Healthcare AI News — Diagnostics, Drug Discovery & More
AI in medicine, diagnostics, and drug discovery, with an eye on practical applications for healthcare businesses.
Medical AI, diagnostics, drug discovery, and clinical AI tools — including FDA approvals and the AI reshaping healthcare and biotech.
PwC: Healthcare Costs Set to Jump 9% in 2027, Partly Driven by AI-Enhanced Medical Coding
Health plans surveyed by PwC are projecting a 9% rise in commercial healthcare costs in 2027 — the highest medical cost trend in nearly two decades. One contributing factor identified is AI-enabled documentation and coding tools that allow providers to capture greater billing specificity and reimbursable severity without proportionate increases in actual care intensity. Drug prices and ongoing regulatory pressures from the No Surprises Act are also cited as drivers.
See the business automation angle on this story — free with signup.
Abridge Receives Strategic Investment from Eli Lilly, Expands Into Payer and Prior Authorization Workflows
AI clinical documentation startup Abridge announced a strategic investment from pharma giant Eli Lilly — financial terms undisclosed — to support evidence-based care and research. Abridge is also expanding its platform to connect real-time clinical documentation with payer claims workflows, and is co-designing an AI-powered prior authorization solution with Highmark Health. Separately, Nvidia is developing a new AI healthcare model in partnership with Abridge, expected to be ready later this year.
See the business automation angle on this story — free with signup.
Nature Study: General-Purpose LLMs Outperform Specialized Clinical AI Tools on Medical Benchmarks
A peer-reviewed study published in Nature Medicine tested frontier LLMs (GPT-5.2, Gemini 3.1 Pro, and Claude Opus 4.6) against specialized clinical AI tools (OpenEvidence and UpToDate Expert AI) across 500 medical knowledge questions, 500 clinical alignment items, and 100 real physician queries. Frontier general-purpose models outperformed the specialized clinical tools in all three evaluations, with the clinical tools performing comparably to Google Search's AI Overview. The findings challenge the assumption that domain-specific training or RAG automatically makes clinical AI superior.
See the business automation angle on this story — free with signup.
FDA AI Use Surged 148% in FY2025, Leading All HHS Agencies in Adoption Growth
A new Bipartisan Policy Center report found that AI use cases across the Department of Health and Human Services jumped sharply between fiscal years 2024 and 2025, with the FDA leading at 148% growth year-over-year. The CDC saw an 87% increase, CMS a 78% jump, and NIH a 51% rise in AI use cases. The data underscores accelerating federal health agency adoption of AI even as policy debates around oversight continue.
See the business automation angle on this story — free with signup.
Novartis Partners with Orionis Biosciences to Use AI-Driven Platform for Molecular Glue Drug Discovery
Orionis Biosciences has announced a multi-year collaboration with Novartis to discover and design molecular glue drugs using Orionis's Allo-Glue platform and its AI-driven discovery engine, targeting challenging disease areas with traditionally hard-to-drug proteins. The deal is structured around AI-accelerated target profiling and drug optimization. It represents another major pharma company embedding AI into its core drug discovery pipeline through external partnerships.
See the business automation angle on this story — free with signup.
Vega Health Licenses Clinical Risk Prediction AI Models from Parkland Center for Clinical Innovation
Health AI startup Vega Health has partnered with the Parkland Center for Clinical Innovation to license five clinical AI models — covering areas such as patient risk prediction, population health, and social determinants of health — onto the Vega Health Marketplace. The partnership is designed to commercialize high-value clinical AI that wouldn't be viable as standalone companies. The move reflects a growing marketplace model for healthcare AI where vetted models are distributed across health systems.
See the business automation angle on this story — free with signup.
Penn Engineers Release FireANTs, an Open-Source AI Algorithm Hundreds of Times Faster for Medical Image Analysis
Researchers at the University of Pennsylvania developed FireANTs, an open-source algorithm that combines AI speed with geometric precision to perform dense correspondence matching in complex medical images — a key step in detecting subtle disease progression across MRI or CT scan sets. The algorithm runs hundreds to thousands of times faster than its predecessor ANTs with no loss in accuracy, making real-time clinical use feasible for the first time. The researchers published their findings in Nature Communications and released the tool publicly.
See the business automation angle on this story — free with signup.
Philips Survey: AI Helping Clinicians Catch Medical Errors and See More Patients
A Philips-sponsored survey found that 27% of clinicians reported AI had helped them identify possible medical errors at least three times in the past three months, while 36% said AI tools increased the number of patients they could see weekly. The company's North America CEO noted that healthcare professionals are currently using AI most heavily for administrative tasks like scheduling and data compilation, with more complex clinical decisions remaining in human hands. The findings point to AI-driven labor productivity gains becoming a structural feature of healthcare delivery.
See the business automation angle on this story — free with signup.
FDA's Own AI Review Tool Struggles with Basic Tasks, Raising Concerns
NBC News reported in June 2025 that the FDA's internally developed AI tool, designed to speed up reviews and approvals of medical devices like pacemakers and insulin pumps, is struggling with basic tasks according to people familiar with the project. The tool was intended to accelerate the agency's famously slow approval process by automating document review, but early performance has fallen short of expectations. The news raises broader questions about whether AI can reliably handle high-stakes regulatory analysis at this stage.
See the business automation angle on this story — free with signup.
Radiology AI Dominates FDA Approvals with Over 1,039 Cleared Imaging Tools
The FDA has now cleared over 1,039 AI tools specifically for clinical imaging, representing nearly 80% of all AI-enabled medical devices authorized by the agency to date. A recent December 2025 update added 56 new radiology-specific devices from companies including GE HealthCare, Siemens, Fujifilm, and Qure.ai. The rapid pace of approvals reflects both the maturity of medical imaging AI and the FDA's reliance on the 510(k) substantial-equivalence pathway, which avoids the need for new clinical trials.
See the business automation angle on this story — free with signup.
FDA Updates AI-Enabled Medical Device List, Radiology Maintains Dominant Lead
The FDA updated its list of cleared AI-enabled medical devices through December 2025, with radiology maintaining the highest number of cleared devices — surpassing 1,000 authorizations. Nearly all cleared devices entered via the 510(k) substantial equivalence pathway, avoiding the need for costly new clinical trials. The FDA has also signaled plans to tag devices incorporating foundation models and LLMs in future tracker updates.
See the business automation angle on this story — free with signup.
HHS Issues Sweeping Request for Information on Accelerating AI in Clinical Care
The U.S. Department of Health and Human Services released a broad Request for Information asking how the federal government can use regulatory, reimbursement, and R&D levers to speed up AI adoption in clinical settings — described as the broadest federal action on healthcare AI to date. Separately, ARPA-H is soliciting proposals for agentic AI clinical assistants, hoping to set a new FDA regulatory precedent for generative AI in high-risk medical environments. These moves suggest the federal government is preparing to actively fund and shape AI's role in healthcare.
See the business automation angle on this story — free with signup.
FDA's Internal AI Tool for Speeding Device Approvals Struggles With Basic Tasks
The FDA developed an internal AI tool designed to accelerate reviews and approvals of medical devices like pacemakers and insulin pumps, but two people familiar with the system say it is struggling with simple tasks. NBC News reported concerns from insiders about the tool's reliability, raising questions about whether AI can yet handle the nuanced demands of regulatory review. The issues highlight the gap between AI ambition and operational readiness in high-stakes government contexts.
See the business automation angle on this story — free with signup.
FDA Clears Over 1,450 AI-Enabled Medical Devices Through 2025, Radiology Leading
The FDA has now authorized more than 1,451 AI and machine learning-enabled medical devices through the end of 2025, with 295 cleared in 2025 alone and radiology maintaining a dominant lead in approvals. The updated list includes standalone software and imaging hardware with embedded AI for detecting conditions ranging from diabetic retinopathy to lung nodules. The pace of clearances signals growing regulatory confidence in AI-powered clinical tools.
See the business automation angle on this story — free with signup.
FDA's Own AI Approval Tool Struggling with Basic Tasks, NBC Reports
NBC News reported that the FDA's internally developed AI tool — intended to speed up reviews and approvals of medical devices like pacemakers and insulin pumps — is struggling with simple tasks according to two people familiar with its performance. The tool was designed to reduce the workload on FDA reviewers as the volume of AI-enabled device submissions surges. The early struggles highlight the gap between AI promise and production-ready performance even inside regulatory agencies.
See the business automation angle on this story — free with signup.
Radiology AI Dominates FDA Approvals, Topping 1,039 Cleared Imaging Devices
The FDA added 56 new AI-powered radiology tools in a single December update, bringing the total number of approved AI imaging devices to 1,039 — nearly 80% of all AI-enabled medical technologies the agency has ever authorized. Major approvals came from GE HealthCare, Siemens, Fujifilm, Qure.ai, and DeepHealth. The FDA has also issued new draft guidance covering the full product lifecycle for AI-enabled devices, from design through post-market monitoring.
See the business automation angle on this story — free with signup.
FDA's AI-Powered Internal Review Tool Struggles With Basic Tasks, Raising Approval Process Concerns
A new FDA AI tool designed to speed up the review and approval of medical devices — including pacemakers and insulin pumps — is reportedly struggling with simple tasks, according to people familiar with the system. NBC News reported that while the tool is intended to accelerate the agency's historically slow approval pipeline, its early performance has raised doubts among insiders about its readiness. The FDA has now authorized over 1,450 AI-enabled medical devices, with 295 cleared in 2025 alone.
See the business automation angle on this story — free with signup.