«

人工智能能否通过视网膜扫描诊断阿尔茨海默症?埃里克·托波尔对此寄予厚望。

qimuai 发布于 阅读:30 一手编译


人工智能能否通过视网膜扫描诊断阿尔茨海默症?埃里克·托波尔对此寄予厚望。

内容来源:https://www.wired.com/story/big-interview-event-eric-topol-super-agers/

内容总结:

【医学前沿观察:颠覆传统,AI与免疫健康或重塑人类“健康寿命”】

长期以来,医学界普遍将40岁视为健康管理的分水岭:女性开始接受乳腺筛查,男性需关注前列腺健康;此后十年,肠镜检查提上日程,中老年仿佛步入逐步依赖医疗检查的漫长轨道,直至七八十岁身体机能衰退。

但这一时间线真的科学吗?美国斯克里普斯研究所副所长、心脏病专家埃里克·托波尔在新书《超级老人》及近日接受《连线》杂志专访时提出颠覆性观点:现代医学可能误判了疾病预防的“黄金窗口”。当前,部分中年人或许接受了不必要的筛查,而某些易患结肠癌的年轻人却被忽视。我们能否找到一种更智能、更本质的健康老龄化路径?

托波尔指出,长寿(lifespan)与健康寿命(health span)是截然不同的概念,而两者均与遗传关联甚微。研究发现,同样基因构成的65岁以上人群,可能成为无重大疾病的“健康长者”,也可能受困于心脑血管疾病、癌症或神经退行性疾病。关键差异或许在于免疫系统的健康程度。

“健康寿命应尽可能接近自然寿命,而我们现在正迎来实现这一目标的独特机遇。”托波尔强调。当前美国人均健康寿命仅约63至65岁,而平均自然寿命达80岁,这意味着许多人晚年可能面临长达15年的健康衰退期。据世界卫生组织数据,多数65岁以上老人仅能经历一次“健康生日”。

托波尔认为,变革动力来自三方面:

  1. 人工智能医疗突破:大语言模型可高效分析海量病历,发现人力难以察觉的个体或群体健康风险;视网膜图像分析能提前数年预警动脉疾病、帕金森病等;AI甚至能通过实验室数据趋势,以低成本在胰腺癌确诊前数年实现早期预警。
  2. 生物标志物革新:如p-tau217蛋白可提前10-20年预测阿尔茨海默病风险;“器官时钟”技术能精准监测全身各器官(包括免疫系统)的衰老速率,实现衰老量化评估。
  3. 抗炎疗法新前景:GLP-1类药物不仅调节血糖,更能通过肠-脑轴与免疫系统互动,成为“抑制炎症的完美干预手段”。该药物对心、肾、肝脏具有保护作用,还能抑制吸烟、咬指甲等成瘾行为。明年年初将启动研究,探索其对有阿尔茨海默病遗传倾向人群的预防潜力。

然而,托波尔提醒,健康寿命延长的核心仍在于生活方式:减少超加工食品摄入、重视睡眠质量、亲近自然、坚持有氧与抗阻平衡训练。同时需警惕空气污染、微纳米塑料等环境致炎因素。尽管这些议题未被某些政治议程充分重视,但个人主动管理至关重要。

“生活方式是延长健康寿命最重要、最经济的驱动力。”托波尔在采访尾声郑重呼吁,“如果谁对延长健康寿命不感兴趣,请告诉我——但我想这样的人并不存在。”

(本文基于《连线》杂志对埃里克·托波尔的专访及《超级老人》观点综述)

中文翻译:

几十年来,人们普遍认为一旦年过四十,就该更加关注身体状况。女性通常从这个年纪开始接受乳腺X光检查,男性则需更多留意前列腺健康。随后十年里,结肠镜检查将提上日程,自此便仿佛踏上了逐步增加就医问诊的漫长征途,直至七八十岁时身体机能衰退。

但现代医学的时间线是否完全错了?如果我们正为部分中年人进行本不必要的疾病筛查(他们很可能终生不会罹患这些疾病),却盲目忽视可能易患结肠癌的二十多岁年轻人呢?是否存在一种可能,让我们即使逐渐老去,也能以有意义且无需每日吞服12颗马药丸般巨型药片的方式保持健康?

埃里克·托波尔对此深信不疑。这位斯克里普斯研究所副所长、心脏病学专家兼《超级老人》作者确信,人工智能辅助医疗、生物工程和抗炎意识等领域的创新,有望彻底改变人类衰老的方式。

在周四于旧金山举行的《连线》杂志大访谈活动中,托波尔告诉特稿编辑桑德拉·厄普森:他在撰写《超级老人》时意识到,寿命长度与健康寿命存在差异,且两者都与遗传基因关联甚微。那些"健康长者"(指65岁以上且总体健康的群体)与同龄但面临心脏病、癌症或神经退行性疾病等重大健康挑战的老年人,其基因构成几乎完全相同。

托波尔指出,健康的免疫系统与健康老龄化之间存在显著关联。生活方式同样影响健康状态——他倡导减少超加工食品摄入、注重睡眠质量而非时长、多接触自然环境。他还建议进行有氧运动、抗阻训练及平衡训练,这些都能增强身体随年龄增长所需的韧性。

托波尔强调,人们应尽可能避免接触空气污染、微纳米塑料及永久性化学物质等环境压力源,这些都会引发炎症反应。他特别提到,尽管特朗普总统和卫生部长小罗伯特·F·肯尼迪推行"让美国再次健康"议程,却未正视这些问题。

托波尔表示,普通美国人的健康寿命约为63至65岁,而预期寿命可达80岁。这意味着多数美国人将在相对不健康的状态下度过最后15年人生。世界卫生组织数据显示,大多数老年人65岁后仅能经历一次"健康生日"。

"我们应尽可能延长健康寿命,使其接近实际寿命,我相信这可以实现。"托波尔告诉厄普森,"这是医学史上的独特时刻。部分归功于多模态人工智能,部分源于我们获得了新的数据维度。我们首次拥有追踪全身各器官(包括免疫系统)衰老速度的器官时钟,首次掌握能提前10-20年预警阿尔茨海默病风险的生物标志物p-tau217。当代生物医学最重大的突破,正是量化衰老指标的能力。"

托波尔认为人类仅刚开始窥见GLP-1药物的真正潜力,称其仍处于"药物家族发展的早期阶段"。这类药物不仅建立肠道与大脑间的信号传导,还能与免疫系统对话,堪称"抑制炎症的完美干预手段",对心、肾、肝脏皆有益处,甚至能帮助戒除吸烟、咬指甲等成瘾行为。他透露明年年初将启动研究,探索GLP-1药物能否对具有阿尔茨海默病遗传倾向、50岁前开始用药的人群产生预防效果。

托波尔对人工智能的医疗应用前景"相当乐观"。大型语言模型能批量审阅病历,发现连最敏锐的医生都可能忽略的个体或群体健康问题。AI可分析海量个人医疗数据,仅凭视网膜图像就能提前数年预警动脉疾病、帕金森病或阿尔茨海默病风险。他甚至指出,AI能以较低成本检测预示胰腺癌的实验室指标趋势,较传统诊断提前数年,这或将引发医疗行业的革命性变革。

尽管众多网红和科技人士热衷通过生物黑客技术干预衰老进程,托波尔认为其实并不复杂。他在访谈总结时强调:"生活方式才是关键所在。这是影响健康最重要的驱动力,也是延长健康寿命最经济的方式。如果在座各位有谁对延长健康寿命不感兴趣——请务必告诉我。"

英文来源:

For decades now, it’s been fairly well established that once you turn 40 you should start paying more attention to your body. That’s when women are supposed to start getting mammograms and men are supposed to start paying a bit more attention to their prostates. Over the next decade, you’ll start getting colonoscopies, and from then on out, it feels like a gradual march of doctor’s appointments and tests until your body collapses sometime in your seventies or eighties.
But what if modern medicine has the timeline all wrong? What if we’re testing some middle-aged people unnecessarily for diseases they’ll most likely never get, while blindly ignoring twentysomethings who might be prone to colon cancer? Is there a way that, even as we age, we can stay healthy in a way that’s both meaningful and not reliant on taking 12 horse-sized pills every morning?
Eric Topol certainly thinks so. The cardiologist, vice president of Scripps Research, and author of Super Agers is convinced that new innovations in AI-assisted medicine, bioengineering, and anti-inflammatory awareness could have potential to revolutionize the way people age.
During WIRED’s Big Interview event in San Francisco on Thursday, Topol told features editor Sandra Upson that while he was working on Super Agers he learned that there’s a difference between lifespan and health span and that neither has much to do with genetics. Someone who’s “wellderly,” or over 65 and generally healthy, has pretty much the same genetic makeup as someone who’s elderly and facing major health challenges, like heart disease, cancer, or a neurodegenerative disorder.
Instead, Topol said, there appears to be a correlation between having a healthy immune system and aging healthily. Lifestyle can influence how healthy you are, too, with Topol advocating for eating a diet low in ultra-processed foods, focusing on sleep quality over sleep quantity, and getting out in nature. He also recommended getting exercise, focusing both on aerobic work and resistance and balance training, which can help the body become more resilient as you age.
If possible, Topol said, people should avoid environmental stressors, like air pollution, micro and nano plastics, and forever chemicals, all of which Topol said are pro-inflammatory. All of these, Topol noted, are not being addressed by President Trump and health secretary Robert F. Kennedy Jr., despite their Make America Healthy Again agenda.
For the average American, Topol said, health span is about 63 to 65 years. Lifespan, on the other hand, is about 80 years. That means most Americans will spend the last 15 or so years of their lives in relatively poor health, with one World Health Organization stat saying that most elderly people will only experience one “healthy birthday” after the age of 65.
“Health span should be extended as close as we can to lifespan, and I think we can do it,” Topol told Upson. “This is a unique moment in medicine. Part of it is because we have multimodal AI, but part of it is because we have new layers of data. We never had organ clocks, which track the pace of aging for every organ of your body, including your immune system. We never had biomarkers like p-tau217, which tells us about our risk of Alzheimer's 10, 15, even 20 years in advance. The biggest jump in recent biomedicine is the ability to quantify metrics of aging.”
Topol believes humans are only starting to see the true potential of GLP-1s, adding that it’s still the “early innings for this family of drugs.” GLP-1s create signals from a body’s gut to its brain, but they also talk to a body’s immune system, which Topol says can provide the “perfect intervention for bringing down inflammation,” creating benefits for the heart, kidneys, and liver, but also limiting someone’s addiction to everything from cigarettes to nail-biting. Topol said there’s a study scheduled for early next year that will examine whether GLP-1s could have an impact on preventing Alzheimer’s in people with preexisting dispositions for the disease but who can begin to take the drug before they turn 50.
Further innovations in the world of medicine could also come through the use of AI, which Topol said he’s “pretty bullish” about. Large language models can review medical records en masse, noticing problems in single individuals or groups of people that even the most discerning physicians might not. AI can analyze billions of points of individual data in a person’s medical chart and, using just an image of someone’s retina, find risks for arterial issues, Parkinson’s, or Alzheimer’s years before anyone else could. Topol said AI can even be used to pick up lab trends that indicate pancreatic cancer years before a diagnosis could be made and at a relatively low cost, something that could be revolutionary for the healthcare industry.
And while much has been made about influencers and tech personalities biohacking their own aging process, Topol said it’s really not that hard. “Lifestyle is what’s important,” he concluded during the interview. “It’s the biggest driver we have, and it’s the most inexpensive way to extend your health span. And if there’s anyone here who’s not interested in extending their health span, please, let me know.”

连线杂志AI最前沿

文章目录


    扫描二维码,在手机上阅读