AI in Healthcare

Today's Stories

1. 1

   Japanese startup to test pig kidney transplants in humans by 2028

   PorMedTec, a startup spun out of Meiji University, announced it will conduct clinical trials transplanting pig kidneys into patients at two hospitals—Hokkaido University Hospital in Sapporo and Shonan Kamakura General Hospital in Kamakura—as early as 2028. The pigs are genetically engineered by U.S. biotech firm eGenesis and have undergone 69 gene edits to suppress immune rejection and reduce disease transmission risk. Japan faces a severe organ shortage, with more than 300,000 people on dialysis and roughly 15,000 waiting for a kidney transplant, while only about 200 transplants from brain-dead donors occur annually. Cross-species transplants could significantly expand access to organs. The Japanese government has designated this field as a key investment priority in its public-private roadmap.

   Four clinical trials in the United States have already shown promise, with patients reportedly avoiding dialysis treatment for around nine months at the longest. PorMedTec aims to secure production and marketing authorization after confirming safety in the Japanese trials.

2. 2

   Medical robots & AI take center stage at London festival debate

   A live podcast recording at the Great Exhibition Road Festival in London brought together three leading UK academic researchers—George Mylonas (Imperial College London), Antonia Tzemanaki (University of Bristol), and Tom Vercauteren (King's College London)—to discuss robotics and AI in medicine and healthcare, more than 40 years after the first robot-assisted surgery. As medical tools become increasingly autonomous, the conversation addresses critical questions about ethics, regulation of technologies that can learn and change over time, and fair access to cutting-edge medical devices. These concerns directly shape how hospitals and healthcare systems can safely adopt and deploy new surgical and diagnostic tools.

   The researchers cover a range of surgical applications—from robotic needle steering and tumor detection to rehabilitation and wearable robotics for physical therapy. Tom Vercauteren's spin-out company, Hypervision Surgical, already has its optical imaging technology in use across hundreds of hospitals worldwide, showing one path from research to clinical practice.

3. 3

   How FDA Clearance of AI Opioid-Respiratory Monitoring Could Shape Danaher's (DHR) Clinical Data Advantage

   How FDA Clearance of AI Opioid-Respiratory Monitoring Could Shape Danaher's (DHR) Clinical Data Advantage

4. 4

   Stanford deploys AI 'scientist' agents to accelerate drug discovery

   A Stanford team led by associate professor James Zou has deployed thousands of autonomous AI agents in a virtual biotech environment that simulates the full lifecycle of drug development, handling everything from initial discovery through safety testing and clinical trial design. Drug discovery is notoriously inefficient, with 90% to 95% of projects failing and a single successful drug taking over a dozen years and up to $1 billion（約1600億円） from initial discovery to patient distribution. These agentic AI agents maintain continuity across disconnected workflows, addressing a core source of knowledge loss and delay in traditional development.

   The research will be discussed at VB Transform 2026, where the potential to reshape how pharmaceutical projects move from discovery to patient distribution may become clearer.

5. 5

   AWS shows how to build multi-tenant AI agents safely

   AWS published a blog post demonstrating how to build multi-tenant AI applications using Amazon Bedrock AgentCore, with a healthcare example that implements two service tiers—Basic (using Mistral Ministral 3 8B Instruct for small clinics) and Premium (using OpenAI GPT OSS 120B with web search for hospitals and specialty centers). Multi-tenant AI systems face real operational risks: customer data exposure, inconsistent service quality across pricing tiers, and hidden cost overruns. This post addresses those challenges by showing how to enforce complete tenant isolation through document scoping, memory separation, model access control, and granular cost attribution—all without building custom isolation infrastructure.

   The solution uses a pool model where tenants share underlying compute resources (rather than dedicated silos), maximizing efficiency while maintaining logical isolation through scoped identifiers, access policies, and data partitioning. Sample code is available on GitHub at https://github.com/aws-samples/sample-agentcore-and-multitenancy-blog.

6. 6

   Insilico Medicine, SK Bio ink $2.5B AI drug discovery pact

   Insilico Medicine and SK Biopharmaceuticals struck a $2.5B deal to collaborate on AI-powered drug discovery, with a focus on neuroimmune therapies. The partnership will combine Insilico's AI platform with SK Bio's pharmaceutical expertise and development capabilities. The deal signals growing confidence in AI's ability to accelerate drug discovery, a traditionally lengthy and expensive process. For pharmaceutical companies, working with AI specialists may reduce the time and cost needed to bring new treatments to market, potentially making drug development more accessible.

   The collaboration targets neuroimmune therapies, a specialized area within drug development. Insilico's co-CEO framed the ambition as aiming to become "the SpaceX of the pharmaceutical industry," suggesting the partnership aims to demonstrate transformative efficiency gains in how new drugs are discovered and developed.