---
id: "2026-10-19-fahr-face-cancer-survival-face-photos"
url: "https://postcutoff.com/e/2026-10-19-fahr-face-cancer-survival-face-photos/"
as_of: "2026-10-20T07:28:00+02:00"
date: "2026-10-19"
date_precision: day
category: science
importance: 3
confidence: medium
status: [Result confirmed]
verification: Peer-reviewed
sources: 2
editor: human editor
human_review: null
version: null
---

As of: 2026-10-20 07:28 CEST.
Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/
Canonical page: https://postcutoff.com/e/2026-10-19-fahr-face-cancer-survival-face-photos/

# FAHR-Face: Mass General Brigham foundation model estimates biological age and cancer survival from a face photo (Lancet Digital Health)

A study published on Oct 19, 2026 in Lancet Digital Health presents FAHR-Face, a foundation model trained on about 40 million face images and adapted to estimate biological age and survival of cancer patients from a single photo. Of the 122 highest-risk patients in a 3,820-patient test group, 64 died within 90 days of their photo. Researchers say it is not ready for clinical use; an ethicist warned about insurers and employers.

## Key facts
- Name: FAHR-Face (foundation AI models for health recognition using face photographs); team led by radiation oncologist Ray Mak, first author Fridolin Haugg (Mass General Brigham)
- Pre-training: ~40 million face images without a specific task, then adapted to biological age and cancer survival
- Survival test: 3,820 cancer patients; 64 of the 122 with the highest risk scores died within 90 days of the photo; age findings used >38,000 cancer patients
- Adding face-age and face-survival scores to a clinical prognostic model predicted survival better than either alone
- Model attention: age on the nasolabial folds and forehead; death risk on the bridge of the nose and below the eyes (not explained)
- Robust to expressions, makeup and lighting; rhinoplasty made people look older after surgery
- Limits: trained mainly on a majority-white Boston-area population; patent pending; Mak: 'There's still a lot of development ahead and clinical testing'
- Matthew Liao (NYU, not involved): 'If it gets it wrong, we might be sort of not providing care to someone who actually needs it'

## What happened

Mass General Brigham's team pre-trained a general face model and then adapted it to two health tasks. The figures above come from
Scientific American's report; the Guardian article and the paper were not readable for us.

## Why it matters

A camera-only prognostic signal is cheap to collect and could help clinicians, but it also raises the risk of face photos being used
to judge people's health by insurers or employers, as outside experts warned.

## Your AI and this story

- GPT-6 Astra (training cutoff April 2026): 172 days after its cutoff
- Claude Opus 5.5 (training cutoff June 2026): 111 days after its cutoff
- Gemini 4 Argon (training cutoff March 2026): 202 days after its cutoff
- Grok 4.7 (training cutoff May 2026): 141 days after its cutoff

## Sources

1. [Scientific American: Can AI predict whether you'll die from cancer from just an image of your face?](https://www.scientificamerican.com/article/can-ai-predict-whether-youll-die-from-cancer-from-just-an-image-of-your-face/) (scientificamerican.com, press)
2. [The Guardian: New AI tool can predict cancer survival by a single patient selfie, study suggests](https://www.theguardian.com/technology/2026/oct/19/ai-tool-predict-cancer-survival-patient-selfie) (theguardian.com, press)

## Changes

- 2026-10-20 (filed): Created (Google News; Scientific American read)