読書リサーチレーダー投資フレームワーク
ログイン / 新規登録
← 読解に戻る
ニューストピック · 麻省总院面部AI预测重大1 分で読めるDisconfirmAI 編集部

US researchers report AI predicts cancer survival from single selfie

Researchers at Mass General Brigham developed an AI model that estimates biological age and predicts cancer survival rates using only facial photos, representing preliminary research results.

2026-10-20 15:00 GMT+8

A team at Mass General Brigham, a US academic medical center, has published research in Lancet Digital Health demonstrating that artificial intelligence can estimate a patient's biological age and directly predict cancer survival outcomes from a single standard facial photograph.

The researchers trained a foundation model named FAHR-Face on over 40 million facial images across diverse races and ages. From this base, they derived two tools: FAHR-FaceAge, which estimates how old a person looks compared to their chronological age, and FAHR-FaceSurvival, which bypasses age estimation to predict survival probabilities directly. According to reports citing the study, among US cancer patients undergoing radiotherapy, those whose faces appeared five or more years older than their actual age had a 21% higher risk of death during an average four-year follow-up period. Conversely, those appearing younger had a 17% lower risk.

Hugo Aerts, director of the Medical AI program and a senior author, stated that this appears to be the first demonstration of an AI model predicting cancer survival directly from a simple photo. The study also found that lifestyle factors like smoking and heavy drinking correlate with faster facial aging. The team emphasizes that these tools are intended to assist personalized treatment decisions rather than replace physician judgment, and further validation is required before widespread clinical use.

ソース

ログイン / 新規登録
読書リサーチレーダー投資フレームワーク
technews.tw↗
トピック

麻省总院面部AI预测

美国麻省总医院布莱根(Mass General Brigham)团队开发出一款名为FAHR-Face的人工智能基础模型,声称仅凭一张普通面部照片即可估算患者的“生理年龄”并直接预测癌症存活率。该研究于2026年10月发表于《柳叶刀·数字健康…

3 ステップ0 件更新日 2026-10-21 01:51

経緯

  1. 2026-10-19麻省总医院布莱根团队在《柳叶刀·数字健康》发表论文,介绍FAHR-Face模型及其衍生的FaceAge和FaceSurvival工具,展示从单张照片预测癌症存活率的初步成果。theguardian.com ↗
  2. 2026-10-20多家媒体跟进报道,披露具体数据:外貌老化5岁以上者死亡风险高21%,年轻5岁以上者风险低17%;同时提及生活方式(吸烟、饮酒)与面部老化的关联。news.sciencenet.cn ↗
  3. 2026-04-28同一团队在《自然·通讯》发表后续研究,提出“面部衰老率”(FAR)概念,利用连续照片追踪生物年龄变化,发现更高的FAR与更差的总体生存率相关。nature.com ↗

未解決の問題

  • 该模型在不同种族、不同癌症阶段以及非放射治疗场景下的泛化能力如何?
  • 面部特征与癌症预后之间的生物学机制是什么?是因果关系还是伴随现象?
  • 监管机构是否会批准此类基于图像的AI作为独立的医疗器械或辅助诊断工具?
トピック全文を読む →
リサーチを続ける →