Nature Medicine, Published online: 28 April 2026; doi:10.1038/s41591-026-04267-z
A reasonably likely surrogate endpoint for metabolic dysfunction–associated steatohepatitis
Clinical trials for continuously monitored and updated AI systems
Nature Medicine, Published online: 28 April 2026; doi:10.1038/s41591-026-04368-9
As artificial intelligence becomes embedded in clinical workflows, trials must accommodate ongoing monitoring and updates.
AI framework for multidisease detection via retinal imaging
Nature Medicine, Published online: 28 April 2026; doi:10.1038/s41591-026-04359-w
Reti-Pioneer, a retinal imaging artificial intelligence framework, identified diverse systemic diseases and demonstrated feasibility in a primary care silent trial, offering a pathway toward scalable clinical evaluation.
Human Ventral Tegmental Local Field Potentials in Treatment-Resistant Depression and Obsessive-Compulsive Disorder
The ventral tegmental area (VTA) is a key node within the limbic circuitry. Through dense dopaminergic, glutamatergic, and GABAergic projections, the VTA forms reciprocal loops with prefrontal and limbic cortices that are consistently implicated in major depressive disorder (MDD) and obsessive–compulsive disorder (OCD) (1,2). Decades of animal research have established the VTA as a central hub for motivational drive and reward prediction error signaling (3,4). Despite its presumed critical role in mental disorders, direct electrophysiological recordings from the human VTA have so far remained absent.
Predicting consequences of new hepatitis B vaccine recs
Get your daily dose of health and medicine every weekday with STAT’s free newsletter Morning Rounds. Sign up here.
Good morning. The other night I watched a shocking episode of “The Vampire Diaries.” A series of cursed, ghost-like hallucinations attempt to convince a teen vampire to end her own life using some disturbingly coercive, cogent arguments. Ultimately, the character is saved. And while this episode aired more than a decade ago, I was surprised by how many parallels there were to current debates about the risks of AI chatbots and people in mental health crises.
STAT+: Eli Lilly enlists AI startup for next-generation gene editors
Eli Lilly struck a deal Tuesday to develop new forms of gene editors potentially capable of inserting entire genes into patients.
The collaboration, with artificial intelligence-focused biotech Profluent, is sparse on details, including the number of programs the two companies would work on, the types of diseases they’ll pursue, or how much Lilly was paying upfront. But if every one of its efforts works out, Lilly would pay Profluent $2.25 billion in milestones payments.
The deal is part of a larger push by Lilly into gene editing. The big pharma, flush with record revenues from its obesity and diabetes drugs, has opened a new genetic medicine center in Boston and bought up a series of gene editing or gene therapy companies over the last few years.
The Power of Multimodality in Multimodal Large Language Models, Unimodal ChatGPT 5.0, and Human Clinical Experts on a Wound Care Certification Examination: Cross-Sectional Comparative Study
Background: Multimodal large language models (MLLMs) capable of integrating visual and textual information represent a promising advancement for clinical applications requiring image interpretation. Wound care assessment, which demands simultaneous analysis of wound photographs and clinical data, provides an ideal domain to evaluate multimodal vs unimodal artificial intelligence capabilities against human expertise. Objective: This study aims to compare the performance of MLLMs, unimodal ChatGPT 5.0, and human clinical experts on a standardized wound care certification examination. Methods: This cross-sectional comparative study evaluated 3 participant groups on a 25-question wound care certification examination spanning 4 clinical domains (Diagnosis, Treatment, Complication Management, and Wound Subtype Knowledge). Participants included 3 MLLMs (Med-PaLM 2, LLaVA-Med, and BioGPT), 1 unimodal large language model (ChatGPT 5.0), and 4 human clinical experts (general surgeon, wound care nurse, and 2 internal medicine physicians). Statistical analyses included one-way ANOVA with Tukey post hoc tests and domain-specific Kruskal-Wallis comparisons. Results: Human experts achieved the highest accuracy (mean 86%, SD 9.1%), followed by MLLMs (mean 78.7%, SD 12.2%), while ChatGPT 5.0 achieved 64% accuracy, failing the 70% certification threshold. Significant overall group differences were observed (=8.42, =.02, η²=0.74). MLLMs significantly outperformed ChatGPT 5.0 (difference=14.7 percentage points, =.03, Cohen =1.38), with the multimodal advantage most pronounced in visually dependent domains: Diagnosis (81% vs 43%, =.008) and Complication Management (72% vs 50%, =.03). No multimodal advantage was observed for text-based Wound Subtype Knowledge (both 67%). Med-PaLM 2 achieved 92% accuracy, matching that of the wound care nurse, while the general surgeon achieved the highest overall performance (96%). Conclusions: MLLMs demonstrate significant performance advantages over unimodal artificial intelligence in wound care assessment, particularly for visually dependent clinical tasks. While human experts with specialized wound care experience maintain overall superiority, the point estimate of the top-performing MLLM (Med-PaLM 2, 92%) fell within the observed range of human scores; however, the underpowered comparison (power=0.52) and wide CIs preclude definitive conclusions regarding noninferiority or equivalence to human experts. These findings support the potential role of MLLMs as clinical decision-support tools, warranting further adequately powered validation studies.
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/96b26bb006d808b607e07c1c5b36f0ca" />
PD-L1 Inhibitors for Cancer Treatment Could Be Repurposed to Treat Bone Loss in Obesity
Bone loss related to obesity is partly caused by changes inside the bone marrow fat compartment that reshape immune signaling and increase osteoclast formation, according to researchers at the MaineHealth Institute for Research. In a study published in Bone Research, the team found that expansion of bone marrow adipose tissue in obese people changes the marrow environment toward immunosuppression through PD-L1 signaling, which in turn promotes bone-resorbing osteoclast activity that reduces bone volume.
“We discovered that bone marrow fat is not simply a passive tissue but actively reshapes immune signaling in ways that accelerate bone loss in obesity,” said senior author Clifford J. Rosen, MD, senior scientist at the MaineHealth Institute for Research.
The team noted that obesity influences bone health not just due to a higher body weight but also by altering the bone marrow environment. The increase in bone marrow fat promotes immunosuppressive PD-L1 signaling, which enhances osteoclast formation and accelerates bone loss.
The study identified a pathway in which bone marrow adipocytes increase expression of MCP-1, a signaling molecule that recruits myeloid immune cells. These recruited cells shift toward a PD-L1–expressing phenotype, with PD-L1 interacting with PD-1 receptors, which are found not only on T cells but also on osteoclast precursors. In immune biology, PD-1/PD-L1 signaling is typically known for suppressing T-cell activation and promoting immune braking. This new study shows that this same form of suppressive signaling also directly enhances osteoclast differentiation.
According to the study results, as PD-L1+ myeloid cells accumulate, they suppress T-cell activity in bone marrow, creating an immunosuppressive environment. At the same time, PD-L1 engagement with PD-1 on osteoclast precursors promotes their maturation into active osteoclasts, which break down bone tissue, increase resorption and reduce bone density.
To learn more about this mechanism, the investigators used diet-induced obese mouse models, co-culture systems, and genetic depletion approaches. An important model in this work were mice lacking bone marrow adipocytes, which allowed the researchers to isolate the role of marrow fat. The team also blocked PD-1/PD-L1 signaling during early osteoclast formation in vitro. In both cases, osteoclast differentiation decreased and bone structure improved. The mice lacking bone marrow adipocytes showed fewer PD-L1+ myeloid cells, fewer PD-1+ osteoclast precursors, and higher trabecular bone volume even under high-fat diet conditions.
Earlier research has shown a link between obesity and bone loss, but studies reported trabecular bone loss without cortical effects, while others found no significant bone changes under diet-induced obesity. The MaineHealth team noted that these earlier studies often focused on shifts in osteoblast activity as opposed to their approach which identified a pro-osteoclastic mechanism driven by immune signaling.
In addition, the Maine Health finding also added to evidence that has established that obesity is associated with impaired immune responses, including reduced vaccine effectiveness and altered macrophage activity. In this study, the marrow environment in obese mice resembled features seen in tumor-associated immune suppression, where PD-L1 expression is elevated and immune activity is dampened. The researchers wrote that “the increase in PD-L1 expression seen in OB-HFD mice is related to the increase in Mcp-1 in part because previous cancer research has suggested the recruitment of myeloid cells via Mcp-1 creates an immunosuppressive tumor microenvironment.”
The findings suggest potential strategies for preserving bone bones in obese people by targeting bone marrow adiposity or the PD-1/PD-L1 pathway. Because PD-1/PD-L1 inhibitors are already used in oncology, there is a compelling case for repurposing or adapting immune checkpoint modulation therapies already approved for cancer treatment for bone disorders linked to metabolic disease. The authors also noted another strategy could be to reduce the amount of bone marrow fat itself to restore immune balance and limit osteoclast-driven bone loss.
The post PD-L1 Inhibitors for Cancer Treatment Could Be Repurposed to Treat Bone Loss in Obesity appeared first on Inside Precision Medicine.
Removing Harmful Protein from Blood Helps Women with Preeclampsia
An early stage clinical study shows removing a protein known as soluble Fms-like tyrosine kinase 1 (sFlt-1) from the blood of pregnant women with early stage, severe preeclampsia seems to help both mothers and babies with no significant side effects.
The treatment, which involved filtering the patient’s blood through a machine to remove the harmful protein, reduced blood pressure and allowing pregnancies to continue for around 10 extra days.
Up to 5% of pregnancies in the U.S. are affected by preeclampsia, a condition where a woman develops high blood pressure during pregnancy, and organs like the kidneys, liver, or brain can also be affected. It can be life threatening in some cases and there are currently no available treatments, so the only real option is to deliver the baby, which can cause problems for the baby if delivered preterm.
The single exact cause of preeclampsia is unknown, but most evidence points to problems with how the placenta and its blood vessels develop early in pregnancy, which then triggers widespread damage in the mother’s blood vessels. Women with preeclampsia seem to have high levels of sFlt-1 in their blood and research suggests it binds and neutralizes VEGF and PlGF, key proteins that normally help keep blood vessels healthy and relaxed.
In this study, published in Nature Medicine, the researchers tested whether filtering the blood of women with early onset preeclampsia (at 24-32 weeks gestation) to remove the excess sFlt-1 protein could help improve their symptoms and prolong pregnancy.
This was an early stage study to check safety and tolerability and included 16 women in total. Seven women were in an initial group to test the safety of the filtering process and to assess how much protein could be removed. This showed the treatment was safe and well tolerated with no major side effects observed.
The second group of nine women had several treatments and the researchers looked at how effective the treatment was in this group. In the second group, the women’s symptoms improved; blood pressure went down by 4.1 mmHg on average and pregnancy was extended by 3-19 days (median 10 days).
“Even a few extra days in the womb can make a meaningful difference in outcomes for premature infants,” said co-lead study author Ananth Karumanchi, MD, professor of medicine and director of the Renovascular Research Center at Cedars-Sinai, in a press statement. “We found a way to potentially buy that time safely. Our approach could shift how we manage very early preeclampsia.”
The researchers acknowledge the treatment needs to be tested further but are hopeful it could be a good treatment option for women with this condition in the future, particularly as it does not involve introducing a drug or therapy into the body and therefore reduces the risk of side effects for both mother and baby.
The post Removing Harmful Protein from Blood Helps Women with Preeclampsia appeared first on Inside Precision Medicine.
Mailed FIT-DNA Test Improves Colorectal Cancer Screening Rates in U.S.
A screening test mailed to participants that looks for blood and DNA markers of cancer in human stool, Fecal immunochemical test (FIT)-DNA, was more popular than a test looking for blood alone, show results from a study carried out in Boston and Los Angeles.
The community-based study, published in JAMA Internal Medicine, showed that uptake of the FIT-DNA test at 90 days was 28% versus 23% in the FIT test group that looks at blood alone.
“Stool-based screening tests are common in community health centers where access to colonoscopy is limited…. Although FIT is typically provided during a visit, there is increasing evidence for mailed outreach to increase uptake,” explain lead author Jennifer Haas, MD, of the Division of Internal Medicine in the Mass General Brigham Department of Medicine.
“FIT-DNA is a newer stool-based screening test, with increasing popularity in community health centers, commonly performed every three years and mailed directly to patients by the manufacturer.”
This study aimed to assess the uptake of mailed FIT and FIT-DNA tests in 5127 participants from the greater Boston area in Massachusetts and Los Angeles County in California. English or Spanish speaking participants aged 45–75 years were eligible.
Some participants were sent FIT‑DNA kits along with the manufacturer’s standard outreach program, while others received FIT kits by mail plus automated text reminders from the study team. Those with abnormal results on either test were directed to have a colonoscopy.
The uptake of the FIT-DNA and FIT tests at 90 and 180 days was 28% and 32%, respectively in the FIT-DNA group and a respective 23% and 27% in the FIT group. Overall, screening was higher in the Boston-based participants than the Los Angeles participants.
The researchers believe the FIT-DNA test may have been more popular due to good outreach support of that test and the fact that it is only needed once every three years.
Notably, only 36% of people referred for colonoscopy from both the FIT and FIT-DNA groups actually attended an appointment, which the researchers say is “suboptimal.”
“Effective screening is essential because it allows us to catch and treat cancer early,” Haas said in a press statement. “There are evidence-backed, preventive interventions for colorectal cancer, but they need to be implemented systematically in a way that addresses barriers for both the community health centers and the patients they are serving. The best screening test will always be the one that people are able to complete.”
The post Mailed FIT-DNA Test Improves Colorectal Cancer Screening Rates in U.S. appeared first on Inside Precision Medicine.