Yazılar

U.S. FDA to Roll Out AI Tools Across All Centers Following Successful Pilot

/in /tarafından

The U.S. Food and Drug Administration (FDA) announced it will immediately begin deploying artificial intelligence tools internally across all of its centers, with full integration expected by June 30. The move follows a successful generative AI pilot aimed at supporting scientific reviewers in accelerating the drug review process.

WHY IT MATTERS:
The FDA typically has 6 to 10 months to evaluate a drug approval application. The newly tested generative AI tools are designed to ease the burden on scientists by automating repetitive and time-consuming tasks, thereby streamlining the overall review process and potentially speeding up access to life-saving treatments.

In a statement, the agency emphasized that the focus of future AI enhancements would be on usability, better document integration, and center-specific output customization — all while upholding strict data security and FDA compliance standards.

KEY QUOTE:
“Future enhancements will focus on improving usability, expanding document integration and tailoring outputs to center-specific needs, while maintaining strict information security and compliance with FDA policy,” the FDA said.

CONTEXT:
The announcement comes just a day after Wired reported that the FDA had been in discussions with OpenAI, the maker of ChatGPT, regarding potential AI collaborations. The report also mentioned that representatives from Elon Musk’s Department of Government Efficiency had attended multiple meetings with both the FDA and OpenAI in recent weeks.

WHAT’S NEXT:
The FDA plans to monitor the system’s performance closely, solicit feedback from its users, and refine the tools accordingly. The agency has committed to releasing more information about the AI implementation and its outcomes in June.

This marks one of the most significant government-level adoptions of generative AI to date and could signal a broader shift toward AI-assisted regulatory workflows in the healthcare and pharmaceutical sectors.

Dexcom Introduces AI-Powered Reports for Stelo Glucose Monitor, Offering Personalized Insights

/in /tarafından

Dexcom has launched a new artificial intelligence feature for its Stelo continuous glucose monitor (CGM), giving users a more personalized view of how their meals, sleep, and activities affect their glucose levels. This AI-driven addition, which debuted on Tuesday, is part of Dexcom’s broader efforts to enhance user engagement and provide deeper insights into glucose management.

Stelo, an over-the-counter CGM that monitors real-time blood sugar levels by inserting a small sensor under the skin, was first introduced in August. Unlike traditional CGMs, Stelo is designed for adults who do not take insulin, opening up a new consumer market for the company. This latest feature aims to make Stelo more valuable and personalized for everyday users, with Dexcom focusing on enhancing its capabilities for a broader audience.

Jake Leach, Dexcom’s chief operating officer, emphasized that user feedback had driven the company’s decision to integrate more advanced AI features. “The No. 1 feedback we get is users want to see more,” Leach explained in an interview with CNBC. “They’re making an investment and wearing the product, and they want to be able to take the most advantage of all the data that they’re generating.” The new AI-generated reports reflect this need for deeper, actionable insights.

Dexcom is leveraging Google Cloud’s Vertex AI platform and its Gemini models to build the new AI features. These tools enable developers to synthesize diverse data types, a complex task in healthcare. The company is proceeding cautiously, ensuring the AI platform adds value without compromising the reliability of its core CGM products, which are essential for managing serious health conditions.

While Stelo users have always had access to weekly insights reports, these reports were traditionally formatted in a standard template. The new AI-powered version promises to offer a more personalized experience, drawing on individual user data to provide tailored feedback. For example, if a user’s activity level is low after meals, the report will include specific recommendations to help improve glucose management.

It is important to note that while the AI feature offers personalized insights, it does not provide medical advice. Dexcom has developed the feature with guidance from the U.S. Food and Drug Administration’s (FDA) AI framework. The FDA approved Stelo for use in March, and the company is now looking at expanding the use of generative AI across its other CGM products.

Looking forward, Dexcom plans to evolve its AI capabilities to provide real-time feedback rather than just weekly summaries. The company also envisions using the AI platform to predict potential issues before they arise, much like a “check engine light” in a car, providing early warnings and suggestions for further consultation with healthcare providers.

Chris Sakalosky, vice president of strategic industries at Google Cloud, noted the potential for the technology to offer predictive insights. “It gives you a sense for what could be going on, and recommendations of where you might want to go to seek more advice,” Sakalosky said.

The updated AI-powered reports are already rolling out to Stelo users this week, marking an important step in Dexcom’s mission to make glucose monitoring more intuitive and informative for consumers.

 

Eli Lilly’s Alzheimer’s Treatment Approved in China

/in /tarafından

China’s medical regulator has approved Eli Lilly’s Kisunla, an Alzheimer’s treatment for early-stage disease, offering patients an additional therapeutic option after the approval of Leqembi, developed by Eisai and Biogen, earlier this year. This marks China’s entry into a select group of major markets—including the United States, Japan, and the UK—where Kisunla is approved, Eli Lilly announced on Tuesday.

Kisunla, like Leqembi, targets the clearance of beta-amyloid, a protein linked to Alzheimer’s disease, from the brain. In a late-stage clinical trial, Kisunla demonstrated a 29% reduction in the progression of memory and cognitive decline compared to a placebo. However, the treatment carries safety concerns, with reports of brain swelling in nearly 25% of patients and brain bleeding in approximately 33%, though most cases were deemed mild.

In response to these risks, Kisunla’s U.S. prescribing label includes the FDA’s strongest “boxed” safety warning, similar to Leqembi’s label. Lilly has addressed concerns by introducing a gradual dosing schedule, which has been shown to reduce the likelihood of severe brain swelling.

A key distinction between Kisunla and Leqembi lies in their dosing approach. Kisunla offers finite dosing, allowing patients to discontinue the treatment once brain scans confirm the absence of amyloid plaques, whereas Leqembi does not.

Kisunla is currently under review by the European Union’s drug regulator. In contrast, Leqembi was rejected by the EU earlier this year due to concerns that the risks of brain swelling outweighed its modest benefits in slowing cognitive decline.

According to the World Health Organization, Alzheimer’s is the leading cause of dementia, accounting for 60%-70% of cases globally.