diagnosed at an advanced stage because there are no prominent early signs. However, a newly developed AI model from the Mayo Clinic and the University of Texas MD Anderson Cancer Center could change that—a new CT scan-based system can help with the timely diagnosis of pancreatic cancer.

Can AI help with pancreatic cancer diagnosis?

REDMOD (radiomics-based early detection model) was tested on CT scans of patients who were later diagnosed with pancreatic cancer. Researchers found that in nearly three out of four cases, REDMOD successfully identified the most common form of pancreatic cancer 16 months before diagnosis. It nearly doubled the detection rate of specialists reviewing scans without AI assistance. In some cases, REDMOD recognised suspicious tissue patterns more than two years before diagnosis. Researchers said that the AI system can detect cancer up to three years in advance.

How does AI work for pancreatic cancer diagnosis?

Experts found that the greatest barrier to saving lives from pancreatic cancer was the inability to detect the tumour when it was still curable. AI can identify cancer signs from a normal-looking pancreas, and it can do so reliably across clinical settings. Researchers used 969 CT scans of the pancreas as training data for REDMOD to help it detect early-stage cancer signs.

Instead of looking for a prominent tumour, the model analysed radiomic patterns that disrupt tissue structure and texture—changes that are too subtle for the human eye to detect. Many cancers begin when normal cells acquire DNA mutations, which affect how cells divide and grow. However, it can take years before these changes develop into a tumour that produces symptoms or becomes clearly visible on a scan.

How accurate was REDMOD at diagnosing pancreatic cancer?

REDMOD was tested on a varied set of CT scans after training — 63 from patients who later developed cancer but were scanned before diagnosis, and 430 from healthy individuals. Out of the 63 cases, REDMOD flagged 46 as suspicious, resulting in a 73 per cent success rate. All these scans had previously been given the all-clear by radiologists who evaluated them at the same time as REDMOD.

Out of the 430 healthy individuals, 81 were identified as suspicious cases by REDMOD. This means that if AI were deployed in a real-world scenario, some individuals might be recommended for additional tests before receiving a final all-clear. A similar performance was observed in two other datasets from different hospitals using different equipment. For patients who had multiple scans available, AI produced consistent results, even when the scans were taken months apart.

Are there any drawbacks of REDMOD?

The sooner REDMOD is implemented in clinical practice, the earlier it can detect pancreatic cancer. It could identify tumours at a stage when treatment is still possible. This could significantly improve survival rates and reduce cancer-related mortality. However, researchers now aim to test the effectiveness of AI in larger and more diverse populations.

The research has been published in the Gut.