Have you ever experienced a moment in your life that was so painful and confusing, that all you wanted to do was learn as much as you could to make sense of it all? When I was 13, a close family friend who was like an uncle to me passed away from pancreatic cancer. When the disease hit so close to home, I knew I needed to learn more. So I went online to find answers. Using the Internet, I found a variety of statistics on pancreatic cancer, and what I had found shocked me. Over 85 percent of all pancreatic cancers are diagnosed late, when someone has less than a two percent chance of survival. Why are we so bad at detecting pancreatic cancer? The reason? Today's current "modern" medicine is a 60-year-old technique. That's older than my dad. (Laughter) But also, it's extremely expensive, costing 800 dollars per test, and it's grossly inaccurate, missing 30 percent of all pancreatic cancers. Your doctor would have to be ridiculously suspicious that you have the cancer in order to give you this test. Learning this, I knew there had to be a better way. So, I set up scientific criteria as to what a sensor would have to look like in order to effectively diagnose pancreatic cancer. The sensor would have to be: inexpensive, rapid, simple, sensitive, selective, and minimally invasive. Now, there's a reason why this test hasn't been updated in over six decades. And that's because when we're looking for pancreatic cancer, we're looking at your bloodstream, which is already abundant in all these tons and tons of protein, and you're looking for this miniscule difference in this tiny amount of protein. Just this one protein. That's next to impossible. However, undeterred due to my teenage optimism -- (Laughter) (Applause) I went online to a teenager's two best friends, Google and Wikipedia. I got everything for my homework from those two sources. (Laughter) And what I had found was an article that listed a database of over 8,000 different proteins that are found when you have pancreatic cancer. So, I decided to go and make it my new mission to go through all these proteins, and see which ones could serve as a bio-marker for pancreatic cancer. And to make it a bit simpler for myself, I decided to map out scientific criteria, and here it is. Essentially, first, the protein would have to be found in all pancreatic cancers, at high levels in the bloodstream, in the earliest stages, but also only in cancer. And so I'm just plugging and chugging through this gargantuan task, and finally, on the 4,000th try, when I'm close to losing my sanity, I find the protein. And the name of the protein I'd located was called mesothelin, and it's just your ordinary, run-of-the-mill type protein, unless, of course, you have pancreatic, ovarian or lung cancer, in which case it's found at these very high levels in your bloodstream. But also, the key is that it's found in the earliest stages of the disease, when someone has close to 100 percent chance of survival. So now that I'd found a reliable protein I could detect, I then shifted my focus to actually detecting that protein, and thus, pancreatic cancer. Now, my breakthrough came in a very unlikely place, possibly the most unlikely place for innovation -- my high school biology class, the absolute stifler of innovation. (Laughter) (Applause) And I had snuck in this article on these things called carbon nanotubes, and that's just a long, thin pipe of carbon that's an atom thick, and one 50,000th the diameter of your hair. And despite their extremely small sizes, they have these incredible properties. They're kind of like the superheroes of material science. And while I was sneakily reading this article under my desk in my biology class, we were supposed to be paying attention to these other kind of cool molecules, called antibodies. And these are pretty cool because they only react with one specific protein, but they're not nearly as interesting as carbon nanotubes. And so then, I was sitting in class, and suddenly it hit me: I could combine what I was reading about, carbon nanotubes, with what I was supposed to be thinking about, antibodies. Essentially, I could weave a bunch of these antibodies into a network of carbon nanotubes, such that you have a network that only reacts with one protein, but also, due to the properties of these nanotubes, it will change its electrical properties, based on the amount of protein present. However, there's a catch. These networks of carbon nanotubes are extremely flimsy. And since they're so delicate, they need to be supported. So that's why I chose to use paper. Making a cancer sensor out of paper is about as simple as making chocolate chip cookies, which I love. (Laughs) You start with some water, pour in some nanotubes, add antibodies, mix it up, take some paper, dip it, dry it, and you can detect cancer. (Applause) Then, suddenly, a thought occurred that kind of put a blemish on my amazing plan here. I can't really do cancer research on my kitchen countertop. My mom wouldn't really like that. So instead, I decided to go for a lab. So I typed up a budget, a materials list, a timeline, and a procedure, and I emailed it to 200 different professors at Johns Hopkins University and the National Institutes of Health -- essentially, anyone that had anything to do with pancreatic cancer. I sat back waiting for these positive emails to be pouring in, saying, "You're a genius! You're going to save us all!" And -- (Laughter) Then reality took hold, and over the course of a month, I got 199 rejections out of those 200 emails. One professor even went through my entire procedure, painstakingly -- I'm not really sure where he got all this time -- and he went through and said why each and every step was like the worst mistake I could ever make. Clearly, the professors did not have as high of an opinion of my work as I did. However, there is a silver lining. One professor said, "Maybe I might be able to help you, kid." So, I went in that direction. (Laughter) As you can never say no to a kid. And so then, three months later, I finally nailed down a harsh deadline with this guy, and I get into his lab, I get all excited, and then I sit down, I start opening my mouth and talking, and five seconds later, he calls in another Ph.D. Ph.D.s just flock into this little room, and they're just firing these questions at me, and by the end, I kind of felt like I was in a clown car. There were 20 Ph.D.s, plus me and the professor crammed into this tiny office space, with them firing these rapid-fire questions at me, trying to sink my procedure. How unlikely is that? I mean, pshhh. (Laughter) However, subjecting myself to that interrogation -- I answered all their questions, and I guessed on quite a few but I got them right -- and I finally landed the lab space I needed. But it was shortly afterwards that I discovered my once brilliant procedure had something like a million holes in it, and over the course of seven months, I painstakingly filled each and every one of those holes. The result? One small paper sensor that costs three cents and takes five minutes to run. This makes it 168 times faster, over 26,000 times less expensive, and over 400 times more sensitive than our current standard for pancreatic cancer detection. (Applause) One of the best parts of the sensor, though, is that it has close to 100 percent accuracy, and can detect the cancer in the earliest stages, when someone has close to 100 percent chance of survival. And so in the next two to five years, this sensor could potentially lift the pancreatic cancer survival rates from a dismal 5.5 percent to close to 100 percent, and it would do similar for ovarian and lung cancer. But it wouldn't stop there. By switching out that antibody, you can look at a different protein, thus, a different disease -- potentially any disease in the entire world. So that ranges from heart disease, to malaria, HIV, AIDS, as well as other forms of cancer -- anything. And so, hopefully one day, we can all have that one extra uncle, that one mother, that one brother, sister, we can have that one more family member to love. And that our hearts will be rid of that one disease burden that comes from pancreatic, ovarian and lung cancer, and potentially any disease. But through the Internet, anything is possible. Theories can be shared, and you don't have to be a professor with multiple degrees to have your ideas valued. It's a neutral space, where what you look like, age or gender -- it doesn't matter. It's just your ideas that count. For me, it's all about looking at the Internet in an entirely new way, to realize that there's so much more to it than just posting duck-face pictures of yourself online. (Laughter) You could be changing the world. So if a 15 year-old who didn't even know what a pancreas was could find a new way to detect pancreatic cancer -- just imagine what you could do. Thank you. (Applause)
你是否经历过 非常痛苦和困惑的时刻 恨不得 立刻就能搞懂一切 我13岁时,有一位家庭挚友 亲如我的叔叔 因为胰腺癌过世了 当这么亲近的人遭逢病变, 我明白我需要更深入暸解 于是我上网找答案 在互联网上找到了 各种胰腺癌的统计数据 结果让我大吃一惊 85%以上的胰腺癌 到晚期才被诊断出来 晚期的存活率只有2% 检测胰腺癌为何成效不佳? 原因是现代的医疗方法 是60年前的老旧技术 比我的父亲还要老 (笑声) 而且很昂贵 检测一次需800美元 而且极不准确 30%的胰腺癌检测不出来 除非医生反常地怀疑 病人罹癌才会安排检测 我得知后,觉得一定有更佳方法 所以我建立了科学标准 界定了传感器的必备功能 以便有效地诊断胰腺癌 传感器必须便宜、快速 简单、敏感、又能筛选 并且是微创的检测 但是现有的检测方法 六十多年不变也有原因 因为诊断胰腺癌 是透过验血 血液富含巨量蛋白质 要找出微小的差异 微量的蛋白质 单单一种蛋白质 几乎不可能 然而我年少乐观所以不怕困难 (掌声) 我上网找青少年的两位益友 谷歌和维基百科 我从中得到了所需的资料 我读到一篇文章 该文列举了8000多种蛋白质 是胰腺癌患者血液裡有的 所以我开始了新的任务 逐一筛选这些蛋白质 找出胰腺癌的生物标志物 为了让任务更简单一点 我订定了下列科学标准 首先,符合条件的蛋白质 必须存在于所有类型的胰腺癌 早期血液裡的含量要高,但仅限于患者 我按部就班地执行艰巨的任务 终于在第4000次筛选的时候 在我快发疯的时候 找到了符合标准的蛋白质 这种蛋白质叫做 间皮素(mesothelin) 是再普通不过的蛋白质 但是如果患了胰腺癌 卵巢癌或肺癌,自然另当别论 患者血液里间皮素的浓度极高 但关键还在于 间皮素在胰腺癌最早期就出现 在此时期几乎100%的病人 都可以存活 既然找到了赖以检测的蛋白质 我把焦点转向如何测出该蛋白质 进而诊断胰腺癌 突破来自意想不到的地方 或许是最不可能创新的地方: 高中生物课 绝对是扼杀创新的地方 (笑声)(掌声) 我读到关于碳纳米管的文章 是一种长薄的管状碳 只有一个原子的厚度 是头发的直径的五万分之一 尽管碳纳米管的尺寸极小 却有惊人的性质 有点像物质科学的超级英雄 我偷偷地阅读这篇文章 是在上生物课的时候 当时应该专心学习的是 另一种叫做抗体的酷分子 他们之所以酷是因为 只与特定的蛋白质起反应 但抗体远不如碳纳米管有趣 然后我在课堂上 突然灵光一现 我能把读到的 有关碳纳米管的特性 和我应该学习的抗体结合起来 基本上我可以把抗体 植入碳纳米管的网络中 因此这种网络 只会与一种蛋白质起反应 但也由于碳纳米管的特性 网络的电性会改变 根据蛋白质的含量而变 然而,有一个弱点 碳纳米管的网络非常脆弱 这么纤弱的东西需要支撑物 因此我决定使用纸张 用纸制作癌症传感器 就像制作巧克力饼干一样简单 我很喜欢 要先把纳米管倒入水中 再加入抗体混合起来 把纸浸入混合液,然后弄干 就可以检测癌症 (掌声) 然后我突然想到 我这个妙计有点瑕疵 我没有办法在厨房的台面上 研究癌症 妈妈应该不会乐意 所以我决定找一间实验室 我拟定预算和材料列表 实验进度和程序 寄给200位不同的教授 约翰霍普金斯大学的教授 以及国家卫生研究院的人员 基本上与胰腺癌相关的任何人员 我坐在那儿期待赞美的电邮涌入 说“你真是天才” “你会拯救全人类” (笑声) 然后现实来临 在一个月内 200件电邮中有199件遭到回绝 一位教授甚至费心检查了整个程序 真不知道他哪来的时间 说明为什么我的每一个步骤 都像犯了有生以来最大的错误 显然教授们对此事的看法 不如我自认的那么重要 然而还有一线曙光 一位教授说“年轻人,也许我能帮你” 于是我就锁定他了 (笑声) 因为你总不好拒绝一位小孩 所以三个月后 我终于和他敲定了严格的期限 我去他的实验室 我兴奋极了,然后坐下来 开口说话 五秒钟后他召来另一位博士 接着一群博士涌入小办公室 他们连番问我犀利的问题 结束时我觉得像在小丑车上 20位博士再加上我和教授 挤在小小的办公室里 他们连珠炮似地质问着我 想要搞垮我设计的程序 他们能得逞吗?谁怕谁 (笑声) 然而我被审讯的时候 回答了所有的问题 不少是用猜的,但我都答对了 我终于有了需求的实验室 但我不久之后就发现 我曾经认为绝妙的程序 似乎有100万个漏洞 在七个月内 我煞费苦心地弥补每个漏洞 成果是这张小的纸制传感器 成本三美分,检测时间五分钟 速度快了168 倍 成本便宜26000倍 敏感了400倍 远超乎目前的胰腺癌检测标准 (掌声) 传感器的最佳效能之一 是近乎100%的准确度 并且可以在癌症最早期测出 最早期的存活率接近100% 所以未来两到五年 传感器可能提高胰腺癌的存活率 从令人沮丧的5.5% 到接近100% 对于卵巢癌和肺癌也将有类似效果 但是不止于此 如果变更抗体 可以检测不同的蛋白质 从而检测不同的疾病 也有可能适用于世上任何疾病 所以从心脏疾病 到疟疾、 艾滋病毒、 艾滋病 以及其他形式的癌症——任何疾病 希望有一天 我们都可以多保有一位叔叔 一位母亲、 一位兄弟姐妹 可以多保有一位挚爱的家人 可以免除疾病造成的心里负担 包括胰腺癌、 卵巢癌、肺癌 任何疾病都有可能 透过互联网一切都有可能 理论是可以共享的 你不必成为多个学位的教授 想法才会受到重视 互联网是没有偏见的地方 长相、年龄、性别 都不重要 你的想法才重要 对我来说重点是 要全新看待互联网 要认识到其功能太大了 不只是上网贴自己的嘟嘴照片 你可以改变世界 如果有一位15岁的少年 连胰腺是什么都不知道 却能找出检测胰腺癌的新方法 想想看你能做些什么 谢谢 (掌声)