AI Reads 6 Chest CTs at Once: 8.74M Tokens, 85 Calls, 72 3D Calipers Reveal What Reports Missed
npinto · x · 2026-10-30
A striking real case: a patient with 28 lung nodules across 6 chest CTs over 4 years had each report simply say 'stable.' Running Gemini 4 Argon across all six scans at once (8.74M tokens, 85 calls, 72 3D calipers) produced a longitudinal measurement table and curves showing a biphasic post-hysterectomy regression of 12.3%–23.9% — insight invisible in scan-to-scan text reports.
More from Apps
- Cited but not recommended: brands left off ChatGPT lists ~2/3 of the time — sanderssays · 2026-10-30
- Dev: An agent that automates expense reports in existing enterprise software beats another booking app — gethackteam · 2026-10-30
- Procedural idea generation surprises 1 in 10 times, recombination is underrated — Afinetheorem · 2026-10-30
- Wabi's High-Quality Widgets Can Now Be Shared With Friends — joonasvirtanen · 2026-10-30
- YC amplifies folk: AI texting assistant adds 170,000 users in under a week — ycombinator · 2026-10-30
- Screendrop v0.37 lets Claude Code and other MCP agents edit and export your screen recordings — craigsdennis · 2026-10-30