美国Mass General Brigham医疗系统研发了一套AI工具，能够通过患者的单张自拍照预测癌症生存风险1。该系统包含两个互补工具——FAHR-FaceAge和FAHR-FaceSurvival，其中前者估计生物年龄并预测癌症生存率，后者则直接从面部照片预测生存率1。研究团队的训练数据来自超过4000万张不同种族和年龄的面部照片1，研究成果已发表在《柳叶刀数字健康》期刊1。
根据研究结果，在接受放射线治疗的美国癌症患者中，生物年龄看起来比实际年龄老5岁或以上的患者死亡风险高21%，而看起来比实际年龄年轻5岁或以上的患者死亡风险则低17%1。研究还发现，重酒精使用、吸毒和吸烟与生物年龄老化相关1。共同首席作者Hugo Aerts表示："据我们所知，这是AI模型首次直接从简单照片预测癌症生存的演示"1。
Researchers at Mass General Brigham have developed an artificial intelligence system capable of estimating cancer survival outcomes from a single facial photograph 1. The technology comprises two complementary tools: FAHR-FaceAge, which calculates biological age and forecasts survival rates, and FAHR-FaceSurvival, which predicts survival probability directly from facial imagery 1. The AI system was trained using more than 40 million facial photographs representing diverse racial and age groups 1.
Among cancer patients undergoing radiation therapy in the United States, those whose faces appeared at least five years older than their chronological age faced a 21 percent increased risk of death, while patients appearing at least five years younger than their actual age showed a 17 percent decreased mortality risk 1. The research identified correlations between biological aging markers visible in facial features and heavy alcohol use, drug consumption, and smoking 1. According to Hugo Aerts, one of the study's co-lead authors, "To our knowledge, this is the first demonstration of an AI model directly predicting cancer survival from a simple photograph" 1.
The findings have been published in The Lancet Digital Health journal 1.
评论
还没有评论，欢迎留下第一条。