🤖

AI Advances in Healthcare

Sep 30, 2025

Overview

The transcript reviews key advances in AI-driven healthcare, focusing on Epic Systems' health data infrastructure, recent AI startups enhancing benefits navigation and cancer detection, and a noteworthy AI-enabled Alzheimer’s discovery. It highlights the growing impact of AI in streamlining administration, improving diagnostics, and uncovering deep biological mechanisms through big data.

Epic Systems: The Healthcare Tech Giant

  • Epic Systems runs MyChart, supporting 191 million Americans and powering backend operations for most US hospitals.
  • The company remains private, never having lost a customer and eschewing acquisitions or VC funding.
  • Epic’s founder, Judy Faulkner, started the company with minimal capital and maintains long-term ownership via a trust.
  • Epic’s Cosmos database aggregates anonymized clinical data from 295M patients and 15B encounters for large-scale analysis.

AI Startups Transforming Healthcare Administration and Diagnostics

  • Healthee uses AI to answer employee benefits questions and streamline open enrollment, raising $50M in Series B funding.
  • Healthee’s platform helps manage rising employer healthcare costs and serves 15K+ customers, including notable companies.
  • Craif leverages AI to detect cancer via urine microRNA analysis, raised $22M, and is aiming for US clinical trials by late 2026.
  • Craif’s test can potentially identify seven cancer types non-invasively before symptoms appear, with high compliance rates.

AI and Advanced Disease Understanding

  • The next leap is AI linking molecular biology data with large-scale clinical records to uncover mechanisms beyond human reach.
  • UCSD researchers used AI protein-folding models to reveal a hidden DNA-binding role in the Alzheimer’s biomarker PHGDH.
  • The study identified a repurposed molecule, NCT-503, improving memory and anxiety in mouse Alzheimer’s models.

Cosmos Database: Research and Predictive Modeling

  • Cosmos enables research on disease causes, treatment patterns, and public health surveillance by offering an unprecedented dataset size.
  • Researchers have used Cosmos to study rare diseases, maternal outcomes, and track disease trends like COVID.
  • The database supports AI-driven predictive modeling, with hospitals deploying algorithms to identify at-risk patients.
  • Epic develops AI solutions ("Cosmo AI") leveraging Cosmos—integrating predictive tools into MyChart and clinician workflows.

Limitations and Future Integration

  • Cosmos primarily contains clinical rather than molecular/genomic data, limiting some AI-driven discoveries.
  • The true revolution lies in linking Cosmos with genomics, proteomics, and biobank datasets, enabling richer AI-driven insights.