AI Trained on 40M Face Photos Predicts Cancer Survival From a Single Selfie, per Lancet Digital Health
EricTopol · x · 2026-10-20
Eric Topol follows up on a new Lancet Digital Health study: a team at Mass General Brigham trained a face foundation model on over 40 million facial photographs across ethnicities and ages. The model can extract extensive health signals from a single ordinary photo, including cancer outcomes, biological age, lifestyle effects, and survival.
It spawned two complementary tools: FAHR-FaceAge estimates a patient's biological age (how old they look rather than their chronological age), while another directly predicts cancer patient survival odds. It's one of the most comprehensive demonstrations yet of assessing health from a selfie, as reported by The Guardian.
More from Apps
- Grok Bot ships wave of updates: voice calls, X DMs, connectors and more — mattyp · 2026-10-20
- AI agent runs a real car-rental startup's ads with $25,000 for 30 days, in public — socialwithaayan · 2026-10-20
- AI-designed viruses used to treat a human patient for the first time, targeting drug-resistant E. coli — Polymarket · 2026-10-20
- TapMac turns taps on your MacBook into voice dictation and AI text fixes, $7.99 one-time — sharpeye_wnl · 2026-10-20
- Taliri is a $49/yr Chrome extension that teaches languages inside pages you already read — theteknosaur · 2026-10-20
- Profound partners with Moz to show SEO and AEO data side by side per page — gaganghotra_ · 2026-10-20