Researchers at Mass General Brigham have developed a new artificial intelligence tool called FaceSurvival that predicts cancer survival rates using a patient’s selfie. Published in The Lancet Digital Health, the technology aims to standardize visual assessments and support clinical decisions before treatment begins.
FaceSurvival: AI Analyzes Selfies to Predict Cancer Survival
A team of American researchers at the Mass General Brigham institution has created a diagnostic instrument designed to evaluate mortality risks for cancer patients. According to a report published by the British newspaper Daily Mail, scientists have developed an artificial intelligence tool capable of identifying patients facing the highest risk of death simply by looking at their faces, meaning a straightforward self-portrait could soon assist physicians in predicting whether a cancer patient will survive their illness. The technology builds upon previous research indicating that cancer individuals appearing older than their actual chronological years carry a higher likelihood of mortality.
The findings were detailed in a report published by the The Lancet Digital Health journal, signaling a potential radical transformation in how practitioners assess patient conditions prior to initiating therapeutic interventions.
Algorithm Sorts Individuals into Hazard Tiers During Trials
During trials, the algorithm sorted individuals into distinct hazard tiers based on visual data extracted from their portraits. The division revealed stark contrasts in survival rates across the tested population over a short 90-day window.
- More than half of the individuals sorted into the highest risk category died within 90 days.
- Only 2% of the individuals classified in the lowest risk group died during the same timeframe.
Standardizing Visual Assessments in Oncology
In standard medical practices, doctors frequently make rapid evaluations regarding a patient’s health simply by observing them visually, a method recognized as visual assessment that assists in estimating patient frailty or weakness. The newly developed algorithm bridges the gap between intuitive observations and quantitative measurement.
Physicians already evaluate patient health intuitively based on appearance. Our tool provides a means to standardize this intuitive practice and measure it accurately, which could contribute to improving the accuracy of medical predictions and supporting clinical decision-making.
Dr. Grace Lee, colorectal surgery expert and study co-author, via Alarabiya
Experts observe that this technology might help clinicians determine which patients maintain sufficient fitness to undergo intensive cancer treatments and which individuals might require a more moderate therapeutic approach.
