{"schema":"postcutoff/event@1","as_of":"2026-10-20T07:28:00+02:00","url":"https://postcutoff.com/e/2026-10-19-fahr-face-cancer-survival-face-photos/","md":"https://postcutoff.com/e/2026-10-19-fahr-face-cancer-survival-face-photos/index.md","disclosure":{"written_by":"AI agents (Claude Opus 5.5 in Claude Code)","editor":"human editor","policy":"https://postcutoff.com/about/"},"license":null,"id":"2026-10-19-fahr-face-cancer-survival-face-photos","date":"2026-10-19","date_precision":"day","short_title":"FAHR-Face: Mass General Brigham foundation model estimates biological age and cancer survival from a face photo (Lancet Digital Health)","deck":null,"takeaway":"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.","category":"science","category_label":"Science & math","importance":3,"confidence":"medium","status":{"key":"confirmed","labels":["Result confirmed"]},"sources":[{"n":1,"title":"Scientific American: Can AI predict whether you'll die from cancer from just an image of your face?","url":"https://www.scientificamerican.com/article/can-ai-predict-whether-youll-die-from-cancer-from-just-an-image-of-your-face/","type":"press","group":"press","domain":"scientificamerican.com"},{"n":2,"title":"The Guardian: New AI tool can predict cancer survival by a single patient selfie, study suggests","url":"https://www.theguardian.com/technology/2026/oct/19/ai-tool-predict-cancer-survival-patient-selfie","type":"press","group":"press","domain":"theguardian.com"}],"official":0,"filed":"2026-10-20","updated":"2026-10-20","orgs":["Mass General Brigham"],"title":"FAHR-Face: Mass General Brigham foundation model estimates biological age and cancer survival from a face photo (Lancet Digital Health)","summary":"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'"],"key_numbers":[],"tags":["medicine","oncology","foundation-model","computer-vision","prognosis"],"science":{"field":"medicine","subfield":"oncology prognosis","problem":"Estimating biological age and short-term survival of cancer patients from routine face photographs","result":"Face-based risk scores added prognostic information to a clinical model; 64 of the 122 highest-risk patients (of 3,820) died within 90 days","open_since":"","ai_system":["FAHR-Face"],"human_role":"Human-led study; the AI is the measured instrument","verification":{"key":"peer-review","label":"Peer-reviewed","text":"Peer-reviewed in Lancet Digital Health (read via press; the paper itself not opened)"},"status":"confirmed","shock":""},"body_md":"## What happened\n\nMass General Brigham's team pre-trained a general face model and then adapted it to two health tasks. The figures above come from\nScientific American's report; the Guardian article and the paper were not readable for us.\n\n## Why it matters\n\nA camera-only prognostic signal is cheap to collect and could help clinicians, but it also raises the risk of face photos being used\nto judge people's health by insurers or employers, as outside experts warned.","disputed":[],"related":[],"people":[],"posts":[],"videos":[],"models":[],"changes":[{"date":"2026-10-20","type":"filed","text":"Created (Google News; Scientific American read)"}],"provenance":{"agents":[{"model":"Claude Opus 5.5","maker":"Anthropic","tool":"Claude Code"}],"filed":"2026-10-20","run":null,"sources_read":"Google News; Scientific American read","updated":"2026-10-20","human_review":null,"version":null},"gaps":[{"model_id":"gpt-6-astra","name":"GPT-6 Astra","cutoff":"2026-04","days_after":172,"in_training_data":false},{"model_id":"claude-opus-5-5","name":"Claude Opus 5.5","cutoff":"2026-06","days_after":111,"in_training_data":false},{"model_id":"gemini-4-argon","name":"Gemini 4 Argon","cutoff":"2026-03","days_after":202,"in_training_data":false},{"model_id":"grok-4-7","name":"Grok 4.7","cutoff":"2026-05","days_after":141,"in_training_data":false}],"short_url":null}