started · updated
Mass General Brigham researchers develop AI to predict cancer survival from facial photos
Researchers at Mass General Brigham have developed a foundation AI system called FAHR-Face that uses facial photographs to estimate biological age and predict survival outcomes for cancer patients. The system was trained on a database of over 40 million facial images and adapted into two specialized tools: FAHR-FaceAge and FAHR-FaceSurvival.
FAHR-FaceAge estimates biological age by analyzing features such as the forehead, temples, and wrinkles around the mouth and nose. The study found that patients appearing at least five years older than their actual age faced a 21% higher mortality risk. FAHR-FaceSurvival predicts survival chances directly by focusing on the area beneath the eyes and around the nasal bridge.
In a study of US cancer patients undergoing radiotherapy, the survival model showed significant predictive power. Patients in the highest-risk group were three times more likely to die during the study period compared to those in the lowest-risk group. The research, published in The Lancet Digital Health, suggests these tools could eventually assist clinicians in assessing a patient's physical resilience before recommending intensive treatments like surgery or radiation. However, researchers emphasized that the technology is intended to support, not replace, professional medical judgment.
Entities
Harvard Medical School · Harvard University · Hugo Aerts · Lancet Digital Health · Mass General Brigham · Massachusetts · Raymond Mak · The Lancet Digital Health · United States
Claims
What the coverage asserts, and how many sources carry each claim.
- [● 2 SOURCES] In a test group of 3,820 cancer patients, over half of those with the highest risk score died within 90 days of the photograph being taken. www.ithome.com · www.scientificamerican.com
- [● 19 SOURCES] The tool can estimate a person's biological age. rr.pt · www.ithome.com · www.independent.co.uk · www.scientificamerican.com · www.ghanamma.com · +13 more
- [● 21 SOURCES] Researchers developed an AI tool capable of predicting cancer patient survival rates from a single facial photograph. rr.pt · www.ithome.com · www.independent.co.uk · www.scientificamerican.com · www.ghanamma.com · +15 more
- [● 35 SOURCES] The AI model was trained using more than 40 million facial images. www.leinetal24.de · www.mannheim24.de · www.op-online.de · rr.pt · www.ithome.com · +29 more
- [● 13 SOURCES] The AI tool's age estimation error is approximately five years. www.ithome.com · www.businesstimes.com.sg · economia.com.py · www.calibre800.com · capitalul.ro · +8 more
- [● 8 SOURCES] Patients appearing at least five years older than their chronological age had a 21% higher risk of death among those undergoing radiotherapy. rr.pt · columnadigital.com · technews.tw · www.noticiasdecoimbra.pt · economia.com.py · +3 more
- [● 32 SOURCES] The AI system includes two tools: FAHR-FaceAge and FAHR-FaceSurvival. rr.pt · columnadigital.com · technews.tw · www.g4media.ro · www.ghanamma.com · +26 more
- [● 12 SOURCES] Smoking showed the strongest pattern of facial aging in the study. technews.tw · startsat60.com · www.independent.co.uk · www.novilist.hr · libertas.mk · +6 more
- [● 7 SOURCES] The mortality rate for the highest risk group was more than three times higher than the lowest risk group. www.leinetal24.de · www.mannheim24.de · www.op-online.de · economia.com.py · internewscast.com · +2 more
- [● 28 SOURCES] The FAHR-FaceSurvival tool predicts cancer survival directly without age estimation. www.leinetal24.de · www.mannheim24.de · www.op-online.de · capitalul.ro · estamosaqui.mx · +23 more
- [● 4 SOURCES] More than 50% of high-risk patients died within three months. www.leinetal24.de · www.mannheim24.de · www.op-online.de · nlc.hu
- [● 3 SOURCES] The system was tested on a group of 3,820 US cancer patients to assign them to risk groups. www.leinetal24.de · www.mannheim24.de · www.op-online.de