医疗创新中的圣杯 | 斯科特·戈特利布博士

我们认为这些真理不言而喻，即人人生而

We hold these truths to be self evident. That all men

平等。请开始你的服务介绍。作为国会议员，我有幸在办公室接待许多非常有趣的人物，他们都是各自领域的专家。这就是本播客的宗旨。

are creed. Take your service covering. As a member of congress, I get to have a lot of really interesting people in the office. Experts on what they're talking about. This is the podcast.

深入探讨各类议题：中国、生物恐怖主义、全民医保等，通过深入浅出的方式解析复杂问题。我们认为我们认为我们认为这些真理我们认为这些真理——丹·克伦肖。

For insights into the issues. China, bioterrorism, Medicare for all, in-depth discussions. Breaking it down into simple terms. We hold We hold We hold these truths We hold these truths Dan Crenshaw.

鹰已着陆。

The eagle has landed.

欢迎大家回来。我们再次回到《坚守真理》——你最喜爱的播客节目，请给它五星好评，因为它就是你最爱的播客。今天我们有位常客——斯科特·戈特利布博士。

Welcome back everybody. We are returning to Hold These Truths, your favorite podcast, and you're gonna give it five stars because it's your favorite podcast. And we've got a repeat offender here. We've got Doctor. Scott Gottlieb.

他是特朗普总统任内的前FDA局长，现为美国企业研究所研究员、美国国家医学院院士，曾在美国卫生与公众服务部任职，并担任纽约大学医学院教授。自上次对话后，我们又邀请过他几次。记得2020年新冠疫情时期，戈特利布博士频繁出现在新闻中，成为公众寻求客观信息时最信赖的权威面孔之一。他近期正在撰写新书《解读细胞治疗革命》。

He's the former FDA commissioner under President Trump, currently an AEI fellow, member of National Academy of Medicine, served previous stints at the Department of Health and Human Services, professor at New York University School of Medicine. Since the last time we spoke, we've spoken to him a couple times. You know, the first time we may remember doctor Gottlieb, he was all over the news during '20 the year 2020 talking about COVID. He he was he was one of those, I I think, familiar faces that people look to for just to to just get some balanced takes on the information that everybody was receiving at rapid fire rate. He's been working on a new book recently called The Making Sense of Cell Therapy Revolution.

这正是我们能源与商业委员会下属卫生小组委员会关注的领域。虽然话题专业性强，但关于细胞治疗的原理及其对未来世代的意义——从治疗阿尔茨海默症、心脏损伤到癌症，甚至可能逆转衰老效应（这样我们都能变成吸血鬼了）——都将带来深刻启示。好了，关于他的成就就先介绍到这里。

This is something we take an interest in on on the health care subcommittee on on energy and commerce. So it's gonna be nerdy, but but I think there's gonna be some interesting takeaways from what cell therapy is and how it can how it can help our future generations. And, you know, by help, I mean ranging a range of diseases from Alzheimer's to heart damage to to cancer, maybe even reverse the effects of aging, and then we can all be vampires. But, alright, I'll stop there. There's a lot to say more about his bio.

斯科特，感谢你的参与。

But Scott, thanks for being on.

非常感谢。谢谢邀请。

Thanks a lot. Thanks for having me.

那么，我们就从热潮开始谈起吧。虽然书还没出版，但为什么选择细胞疗法这个主题？你完全可以写很多不同题材的书。我本该知道但忘了问——你在医学领域具体专攻什么方向？

So, I mean, let's just start with with the boom. It's not out yet, but why why cell therapies? You could write a book about a number of things, I'm sure. I don't I don't know what your I I should know this, but I forgot. What's your what did you specialize in as just in the medical field?

这个专业背景和这本书有关联吗？

And then did that relate to this book at all?

是的。我主攻普通内科医学，曾作为住院医师负责院内患者诊疗，包括重症监护病房。后来我进入FDA工作（2017-2019年），期间我们批准了首批基因疗法——当时共通过三项：其中一项是针对儿童白血病的治疗方案。

Yeah. General internal medicine, and I practiced as a hospitalist taking care of patients inside the hospital, inside the intensive care unit. Well, coming out of the FDA so I was at FDA from 2017 to 2019, and when I was there, we approved the first gene therapies. We approved three of them at that time. One was a treatment for pediatric form of leukemia.

我们还批准了针对遗传性失明的基因疗法。那段时期我不仅见证了已获批疗法的成效，更从研发管线中看到了细胞与基因疗法更广泛的变革潜力——这正是我决定写这本书聚焦该技术发展的原因。想想我们日常医疗行为，作为住院医师开药时，多数时候只是在治疗疾病症状，偶尔幸运地能延缓病程进展。

We also approved gene therapy for a treatment, a treatment for an inherited form of blindness. And what I was seeing with respect to not just what we were able to prove over that time period, but what was in the pipeline was the transformative potential of these new modalities of cell and gene therapy more broadly. And that's why I wanted to write a book focusing on the development of this technology. If you think about what we do in medicine and what I do as a hospitalist, a lot of what we do when we prescribe medications is treat the symptoms of disease. And sometimes we're fortunate enough to be able to slow the progression of a disease.

临床医学中能真正治愈疾病的疗法凤毛麟角。除了某些可治愈的癌症（肿瘤领域存在根治性疗法）和用抗生素就能根治的感染性疾病外，我们绝大多数药物治疗都只是缓解症状，至多延缓病情恶化。外科手术或许能治愈疾病，但这在内科领域并不常见。而细胞与基因疗法首次大规模提供了逆转疾病影响的可能性，在某些情况下更能实现根本性治愈。

There's very few places in clinical medicine where you can actually prescribe a therapy to cure a disease. Outside of oncology, certain forms of cancer that you can cure, there's curative therapy available for certain forms of cancer. And outside of infectious diseases where you can treat an infection, for example, with antibiotics and cure the infection, Most of what we do in medicine is prescribe drugs that treat the symptoms of the disease, and in some cases hopefully slow the progression of the disease. And surgery can potentially cure a disease, but not necessarily in medicine, internal medicine in the field that I practice in. And what cell and gene therapy offers, really for the first time on a broad basis, is the ability to reverse the effects of disease, and in some cases, develop a fundamental cure for disease.

我们已经从部分改良疗法中看到了成效，目前研发管线非常充实。我认为这将彻底改变医学领域，而我们才刚刚起步。正因如此，我决定聚焦于此。

And we're already seeing that with some of the therapies that have been improved, and the pipeline is very full right now. So I think this is gonna fundamentally transform medicine, and we're just at the early stages of it. And that's why I wanted to focus on it.

比如大多数人熟悉的干细胞疗法，但我完全不清楚它是否有效。我认为还需要更多临床数据验证，其具体适应症范围也尚不明确。有时人们会提取脂肪中的细胞进行培养——干细胞疗法其实分很多种类。

So, I mean, one kind of cell therapy that I think most people are familiar with is stem cell therapy, but I have no idea whether it works or not. I I think there's a lot more clinical data that needs to happen, and and it's unclear what it what it's useful for and what it what it isn't. Sometimes people take it out. There's different types of cell therapies or stem cell therapies. Sometimes they take it out of fat, kind of reproduce it.

脂肪来源的被认为是更高质量的。还有些情况会直接从骨髓或脂肪中提取后重新注射到关节里，可能让关节感觉好些，但效果并不明确。你对此总体看法如何？除了这种干细胞热潮外，细胞疗法还包含哪些内容？

That's considered a a more high quality one. In in other cases, they literally just take it out of maybe your marrow or your fat and then just reinject it into your joints, then, you know, maybe your joints feel better. It's kinda unclear. What's your general take on that? And then what else do we mean by by cell therapies beyond just the, again, this sort of, think, what would I call it, yeah, this fad of stem cells?

是的，这个概述很好，涵盖了细胞疗法的不同模式。你提到骨髓来源的细胞，那通常是富含血小板的血浆，很多骨科医生用它进行关节注射以尝试修复关节。你还提到了脂肪干细胞。

Yeah. That was a pretty good overview just laid out of the different modalities, if you will, of cell therapies. So you talked about taking cells out of the the bone marrow. That's generally platelet rich rich plasma, which a lot of orthopedic surgeons are using to inject into joints and try to rejuvenate joints. Then you talked about adipose stem cells.

这些既用于骨科手术也用于整形手术。我认为真正的突破发生在2006年，当时我们开发出了诱导多能干细胞技术——这本质上是将成体细胞逆转成干细胞的能力，让已分化的细胞回到原始状态，从而能分化成其他组织。现在很多研发中的治疗都基于这种技术。

Those are being used also in orthopedic procedures and also some cosmetic surgery. The real breakthrough, I think, came in 2006 when we developed the technology to create what we call induced pluripotent stem cells. And that's basically the ability to take an adult cell, so a cell that you might derive from a person's own body, and basically turn back the clock on it, turn it back into a stem cell. So take a finished adult cell and turn it back into its sort of primordial, you know, original state, a stem cell that can then be induced to form other kinds of tissue. And so a lot of the treatments that are being put into development right now are based on induced pluripotent stem cells.

无论是干细胞、诱导多能干细胞还是脂肪干细胞，广义上都属于再生医学范畴——试图用干细胞修复因疾病或损伤受损的组织。我在书中重点关注的另一类细胞疗法是：改造T细胞等免疫细胞作为靶向治疗，或改造血细胞来治愈镰刀型贫血症、血友病等疾病。通过操控血细胞DNA，我们既能治疗疾病，在某些血细胞本身病变的情况下还能实现根治。最近就有针对镰刀型贫血和血友病的基因疗法获批。

So, you know, those those kinds of modalities, when you talk about stem cells and induced pluripotent stem cells and even adipose stem cells, a lot of that falls within the rubric, broadly speaking, of regenerative medicine, trying to use stem cells to restore tissue that may have become damaged or destroyed as a result of disease or injury. The other cell therapies that are in development, and that's what I focus on very heavily in the book, are the ability to take immune cells like T cells and reengineer them and use them as targeted therapies, or to take blood cells and reengineer them to help cure diseases like sickle cell disease or hemophilia. So that's a whole another field where we're able to, by manipulating the DNA of blood cells, reengineer blood cells either to treat diseases or in some cases, when you're dealing with a disease where the underlying dysfunction is in the blood cell, it's blood cell itself, try to cure the disease. And we just saw a treatment get approved for sickle cell disease, which was a gene therapy delivered to blood cells. We saw one recently approved for hemophilia as well, another blood disease.

你们到底怎么修改细胞DNA的？现代语境下说起来轻巧，好像‘哦直接读改DNA就行’，但实际上根本没人知道具体怎么做吧？

How on earth do you manipulate a cell's DNA? I mean, it's it's we sort of we we talk about that in modern parlance. Like, it's something like, oh, yeah. I mean, just you just manipulate the you just read the DNA. Nobody has any clue how to do that, though, of course.

所以你到底给它注射了什么？是某种特定化学物质还是蛋白质组合？你从哪儿弄到这些蛋白质的？我喜欢尽可能从最基础的东西开始向人们解释，如果可行的话。

So so what what are you injecting it with some kind of specific chemical or or combination of proteins? Where do you even get those proteins? I like to try to explain things to people just from the very basics if it's even possible.

是的。我在这本书中做的一件事就是讨论早期阶段，因此我重点研究了CAR-T疗法的发展。这是一种用于治疗多种癌症的免疫疗法，特别是血癌。科学家们提取T细胞，如你所说，通过操控DNA让T细胞表达能识别癌细胞的新受体。取出这些T细胞后，用新受体重新改造它们，本质上训练它们识别癌细胞，然后大量扩增并回输。当这些携带新受体的T细胞回输到原患者体内时，就会开始攻击血癌细胞。

Yes. One the things that I did with this book was talk about the early so I focused a lot on the development of CAR T therapy. This is a form of immunotherapy being used to treat different forms of cancer, particularly blood cancers where where scientists basically take T cells and, as you said, manipulate the DNA to get the T cell to express a new receptor that can recognize the cancer cells. And so you take these T cells out, you reengineer them with new receptors where you basically train them how to recognize the cancer, and then you you expand them and reinject them in large quantities. And when they get reinfused into the same patient that they were taken from, now with this new receptor that they're expressing, they go on to attack the blood cancer cells.

改造和操控T细胞DNA的技术大约始于1980年代早期，源自斯坦福和MIT的研究。科学家发现可以利用病毒——特别是逆转录病毒（比如HIV就是逆转录病毒）——对病毒本身进行改造，去除有害特性，同时保留病毒的能力。

The technology for trying to reengineer and manipulate the DNA of T cells really started in probably the early nineteen eighties with research coming out of Stanford and then MIT, where scientists figured out how to use viruses, and particularly retroviruses, basically like HIV, HIV is the retrovirus, to take the to reengineer the virus itself and take the harmful features out of the virus, but maintain the aptitude that viruses had, the capability

但他们具体是怎么做到的？是通过反复混合某种化学物质上千次，直到发现'这样操作就能实现改造并达到预期效果'吗？我这样理解对吗，还是完全在胡说八道？

But that they like okay. So like but that's exact so like how did how did they do how did they reengineer it? Or did they mix it like, they mix it with some other chemical a a thousand times until they figured out, okay, when you do it this way, it actually reengineers and has the effect that we want it to have. Is that somewhat accurate, or do I have no idea what I'm talking?

对病毒本身，他们用实验室技术剪接病毒RNA的组件（病毒表达的是RNA而非DNA），剔除某些基因后插入能赋予新特性的基因，然后把这个病毒当作特洛伊木马，将遗传物质注入人类细胞。就像HIV感染细胞时，会注入自身遗传物质，教唆我们的细胞复制新病毒——本质上就是劫持了我们的细胞。

With the virus itself, they used laboratory techniques to splice out components of the virus's RNA. This was the RNA in the virus because viruses express RNA not DNA. And so take out certain genes and then reinsert other genes in the virus that gave it these new features, and then use the virus as sort of a Trojan horse to then inject that that genetic material into human cells. And so what viruses do, if you think about HIV, what HIV does when it infects our cells is it injects its own genetic material into our cells, which then train our cells to make new copies of the virus. So they basically hijack our cells.

如果能利用病毒这种劫持细胞并注入新...

And so if you can use that aptitude of the virus, the ability to hijack a cell and inject new

等等，你说得好像是在给单个细胞做手术似的。是这样吗？

See the only thing you But you kinda make it sound like you're performing surgery on a single cell. Is that?

基本上，我认为这是一个很好的思考方式。

That's basically, I mean that's a good way to think about it.

他们实际上是在剔除不喜欢的RNA，并以某种方式插入他们想要的RNA。我只是...我不想深入探讨这个。我只是好奇，但我们对这些能做到的事情真是令人惊叹——我甚至不确定我们的电视是否能开机，所以我的意思是，我们不知道有多少事情真正完成了，但这很迷人。好吧，那么这些就是另外几种不同类型的细胞疗法。我想我们在对话中也跨越到了所谓的基因疗法，对吧？

And they're kind of literally like poking out the RNA they don't like and somehow putting in the RNA they do like. It's just, I mean it's, I don't wanna dig too much deeper into this. I'm just I'm just curious, but it's it's just amazing what we've been able to do for I don't even know if our TVs turn on, so I mean, know, this is this is we we don't know how much of anything gets done, But it's fascinating. So, okay, so those are the other couple different types of cell therapies. And I think we crossed over in that conversation into what you would call gene therapies then also, right?

因为你刚才谈到了剪接RNA和DNA。

Because you were just talking about splicing out RNAs and DNAs.

当你思考细胞疗法时，很多我关注的细胞疗法，以及许多获得批准的细胞疗法，其实是在操纵血细胞中的基因。你从人体中提取血细胞或免疫细胞（B细胞或T细胞），改变细胞的DNA以赋予其新特性，然后将同一细胞重新注入患者体内。这实际上是一种基因疗法，因为你所做的基因修饰不是直接作用于患者，而是作用于他们的血细胞——提取、修改后再回输。还有其他形式的基因疗法，你提供给患者的不一定是修饰过的细胞，而是载体本身。比如直接将病毒载体输送到患者的某个组织或器官中，希望病毒能将新的遗传物质递送到正确的细胞里。

When you think about how cell therapies, a lot of the cell therapies that I'm I'm focusing on, a lot of the cell therapies that you see being approved, where what you're doing is manipulating the genes in a blood cell. So you're extracting a blood cell from a person or an immune cell, b cell or a t cell, manipulating the DNA of the cell to give it new features, and then reinserting that same cell into the patient. That's really a form of gene therapy because the gene modification you're making isn't directly on the patient, it's on their blood cells, and so you're extracting their blood cells, modifying them, and then reinserting them. There's other forms of gene therapy where what you're delivering to the patient isn't necessarily a modified cell, what you're delivering is the vector itself. So you're delivering the virus directly to the patient into some tissue, some organ in the patient, and then hoping that that virus will then deliver the new genetic material into the right cells.

就像我之前提到的治疗遗传性失明的基因疗法，那种情况下，你是将经过修饰、携带新基因的病毒颗粒直接注射到患者眼睛里。有些基因疗法可能会将病毒直接注入患者的肝脏或胰腺，用于治疗糖尿病或某些肝脏疾病。所以这两种是不同的基因疗法形式。我认为像CAR-T这样的血细胞操作疗法——我经常写到的——属于外源性基因疗法。

And so when you think about, I talked about a gene therapy that treats an inherited form of blindness. In that case, what you're doing is literally injecting the virus particles that have been modified to to insert this new gene directly into the patient's eye. So some of these gene therapies, you might insert the virus directly into a patient's liver or their pancreas if you wanted to try to treat diabetes or try to treat certain diseases that manifest themselves in the liver tissue. So those are, you know, those are two different forms of gene therapy. I think of the manipulation of blood cells like CAR T therapy, which is what I write a lot about, is sort of an exogenous form of gene therapy.

你是在体外实验室环境中对血细胞进行基因改造，然后回输；而内源性基因疗法则是直接将病毒载体输送到患者组织中实现基因修饰。

You're doing the gene manipulation outside the body in a person's blood cells in a laboratory setting, and then endogenous forms of gene therapy, what you're delivering is the viral vector to deliver the gene modification directly into the patient's tissue.

那些T细胞疗法主要与癌症治疗相关吗？

And for those T cell therapies, is that related to the cancer treatments mostly? Just

很多CAR-T疗法确实如此。最初，CAR-T疗法被批准用于治疗多种血癌。现在我们发现，这些对血癌有效的CAR-T疗法，也可能对自身免疫疾病有效。目前正在进行一些针对CAR-T疗法治疗例如狼疮等疾病的临床试验，已经观察到非常显著的早期疗效。

any A lot of the CAR T, yeah. So both. Originally, the CAR T therapies were approved for the treatment of different forms of blood cancer. Now what we're finding is some of these CAR T therapies that are effective at treating blood cancers are also potentially effective at treating autoimmune diseases. So there's some trials going on looking at CAR T therapy and, for example, lupus, where you're seeing some really profound early results.

虽然尚未获批用于这些适应症，但临床数据非常鼓舞人心。CAR-T疗法的作用机制是——特别是这种疗法——它能清除患者的B细胞。对于B细胞癌变的患者，通过靶向B细胞表面CD19抗原的CAR-T疗法，可以彻底清除人体内所有B细胞（包括健康细胞和癌细胞），从而治愈癌症。我们还发现，通过CAR-T疗法清除B细胞后，可能重置了人体免疫系统，因此对某些自身免疫疾病患者能有效消除病症。

None of them approved yet for those indications, but there's really encouraging clinical data. And the belief is what the CAR T therapies do is really so this CAR T therapy in particular, what it does is wipe out a patient's b cells. And so if a patient has a cancer that arises in the b cells by delivering this CAR T therapy, which targets an antigen found on the surface of b b cells called c d 19, you're able to basically ablate or wipe out all of a person's b cells, both their healthy ones as well as the cancerous ones, and that's how you cure the cancer. But what we've also found is by wiping out the b cells using these CAR T therapies, what what we think happens is you reset a person's immune system. And so for certain patients who have autoimmune diseases, you can effectively extinguish the autoimmune disease.

虽然目前仅在早期小规模临床试验中观察到，但患者群体已展现出非常显著的疗效。

And we've seen that in very small, albeit early clinical trials, but very profound results in a group of patients.

那么是否可以这样说：如果说癌症治疗即将迎来突破，很可能将以某种基因疗法的形式出现？这是您现阶段会做出的假设吗？至少对某些癌症而言？

So would it be safe to say that if there is a cure for cancer on the horizon, it's probably going to be in the form of some kind of gene therapy? Is that an assumption that you might make at this point, or at least some cancers?

听着，基于多年来的研究进展，我们已经看到包括CAR-T疗法在内的多种癌症'治愈方案'获得药监局批准。例如2017年我刚加入FDA时获批的CAR-T疗法，就治愈了部分急性淋巴细胞白血病患者——这是一种常见于年轻患者的血癌类型。虽然部分患者未能受益，但确实存在相当数量对其他疗法（包括传统化疗）无效的患者通过该疗法获得治愈。随着这些细胞疗法在靶向治疗上展现出的精准高效，未来会有更多癌症类型被攻克。目前我们在治疗血癌的细胞疗法设计上已取得长足进步。

Look, we've seen, because of all the progress that's been underway for many years, we've seen a lot of what I would call cures for cancer being developed and approved by the agency, including CAR T therapies. In some patients, so the particular CAR T therapy that was approved when I first joined FDA in 2017 cured certain patients who had acute lymphoblastic leukemia, is a form of cancer that affects blood cancer that affects a lot of young patients, it's a pediatric form of blood cancer typically. Some patients who got the therapy didn't benefit from it, but there were a number of patients, and continue to be a number of patients, who fail other therapies that are available, fail conventional chemotherapy, and can be cured by this CAR T therapy. So I think we're gonna see that more and more, where these cell therapies are so effective and so precise in their ability to target cancer that you're gonna see more cancers able to be cured. And really the holy grail right now, we've been good at designing these cell therapies to treat blood cancers.

但尚未能可靠地将这些细胞疗法应用于实体瘤治疗。现在已出现更先进的方法，这需要FDA在监管制度上的创新，也需要科学突破——因为相关技术平台与既往不同。不过近期临床试验已展现出积极成果：有CAR-T疗法正被用于实体瘤治疗，约12个月前公布的胰腺癌治疗数据（众所周知这是致死率极高的癌症）就体现了这一点。

What we haven't been able to do yet in a really reliable fashion is use these cell therapies to also target solid tumors. And you're seeing now better methods for doing that. It's gonna take some regulatory innovation on a part of FDA, but also take some scientific innovation because the platforms for doing that are different than what we've done in the past, but you're seeing some encouraging results in clinical trials where these CAR T therapies are now being trained on solid tumors. There was some data out about 12 ago looking at a CAR T therapy in the treatment of pancreatic cancer, which as you know is a typically deadly form of cancer.

用外行话总结一下：如果我没理解错，所有这些案例中都是提取患者自身细胞，经过您描述过的那些改造流程（具体我就不复述了），再回输体内。从这个角度看是高度个性化的治疗方式。始终如此吗？还是会用到他人细胞？

So I mean, sum this up in layman's terms, you're always, correct me if I'm wrong, in every single one of these cases, you're using the patient's own cells, and then you're putting them through some kind of modification, the kind of modifications that you've already described. I'm not gonna try and repeat them. And then you reinject that into the body. And so it's a very personalized form of treatment in in that sense. Is is that always the case, or do you sometimes use other cells from somebody else?

是的。你可以使用其他细胞，目前正在开发的疗法中有种现成可用的细胞形式。其优势在于，如果能开发出这种现成制备的细胞，就能大规模生产并储备——比如存放在药房，而不必像你说的那样为每位患者个性化治疗：取出患者自身细胞、进行改造、扩增后再回输。你能想象这套制造流程有多复杂昂贵，这也推高了这些疗法的高昂成本。不过使用患者自身细胞的好处是它们不会被识别为外来物。

Yeah. You you can use other cells, and there's treatments in development that are sort of off the shelf forms of cells. And the virtue there is that if you can develop an off the shelf preparation of cells, it's something that you could manufacture on a large scale and stockpile, you know, store in a pharmacy rather than having to personalize this treatment, as you said, for every individual patient, take the patient's own cells out, manipulate them, expand them, and reinfuse them, you can understand how difficult and costly of a manufacturing process that is, and it adds to the high cost of these therapies. The advantage of using someone's own cells though is that they're not gonna be recognized as foreign.

我确实觉得...有时候我们回顾这些疗法时会想，要走到今天这一步，需要经历多少研究、创新和突破啊。光是认识到某些疾病与基因构成有关这点，就得追溯到人类基因组计划，必须绘制出每个人的基因图谱，然后还要理解如何对DNA进行手术——我现在还是无法完全理解你说的概念。用微型手术刀吗？这中间有太多环节了。但我觉得人们确实应该退一步，对现代生活心怀感激。几百年来正是这些微小积累，才让我们现在能直接编辑基因，拯救那些过去会瞬间夺人性命的疾病——要知道几十年前人们甚至不知道病因。

I I do think it's we're taking a step back sometimes looking at these therapies and thinking, like, all of these steps in research and innovation and breakthroughs that had to happen to get to this point. I mean, to even recognize that some diseases are related to your genetic makeup, so so the, know, going back to the Human Genome Project and the necessity to to map out everyone's genetics and then understand how to do surgery on DNA, which I still really don't under just can't conceptualize what you're saying, and then I don't know. Like with like with a really tiny scalpel. I mean, are you even talking about? Or I mean, there's just so many steps, and I think it's it's always good for for people to take a step back and and and show a little gratitude for for the kind of modern lives that we live where people have taken these little baby steps over the courses of hundreds of years to to get to this point where you can literally edit genes and and and save you from a disease that would have killed you just like that and for and then you would have no idea why even just a few decades ago.

这挺有意思的。

That's kind of interesting.

听着，我认为这是个关键点，也是我在书中通过回顾历史试图阐明的观点。以首个CAR-T疗法的开发史为例——我提到2017年首个CAR-T疗法获批用于治疗某种儿童白血病。这种叫Kymriah的药物，其研发历程其实要追溯到三十年前。科学家在1970年代通过骨髓移植研究首次发现T细胞可能对抗癌有效——当癌症患者接受非完全匹配的骨髓移植时，反而更可能治愈。

Look, I think that that's a key point here, and that's the point I I tried to to make in the book by looking back over this history. If you if if you look at the history of the development of the first CAR T therapy so I mentioned that the first CAR T therapy was approved in 2017 for the treatment of this pediatric form of leukemia. If you look at the drug the drug was called Kymriah. If you look at the history that led to the development of that drug, it really stretched back thirty years. The first recognition that T cells could be effective in targeting cancer was made probably in the nineteen seventies, through research that came out of studies with bone marrow transplants where scientists realized that when patients with cancer receive bone marrow transplants, in cases where the bone marrow wasn't a perfect match for that patient, those patients actually were more likely to be cured for the of their cancer.

他们发现骨髓本身具有治疗效果，存在所谓'移植物抗肿瘤效应'。骨髓会攻击肿瘤，激发患者体内的抗癌免疫反应。后来科学家分离出骨髓中负责这种免疫攻击的细胞，发现正是T细胞。于是研究者开始思考：既然T细胞能靶向癌细胞，我们该如何利用这种能力？这才产生了改造T细胞来靶向特定癌症的构想。

And so what they discovered was that there was a therapeutic effect from the marrow itself. There was something called graft versus tumor effect. So the marrow was attacking the tumor, and so was stimulating an immune response against the cancer that the patient had, and then they isolated the cells in the marrow that that were was responsible for that immune attack, and it turned out to be the t cells. And so then science scientists realized, okay, t cells could be effective at targeting cancer cells, how could we harness this ability? And that's where the idea came for trying to take t cells and reengineer them to train them how to target specific kinds of cancer.

我刚才描述的历程确实横跨三十年，期间科学家不断获得渐进式发现，逐步掌握新能力——比如利用逆转录病毒递送基因，或学会直接操控细胞内的基因——最终才催生出首个获批疗法。

So that journey that I just described was literally a thirty year journey where there were these small incremental insights that scientists were making along the way, and then new capabilities that they were acquiring, like the ability to use retroviruses to deliver genes, or figuring out how to manipulate genes themselves in cells that ultimately led to the first approved therapies.

是啊。你知道吗？这个故事适用于我们大多数现代设施，无论是医疗设备还是...你看我们正在用来视频通话的电脑屏幕，光速般的通讯。接下来我们聊聊监管吧，虽然有点专业，但这正好涉及我的工作领域。

Yeah. It's it's and and you know what? That that story applies to most modern amenities that we have, whether it's in health care or you know, this this computer screen that we're using to chat with each other, speed of light. So we'll talk about regulation for a little bit. This is this is a little wonky, but, you know, it does it does serve right into to what I do up here.

我们一直在思考如何能在FDA（美国食品药品监督管理局）更具创新性，或者说如何更有利于创新，而我所在的委员会负责监督FDA。但我知道要找到许多现成的解决方案并不容易，比如直接说‘把事情做对做好就行’。至少从立法角度来看，这确实很难实现。你之前提出的两种不同情境，我认为是展开对话的良好起点——第一种情况是提取患者自身细胞进行轻微改造，这里就涉及到‘轻微’与‘深度’改造的界定问题。

We're always trying to figure out how we can be more innovative or I guess more conducive to innovation at the FDA, and my committee oversees the FDA. But I know it's hard for us to find a lot of easy buttons to say, like, let's just do things right and awesome. It's it's that's hard to find, at least from the legislative perspective. You laid out two different scenarios earlier, which I thought would be a good starting point for this conversation, which was in one case, you're you're taking my own cells. You're you're you're you're manipulating them a little bit, and there was a question of what is a little bit and what is a lot of bit.

这对我们的讨论至关重要。另一种情况则是使用通用细胞制造可上架销售的药物，正如我之前提到的，这种方式效率更高。但两者存在本质区别：前者在我看来更像是骨髓移植——或者说，好吧，这到底算骨髓移植还是...老实说我对骨髓移植了解不深，但这个类比是想说明，这属于将患者自身经改造的细胞回输体内。

That's important for how we talk about this. And then in another case, you know, you're maybe, you're taking just general cells, and you're you're creating a drug that you can put on a shelf, as I mentioned before. It's just more efficient that way. But those are two very different things. In one case, again, seems to me like you're talking about like a bone marrow trans, like a, well, is that a bone marrow transplant or or even just a a is that I guess I don't know about bone marrow transplants to make that analogy, but you're talking about my own cells somewhat somewhat manipulated back into my body.

而后者涉及他人细胞，就更符合药物的特征。

In another case, you're talking about somebody else's cells, so it sounds a lot more like a drug.

是的，骨髓移植这个类比很适合第一种使用自身细胞的情况，对吧？

Yeah, bone marrow transplant is a good analogy for the first scenario where it's your own cells Right,

因为这确实是患者自身细胞。我没疯——虽然没实际操作过，所以可能说得不够专业，但这反而是好事。

because it is your own cells. I'm not crazy. Okay. Haven't done one before, so I I found myself talking a bit out of turn. That's probably a good thing.

我作为患者经历过很多治疗，对此深有体会。现在FDA面临的核心问题是如何监管这类创新：面对这些精妙的突破性技术，比如那些可能拯救生命、改变世界的基因与细胞疗法，目前审批流程究竟需要多长时间？

I've been a patient a lot, and thank thankfully enough for that. Okay. So now there's a real question at FDA over over how that should be regulated, and because you've got all this really neat innovation, and, you know, for it can it can take I don't know. You tell me. Like, how how long is it taking for some of these gene and cell therapies that could potentially save lives, change the world?

这些疗法获得批准要多久？我们的审批方式合理吗？是否过度套用了药物审批模式？我关注这个问题是因为我选区有位选民正在接受我们讨论的这种干细胞疗法——从脂肪中提取干细胞离心分离，让干细胞自行增殖后再回输体内。整个过程几乎没有其他复杂操作，但据我了解却被当作全新药物申请来审批。以我这个非医学人士的常识来看，这似乎有些脱离实际，你怎么看？

How long is it taking to get them approved? Do we have the right way to approve them, or are we thinking about them too much like a drug? I dealt with this because one of my own constituents does this stem cell therapy that we were talking about, and take it from your fat, and they spin it around a centrifuge, extract the stem cells, and then get the stem cells to kind of grow on their own, so you reproduce your own stem cells, then you can reinject those at a later date. There's not a whole lot of other manipulation there going on, and yet it is treated, as I understand it, like a whole new drug application. And then just from my non doctor common sense standpoint that seems a little out of touch, but what's your take on that?

你看，你触及到了核心原则。如果讨论的是现成的细胞制剂，在大多数情况下，我很难想象经过工程改造的现成细胞制剂不被视为药物。但当涉及从患者体内提取细胞、进行某种处理后重新输回同一患者体内——甚至可能回输到原提取部位时，关于何时这种操作会跨越界限成为药物，并非所有人都能明确判断。法规中的定义是：如果从患者体内提取的细胞经过了超出所谓'最低限度处理'的操作——抱歉——那么它就成为药物。所以问题在于：什么叫做'超出最低限度处理'？

Well look, you touched on the right principle. So if you're talking about the off the shelf cell preparations, in most cases those are gonna be, I can't think of a scenario where sort of an off the shelf preparation of cells that have been engineered aren't gonna be a drug. But when you're talking about cells where you're extracting the cells from the patient, doing some form of manipulation to the cells, and then reinfusing them into the same patient, and potentially in the same spot where you extract the cells, it's the question of when that crosses over into becoming a drug isn't always clear to everyone. The way it's defined in regulation is if the cell that's extracted from the patient is more than what we call minimally manipulated, excuse me, then it becomes a drug. And so the question is what is more than minimally manipulated means?

我的理解方式是：如果你改变了细胞的特性，通过某种处理使细胞在重新植入患者体内后表现出新的行为方式，这通常就属于超出最低限度处理。可能是改变细胞的物理特性，可能是添加药物改变其行为，也可能是修改细胞基因——这些显然都超出了最低限度处理。但确实存在一些接近界限的情况，我认为这正是争议的焦点所在。

If you're, the way I think about it and the way I describe it is if you're changing the characteristics of that cell, if you're doing something to the cell where you're changing the behavior of the cell so that it behaves in a new way once it's reinserted into the patient, that is a a general way to think about what is more than minimal manipulation. So it might be changing the physical characteristics of the cell, It might be adding drugs to the cell to to change how it behaves. It might be changing the genes of the cell. That's certainly more than minimal manipulation. But there are these things that fall close to the line, and I think that's where a lot of the debate is.

听起来你描述的情况可能就属于这种边界案例。例如提取脂肪干细胞用于治疗骨科损伤或美容效果时，若对细胞进行离心浓缩后回输，这可能就属于超出最低限度处理——因为你通过移除细胞原有支架改变了其特性，尽管改变程度不大。这类问题往往处于监管的灰色地带，需要企业主动向FDA申请许可。我在FDA任职时，我们曾尝试为某些低风险情形制定特别豁免条款。

And it sounds like you might be describing one of those circumstances with with your constituent. So for example, if you extract adipose stem cells and you might reinsert them to try to treat an orthopedic injury or to have a cosmetic effect, if you're taking the cells, extracting the cells from the patient and spinning them down so you concentrate them and then reinfusing that preparation, that might be more than minimal manipulation in the sense that you're changing the characteristics of the cell because you're taking it out of its original scaffolding, and that will change the characters characteristics of the cell, but not in a very significant way. And so that might be one of these sort of on the line kinds of questions about whether or not it crosses a line and becomes more than minimally manipulated. And these are kind of the gray area questions where I think sponsors, you know, have to come in to the FDA, ask for permission, in certain cases it can be frustrating for them because they don't feel like they're getting back a timely answer. When I was at FDA, we tried to do was carve out certain of these, what we considered low risk circumstances.

就像我描述的这个脂肪干细胞浓缩案例：虽然因移除细胞支架改变了特性而可能属于超出最低限度处理，但用于皮肤或骨科等低风险场景时，我们明确将其划出监管范围。只要符合良好生产规范和无菌操作标准，对患者风险就非常有限。

So, you know, exactly why I described a situation where someone may be taking adipose stem cells, concentrating them. So it's probably more than minimal manipulation because they're changing certain characteristics of the cell by taking it out of its usual scaffolding, but using it in a low low risk way like a like a dermal procedure or like an orthopedic procedure. So we explicitly carved that out, and we said, look. This may be more than minimal manipulation, but for purposes of regulation, we consider it a low risk use of cells. As long as you follow good manufacturing standards, you do it under sterile technique, there's not a lot of risk to the patient.

因此我们选择不予监管，行使执法自由裁量权。我认为监管机构应该更主动地界定这类低风险情形，从而能集中资源监管那些可能给患者带来更高风险的治疗方案。

So we're gonna choose not to regulate that. We're exercising enforcement discretion. And I think that's where the agency can lean forward a little bit more and describe some of more of these sort of low risk circumstances where it chooses not to regulate so that it could focus its activity on the higher risk circumstances where there's more potential risk to patients that are being introduced by some of these therapies?

确实，这反映出监管体系的不确定性。私营领域普遍认为决策标准不够统一——是否给予裁量权，或是将其归入更繁琐昂贵的审批路径。我还见过明显不匹配的案例：评审人员根本缺乏评估特定技术的专业背景，结果他们只能不断否决，要求补充试验，最终导致企业在资金链断裂边缘挣扎，可能永远失去这项技术。这种平衡至关重要。

Yeah, I mean it gets to this uncertainty on the regulatory side and I think a lot of the private sector feels like it's not enough consistency to how these decisions are made, whether that discretion is given or whether it's not or whether they place them under a different pathway that is just far more cumbersome and therefore far more expensive. And I I I've seen other cases too where it just it seemed kind of obvious that the reviewer in charge just wasn't the right person to assess that particular technology. Like, they just didn't have the background, and so what do they so what do they default to? They default to no, and to no again, and to another trial here, and another trial there, and next thing you know, you're in another round of fundraising for your company, which is about to go broke, and then we'll never have that technology again. So that balance is extremely important.

我在FDA任职期间目睹的最大挑战——至今依然存在——是临床审查环节。当需要证明产品安全有效时，开展临床试验生成数据总会面临标准问题：需要证明多大疗效？需要多少病例？特别是细胞治疗这类昂贵疗法，单个CAR-T疗法患者的试验成本就可能高达百万美元。当需要招募100名甚至1000名患者时，这个成本差距就非常惊人了。

Biggest challenge that I saw when I was at the agency, and I think it persists, there's certainly challenges on the clinical portion of the review where you're trying to demonstrate that a product is safe and effective for its intended use. And so you're running clinical trials to try to generate clinical data. You know, and there's always questions about what is the bar, how much of a therapeutic benefit do you need to demonstrate, how many patients is that gonna require, especially when you're dealing with these very expensive modalities like cell therapies, those trials can get hugely expensive. Putting a single patient in a clinical trial for CAR T therapy can literally cost a million dollars. And so, you know, if you're enrolling a 100 patients or asked to enroll a thousand patients, that's a pretty big delta if it's a million dollars for every patient that gets enrolled.

但在我看来，更大的不确定性源于生产工艺问题。FDA过去和现在都很难招募并留住精通这类产品制造工艺的专家，既要理解其中涉及的所有复杂性，又要把握生产过程中的细微调整——比如为提高细胞产量、加快规模化或降低成本而改进工艺流程时，这些微小变化可能如何影响细胞性能。由于监管部门缺乏相关专业知识（这类新兴科学本身也缺乏积累），当研发方需要对细胞制造工艺进行必要改进（这对实现治疗方案的商业化规模至关重要）时，FDA往往要求他们重新提供部分临床数据，以证明细胞特性变化不会影响安全性和有效性。

But the the bigger sort of bucket of uncertainty in my view was around the manufacturing issues, where I think it was very hard for the FDA, and still is hard for the FDA, to recruit and retain people who have expertise in how these products are manufactured, and understanding all the complexity involved in that, and also understanding how small changes in manufacturing, where you might be trying to change your manufacturing process to get a better yield of these cells or be able to scale it up more quickly or make it less expensive, how all those little changes is gonna translate potentially into changes in how the cells perform. And so because the because the agency doesn't have a lot of expertise in that, and there isn't a lot of expertise in that because this is such a new science, what ends up happening is when these sponsors wanna make slight modifications in how they manufacture these cells, which is very important to trying to scale it up in a way that you can actually turn it into a therapy that can be commercialized. In many cases, FDA is going back and asking them to reproduce some of the clinical data to be able to demonstrate that the cells, the characteristics of the cells haven't changed in a way that's changing their safety and efficacy.

这正是许多研发方陷入困境的地方，也是成本激增的症结所在。我们需要加强监管政策制定，让这个流程变得更高效。早期技术都会面临这种问题——记得2000年代初抗体药物刚上市时，FDA同样难以理解生产工艺变更如何影响临床效果，导致企业每次工艺改进都要重新进行大量临床试验。

And that's where I think a lot of sponsors are getting stuck, and that's where a lot of the costs are being added, and that's where we need, I think, need to focus more on more regulatory policy making to make that process itself you know, more efficient. And, you know, this is this is the problem with a lot of early technologies. We have a brand new platform. I remember back in the early two thousands when antibody drugs were coming onto the market, they they encountered some of these same complexities where the FDA really didn't understand how changes and how those antibody drugs were being made translated into changes in how they performed clinically. And so asked sponsors to go back, asked companies to go back and do a lot of new clinical trials every time they made a change in the manufacturing.

这确实阻碍了许多产品的规模化上市。如今我们已掌握相关技术手段，无需通过重复临床试验来证明工艺变更后的抗体药物完全相同。我们现在有足够把握批准抗体药物的仿制药，而不必每次都要求大规模新临床试验。

And it it was an impediment to getting a lot of those products scaled up and on the market as well. Now we have the tools to do that, so you don't need to replicate clinical trials to prove that one antibody is exactly the same as another one, even though the manufacturing process might have changed. And we we feel so certain about our ability to determine that that now we're approving generic forms of antibody drugs without requiring large new clinical trials every time.

这种变革很激进吗？到底是什么技术发生了改变？是实验室技术吗？实验室

Is that aggressive? What is that technology that changed? Is it laboratory technology? Laboratory

技术。没错，主要是应用技术。比如质谱分析这类工具，能通过物理特性判定不同抗体药物的等效性。现在我们通过分析抗体结构就能预测其临床表现。

technology. Yeah. So it's a lot of applied technology. So things like mass spec, a tool for looking at protein drugs, looking at antibodies, being able to determine just based on the physical characteristics that this antibody is gonna perform exactly the same as that antibody. So now we're able to impute by looking at the structure of the antibody, how it's gonna behave.

不过这就引出了政策层面的问题——正如你所说，这没有万能解决方案。这个领域的政策制定往往无法一蹴而就，需要国会多年累积对监管流程的细微调整，最终形成新的运作模式。但要实质性解决我们讨论的这些不确定性，我认为应该加大对监管应用科学的投入。这类研究通常不在NIH或企业的关注范围内，因为无法形成知识产权或带来利润。

So, you know, it leads to a policy point though because you said there isn't sort of a magic bullet. In many times, policy making in this realm isn't satisfying in the sense that there's not one thing you can change and that's suddenly gonna unlock the whole system. It's a lot of small modifications to the regulatory process that's made over years by Congress that, you know, gets accrued and eventually consolidates into new ways of doing things. But I think where we can make a big dent into some of these uncertainties that we're talking about is investing more in that applied science for answering these regulatory questions. And that's not something NIH does typically, it's not something that companies do because they can't own the intellectual property around that, there's not a profit to be had in that.

可以说这片应用技术领域近乎荒漠——没人专注于开发能证明细胞等效性的监管决策技术。这正是NIH应该加大资源投入的领域。虽然NIH更关注药物研发中光鲜的科研环节，但这类应用科学确实是值得填补的重要空白。

And so it's really this sort of barren landscape, if you will, where no one's really focused on the applied technologies to do things like be able to demonstrate that one cell is exactly the same as the other for regulatory decision making purposes. And this is some place where I think NIH could be focusing more of its resources. You know, NIH is looking at getting more into drug development. They want to be engaged in some of the sexier aspects of the sort of scientific enterprise. But this applied science is really a gap where I think they can be focusing more attention.

这很有趣。我是说，他们资金充裕。我不认为他们需要更多资金。他们每年都能获得更多拨款，而且有很多资助项目。

That's interesting. I mean, they have plenty of money. I don't think they need more money. I think they get more money every year. And they have a lot of grant programs.

你会以为他们能在这方面投入一点。说到应用科学，我们指的是那种能够替代临床试验的技术，这些试验——

You would think that they could invest a little bit on that. And by applied sciences, again, so what we mean is is the kind of technology that allows you to sort of replace a clinical trial, which are

极其昂贵。没错。一种利用实验室工具衡量临床疗效的新方法，比如阿尔茨海默病，我们可以通过核磁共振和功能成像来评估患者，诊断他们是否患病及病情进展，而不必依赖需要长时间测量患者迷你精神状态测试表现的临床试验。等到那些测试显现变化时，患者往往已经积累了大量功能障碍。如果能通过影像学更快地检测功能衰退，就能加速阿尔茨海默病治疗的临床研发。

extremely expensive. Exactly. New a new way to measure clinical benefit using a laboratory tool or, you know, think of like Alzheimer's disease, and we could figure out how to use MRI and and functional imaging to assess patients and diagnose when they have Alzheimer's and how it's progressing rather than requiring clinical trials where you're measuring a patient's physical performance on mini mental status tests, which takes a long time. By the time you see changes in that, there has been a lot of disability that's accrued to that patient. So if you can measure, you know, decline in function more quickly using imaging, that would help accelerate clinical development of of treatments for Alzheimer's.

但我们尚未以监管机构认可的方式验证这些工具。美国国立卫生研究院确实在做部分工作，不想说他们毫无作为，但确实应该投入更多资源来验证这些实验室工具。

But we don't we haven't validated those tools in a way that regulators would rely on it. That's something that you know, NIH does do some of this, so don't wanna say they're not doing it, but I think it's someplace where we could be focusing a lot more resources on validating some of these laboratory tools.

说到价格，你提到某些疗法每位患者要花费百万美元。这在临床试验中已经够难了，更让我担心的是如果获批成为正式疗法，每剂仍要百万美元。至少这是单次治疗吧？这些疗法实在昂贵，我首先想问为什么？是因为技术复杂度，还是所需的人力成本？

Speaking of prices, I mean, mentioned a million dollars per patient for some doses. So, I mean, that's hard enough in a clinical trial, but that also makes me assume that if that were approved as therapy, that's still a million dollars per dose. But at least is it a one time dose that we're talking about? Or is it, I mean, these things are really expensive, and I guess my first question is why? Is it just the technology and the amount of labor and personnel that's required to do it?

等我们真正取得突破后，到底要怎么负担得起这些费用？

And then how on earth do we pay for this once we find these breakthroughs?

是的，我说的每位患者百万美元是指CAR-T细胞疗法等临床试验的单例成本。但你说得对，某些已获批的基因疗法定价超过百万美元，甚至达到数百万。这类疗法通常只需单次或少数几次给药。

Yeah, so the million dollars per patient that I was referring to was the cost of a single patient in a clinical trial with something like a CAR T therapy. But you're right, some of these gene therapies that have been approved have been priced at more than a million dollars. They've been multimillion dollar gene therapies. Typically, are. Some of these are administered as a single dose or maybe just a a few doses.

要知道，真正的挑战在于支付这些费用，尤其是在紧缩的保险计划中。我们看到市场上存在显著的差异——拥有优质私人保险的患者能在疗法获批后及时获得治疗，而参与公共医保（特别是医疗补助）的患者则往往无法及时获取。这些保险体系在覆盖审批上迟缓得多。当涉及基因疗法或细胞疗法（如CAR-T疗法）时，及时性可能直接决定生死——这些疗法能彻底治愈或显著延缓像杜氏肌营养不良这类快速致残的遗传病。对先天患病迅速恶化的年轻患者而言，治疗延误意味着终身残疾与相对健康生活的天壤之别。这将成为我们面临的重大挑战。

You know, the the challenge is is gonna be paying for these, especially in austere insurance schemes. And so what we've seen is very differential access in the marketplace where patients who have better insurance, private market insurance, have had more timely access to some of these therapies when they've been approved, and people who are in more insurance schemes, and particularly Medicaid, have not had timely access. Those those insurance schemes have been much slower getting coverage in place. And when you're dealing with a gene therapy or a cell therapy like a CAR T therapy that is literally curative or potentially curative for patients or, you know, helps arrest a disease like Duchenne's muscular dystrophy where patients accrue disability very rapidly, young patients who are born with these inherited diseases accrue disability very rapidly, the timeliness of access can literally be the difference between life and death or a life lived with disability and one lived less encumbered by disease. So this is gonna be, I think, a big challenge for us.

为何如此昂贵？部分原因在于生产成本极高。以小分子药片为例，其生产成本可能仅需几美分到几美元。若售价达五万或十万美元，利润空间极大。但细胞疗法的制造过程可能耗资数万甚至数十万美元，导致产品成本居高不下。

Why do they cost so much? Part of the reason is that they just cost a lot to make, so the cost of goods is very high. If you think of a small molecule, a pill drug, it might cost pennies to make that drug or maybe dollars to make that drug. And, you know, if you're charging a $100,000 for the drug or $50,000 for the drug, a lot of that's profit because the drug didn't cost much to manufacture. With some of these cell therapies, literally the manufacturing process can be tens if not hundreds of thousands of dollars, So the cost of goods is very high.

我们观察到风投正在撤离该领域，因为回报率不佳。尽管这些疗法在临床上具有革命性意义且能根治疾病，但通常针对患者群体较小的罕见病。开发与生产成本极高的同时，适用患者有限，导致投资该领域的企业回报惨淡，许多已缩减相关项目。

And what we've seen in this space, and, you know, you have seen venture capital sort of pull away from these these areas is because the returns haven't been that good. If you're looking at sort of relative investments in different kinds of drug modalities and different kinds of opportunities, while these therapies are hugely exciting from a clinical standpoint and have the ability to cure disease, they're typically targeting very small diseases, some cases rare diseases. And so you're not gonna be prescribing them to a lot of patients. The cost of developing and manufacturing them is very high. And so companies that have invested in these spaces have not had good returns and a lot of them have curtailed some of their programs.

这确实是令人遗憾的现状。虽然现在机会窗口正在扩大，但回溯过去五年，许多大型药企和风投机构都曾暂缓对细胞与基因疗法的投资。

And so that's been one of the unfortunate things. I think it's changing now because the opportunity set's getting larger. But if you look back over the last maybe five years, a lot of the the the big drug companies and the venture capital firms that were investing in cell and gene therapies pulled back for a period of time.

您是否观察到《通胀削减法案》的影响？该法案要求联邦医保对新药进行价格谈判（某些疗法除外），若药企拒绝谈判将面临95%的营收税惩罚。这是否已冲击到细胞基因疗法市场及风投机构的投资布局？

Have you seen any effects from the inflation reduction acts legislation? You know, within that legislation was the the requirement for Medicare to negotiate with new therapies, new drugs. There's some exceptions in that. I don't know if these cell and gene therapies would fall into those exceptions, but it's a it's a requirement to negotiate at the the price that that Medicare wants to set for you, and there's a pretty large punishment if you don't take that negotiation being a a 95% tax on your revenue. So have you seen have you seen that affect the market and the venture capital in the in the especially in the investment space?

对细胞基因疗法尚无直接影响，但法案确实导致全行业研发预算缩减。更直接的影响是改变了投资方向——由于小分子药物将比抗体药等生物制剂更早进入价格谈判阶段，企业正减少对医保报销主力（如老年病治疗）的小分子药物投资。

Yeah. Not not directly with the cell and gene therapies. I mean, indirectly insofar as there has been a reduction in in r and d budgets across the industry as a result of the IRA. But where I've seen that the IRA have a more direct impact is biasing investment away from small molecules because the negotiation gets implemented sooner for small molecules versus biologicals like antibody drugs or even these cell therapies, you are seeing companies pull away from investment in small molecule pill form drugs, particularly small molecules that treat drugs where it's gonna be heavily reimbursed by Medicare. So drugs, you know, for the treatment of conditions that would affect older individuals.

资本重新配置的趋势已在药企交易中显现。投资正转向大分子抗体药物——它们比小分子药物更晚受价格管制，因此能保持更长时间的自由定价。这令人担忧，因为某些疾病靶点（如细胞内靶标或需穿越血脑屏障的脑部靶标）只能通过小分子药物治疗，而抗体药物通常仅能作用于细胞表面靶点。

There's definitely been a reallocation of capital, and you see that in the deals that are getting done in the pharmaceutical industry. So the thing that's gonna get affected first is the deal making where a company is placing new investments, and they're they're now shifting investment towards large molecules, antibody based drugs because those come under the price negotiation later than the small molecule drugs, and so they they're sort of free from the price controls for a longer period of time. And that's unfortunate because there's certain diseases and certain targets, biological targets, that can only be addressed with small molecules. If you're thinking about trying to target something inside a cell, so an intracellular target, if you're thinking about trying to target something in the brain that has to cross the blood brain barrier, small molecules are a very efficient way to do that. When we're talking about large molecules, biologicals like antibody drugs, typically they only target things on the surface of a cell.

它们无法进入细胞内部。因此，有些疾病和整个生物靶点领域只能通过小分子药物来解决，这确实是个值得关注的问题。我认为许多人担心这些领域的投资会减少，而且我们已经看到这种趋势了。

They can't get inside the cell. And so there's whole diseases and whole areas of biological targets targets that can only be addressed with small molecules, and that's a real concern. I think a lot of us have a concern that now there's gonna be less investment in those areas, and we're already seeing it.

那么小分子药物是否总是以药丸形式存在？还是也包括细胞和基因疗法？我们只是在讨论...

So can do small molecules always mean pill form, or can that also mean of cell and gene therapies? We're just talking about

小分子药物确实可以包括这些。细胞和基因疗法属于生物制剂，因此被视为大分子药物。小分子药物通常是口服药丸形式，但有些也需要输注。有些小分子药物是通过注射给药的，比如GLP-1受体激动剂这类用于减重的肽类药物，它们虽然是注射给药，但仍被视为小分子药物，因此会在更短时间内面临价格谈判。

Small molecule could mean yeah. So the so the cell and gene therapies would are biologics, and so they would be considered large molecules. Small molecules are generally pill form drugs, but some of them are also infused. There are small molecule drugs that are infused or injected. You know, the GLP-1s, the drugs that are being used for weight reduction are peptides, so they're an injected drug, but they're considered a small molecule drug, so they're subject to negotiation in that shorter time frame.

事实上，减肥药Wegovy（司美格鲁肽）明年就将面临价格谈判。所以小分子药物可以是注射剂，但通常还是以口服药丸形式为主。

In fact, Wegovy, one of the drugs, semaglutide, is gonna be subject to the price negotiation next year. So small molecules can be injected drugs, but they're typically swallowed as pills.

我也在想这个问题——当然我们无从得知，因为我们无法访问平行宇宙，看不到如果没有通过某项特定法律历史会如何发展。但由于价格谈判的时间限制仍然存在，这些公司...我是说，这很难判断。我认为他们不会公开承认，但如果知道日后会被CMS大幅压价，他们显然有动机从一开始就定更高价格。你们有观察到这种现象吗？

I wonder too, and again, there's no way to know this because we don't have access to the multiverse, and, you know, we we can't see what what what history would have been without some particular law being passed. But because it's a there's still a timeline on when they're subject to that negotiation, do these companies I mean, it's again, it's hard to know. I don't think they would announce it, but they they clearly have an incentive to have a higher price from the get go if they know they're gonna get hammered by by CMS at a certain at a later date. Are you seeing that at all?

是的。市场竞争激烈，所以很难说。目前定价面临很大下行压力。如果你看药品行业的通胀数据，去年是整体药价首次下降的一年。我不认为这是IRA法案导致的。

Yeah. The market's competitive, so it's hard. There's there's a lot of downward pressure on pricing right now. And if you look at inflation in the drug sector, last year was the first year that prices overall fell. And I don't I don't attribute that to the IRA.

我认为这只是药房福利管理机构和保险公司施加的市场竞争压力所致。所以很难说原本会怎样。我认为更容易看到影响的是...随着我们在市场上引入更多价格管控措施（IRA法案显然就是一种价格管控），这会刺激企业以更高的标价上市，因为他们知道市场价格管控会压低药价，回收投资的时间窗口也会更短。

I just attribute it to the competitive pressures in the marketplace exerted by PBMs and insurers. And so it's hard to say what what would have been. I think where it's easier to see an effect I mean, you could certainly there's certainly an argument to be made that as we introduce more price controls in the market place, and the IRA is certainly a price control, it creates incentives to try to launch at higher higher list prices because you know your price is gonna be driven down by the the various price controls that are in the marketplace, you're gonna have a shorter window to recoup

对，我就是这个意思。是的。在我

Right. That's what I mean. Yeah. Where I

是的，不过在我看来最直接的影响还是资本流向以及企业选择投资的领域。要知道，企业不会放弃已有的项目，所以你不会因为《通胀削减法案》就关停那些已进入后期研发阶段的项目。没错。但他们也不会再往这些领域投入新资金了。一旦企业丧失了开发小分子药物的科研能力——比如一旦他们解散了所谓的组合化学团队——就很难再重新涉足这个领域。

yeah, where I see the most immediate impact though is just the flows of capital and what companies are choosing to invest in. And, you know, companies aren't abandoning programs that they had, so you're not gonna just shut down a program that's in later stages of development because of the IRA. Right. But you're not gonna be putting new capital in those areas. And once companies get out of the scientific know how of developing small molecule drugs, so once they get rid of what you call combinatorial tree, they're not gonna just get back into it.

他们将彻底丧失这方面的专业技术和知识储备。

They're gonna lose the expertise and the know how for doing that.

那么在这个领域，我们最该担心哪些疾病呢？尽管我对此有所了解，但我并不清楚哪些疾病更可能用小分子药物治疗。具体是哪些？哪些可能得不到投资？

Then what diseases should we be most worried about in this space? Because I'm not a despite all my knowledge on this, I don't know which diseases are more likely to be treated with small molecule drugs, so what is that? What might not get that investment?

我特别关注的一个领域是精神类疾病，因为许多靶点是细胞内靶点，只能通过小分子药物来干预。还有传染性疾病，以及需要针对脑部进行治疗的疾病。

So one area that I've flagged is mental health diseases where a lot of the targets are intracellular targets and so can only be addressed with small molecules, infectious diseases, things that we have to treat inside the brain.

你是指像阿尔茨海默症这类疾病吗？

Are we talking about like Alzheimer's?

没错。如果你想用小分子药物治疗的话——虽然现在上市的一些治疗脑部疾病的药物是生物制剂，比如治疗阿尔茨海默症等神经退行性疾病的药物——但多数还是小分子药物。凡是需要进入细胞内部靶向细胞内受体的治疗领域都是如此。心血管药物和某些初级保健药物也属于这类，但我最担心的是精神类疾病，因为这类疾病很多靶点都在细胞内，传统上都是靠小分子药物来治疗的。

Yeah. If you're trying to address it with a small molecule, I mean, some of the drugs that are coming onto the market are biologicals that are being used to target brain diseases like Alzheimer's as well, neurodegenerative diseases, but a lot of them are small molecules. So any any anything like that, anytime you're trying to get inside a cell to target some receptor inside the cell, and so, you know, it could be cardiovascular drugs as well and certain drugs you just choose in primary care, but the mental health diseases is one big area that I worry about because a lot of those are intracellular targets that would typically be addressed with small molecules.

你与欧洲同行合作很多，我不知道FDA在欧洲叫什么。但凡事都有利弊。我刚从伦敦参加完情报委员会会议回来，这与医疗无关，但他们的政治局势就像所有国家一样动荡不安，就像我们这里一样。他们最大的问题之一是NHS（国家医疗服务体系），显然处于混乱状态，正在分崩离析，无论你需要什么服务都要排队等上好几个月。我不太清楚具体情况。

You've worked a lot with European counterparts, I don't know what the FDA is called in Europe. But there's pros and cons to it. I just got back from London on an intelligence committee trip, so it had nothing to do with healthcare, but their politics are all riled up just as they are in every country all the time, just like they are here. And one of the biggest issues there is NHS, their national health service, which is apparently in disarray, falling apart, and there's months and months and months of wait times for whatever it is you need. I don't know.

或许我不确定这里面是否有问题，但可能我在寻求一些看法——我们医疗体系的优缺点是什么？因为我确实知道有时创新者更容易选择欧洲。我听私营部门说过，在某些情况下他们与欧洲监管机构打交道更顺利。但同时我认为美国在医疗创新方面绝对是世界第一。

Maybe I'm not I don't know if there's a question in there, but maybe look maybe I'm looking for a comment, and and what are the pros and cons of our of our system? What because I look. I do know that sometimes it's it's easier for innovators to go to Europe. This is I've heard from the private sector that sometimes they just have a an easier time dealing with European regulators in some in some instances. At the same time, I think America's by far number one when it comes to innovation on anything with respect to health care.

那就谈谈这个吧。当我们审视NHS持续衰败的现状时应该警惕什么？还有加拿大实行的单一支付体系（与NHS有所不同）你又怎么看？如果你不介意的话请分享一下你的观点。

So talk to me about that. What should we be wary of when we look at the continuing downfall of NHS and and also what you see in Canada with their what is effectively a single payer system, not quite like NHS. But give give us your thoughts on that if you if you don't mind.

是的。我认为欧洲药品管理局（EMA）多年来在审批治疗重大未满足医疗需求的疾病（特别是影响儿童群体的罕见遗传病）方面更为积极。FDA和EMA在审批治疗严重未满足医疗需求的新型疗法时存在标准差异——当时缺乏有效治疗手段的领域。不过现在这个差距已经缩小很多了。

Yeah. Look. I think that the the European the EMA, the European regulators for a number of years were leaning forward a little bit more on the approval of treatments for significant unmet medical diseases, particularly orphan diseases, inherited diseases that affected pediatric populations. I think so there was a gap between FDA and EMA in terms of, you know, how permissive they were willing to be when it came to the approval of novel therapies that treated very serious unmet diseases where there weren't weren't good available therapy. I think a lot of that gap's been closed.

《21世纪治愈法案》及其实施缩小了大部分差距，FDA也采纳了欧洲的一些做法。我不认为欧洲监管机构现在更灵活——虽然在某些治疗领域他们可能有不同于FDA的侧重点，或许更宽松些。但总体而言，最差情况下双方监管水平相当。我在FDA工作过，某种程度上要为机构辩护：我认为FDA整体上是更包容的监管机构。

I think twenty first century cures and the implementation of it has closed a lot of that gap and that FDA has adopted some of the same approaches that the Europeans have. So I don't think that the European regulators, I would argue, that they're more flexible now. I mean, there may be certain pockets of certain therapeutic certain therapeutic categories where they might have a certain a different focus than the FDA and perhaps are a little bit more accommodating. But I think overall, you know, at at worst, the regulatory authorities are comparable, and I I would argue and, you know, I've I've worked at FDA, I'm gonna defend the agency to some degree. But I I would argue I think FDA is a more accommodating agency overall.

但就欧洲而言，这些新药根本得不到报销。越来越多情况下企业选择不在欧洲上市或延迟上市。特别是那些可能带来革命性疗效的产品，这终将导致治疗结果的差异。多年前我就撰文指出，美国与英国在癌症治疗效果上存在明显差距——虽然很难归因于单一因素（部分源于早期诊断和饮食环境差异），但确实部分源于2000年代中期那批革命性药物在美国获批更快。

With respect to Europe, though, these new products just aren't getting paid for. And in in more and more cases, you're seeing companies just not launch them in Europe or launch them much later in Europe. And, you know, particularly when you're dealing with products that are this potentially transformative in terms of the therapeutic effect that they're delivering, that's gonna start to translate itself into differences in outcomes. And you saw for years, and I wrote about this a number of years ago, you saw a real divide in cancer outcomes between The US and UK. And while it's hard to prescribe exactly what that's attributable to, you know, lot of it's some of it's early diagnosis and differences in diets and environment, but certainly some of it was attributed to the fact that there was this whole spate of drug approvals in February and sort of mid two thousands that were really transformative, and they weren't approved as quickly in The UK as they were in The US.

比如乳腺癌药赫赛汀、结肠癌药爱必妥和阿瓦斯汀，这些对市场影响巨大的药物在英国上市都被延迟。这种差距随时间不断扩大，我认为未来还会继续体现在治疗效果差异上。

You know, Herceptin for breast cancer, Erbitux for colon cancer, Avastin for colon cancer, Drugs that really had a big impact once they reached the market were delayed coming onto the market in The UK. And that's that divide has gotten wider over time. And I think you're gonna continue to see that play out in terms of outcomes.

是的。关于单一支付体系与非单一支付体系的辩论，当然还有不断被提及的事实——美国人会去加拿大和墨西哥购买药品。但我想强调的是，另一方面，加拿大和英国的患者在罹患极其罕见难治的癌症时，会专程前往休斯顿的MD安德森癌症中心求医。他们不会留在本国，而是选择来这里——虽然他们可能在本国能买到某种廉价的知名仿制药，因为两国的医疗体系不同，购药确实更方便。

Yeah. And and gets, and of course, to the debate of single payer systems versus non single payer systems, and and the constant reminder that, well, you know, Americans are going to Canada and going to Mexico to buy their drugs, and I think it's important to say, well, other side of that though is that the Canadians and the British are coming to MD Anderson in Houston when they have a really, really rare difficult cancer. They're they're not staying in their home country. They're they're coming here because they we might be drunk buying some kind of generic, well known drug there because they their system is different. It's just easier to buy it.

但要想获得优质医疗服务，他们必须来美国。那么你更倾向哪边？我宁愿处于我们的处境而非他们的。我认为关键是要在两者间找到健康平衡，特别是要让老旧技术更可负担——这涉及到仿制药审批的讨论，无论是治疗手段还是医疗器械。我们需要建立更具竞争性的体系，但也不能杀鸡取卵，扼杀最初带来医疗进步的创新动力。

But for quality healthcare, they have to come here. And so which side would you rather I'd rather be in our situation than their situation. And I think I think trying to find that healthy balance between the two, making especially making, I think, old technology more affordable, and that's that gets to a discussion about, you know, generic drug approvals and whether those are therapies or medical device, whatever it is, and and making that a more competitive system. But you also can't kill the goose that laid the golden egg and kill the innovation that gives you the healthcare in the first place. Can say that

医疗问题听着，我认为我们本可以采取更多措施让过专利期药物更可负担——特别是在专利到期后，这些药物本应面临仿制药竞争。多数情况下确实如此，但某些领域却未见国会预期的仿制药激烈竞争进入市场，这背后存在政策与监管原因。但观察欧洲在高端疗法（比如我们讨论的细胞与基因疗法）以及新型生物制剂方面的现状，患者根本难以获得这些治疗。

healthcare Look, I think there's is a we could be doing to make older drugs more affordable, particularly after patents have lapsed and drugs should be subject to generic competition. In many cases they are, but in some cases they aren't. Where Congress intended for there to be brisk competition from generic alternatives, you're not seeing that enter the market. And there's some policy and regulatory reasons why in certain cases that's not happening. But when you look at what's happening in Europe with respect to some of these, you know, more expensive but more advanced therapies, the ones we're talking about here, cell and gene therapies, but also a lot of the new biologics that are coming onto the market, patients just aren't getting access to it.

长期以来——我对此持续关注——欧洲患者群体曾强烈意识到自己无法及时获得新疗法，并向政府施压。这种态势现已减弱。因为人们逐渐认识到根本负担不起，药企也放缓了在这些国家的上市速度，远不及在美国或日本的投放节奏。

And, you know, for a long time, and I've followed this pretty closely, for a long time there was a real revolt's the wrong word, but there was a recognition among patients in Europe that they weren't getting timely access to some of these these new therapies, and you saw them pressuring their governments. That's died down. You see less organized pressure from patient groups in Europe who aren't getting timely access to some of these new therapies because I think there's just been this sort of recognition that has set in that they just can't afford them. And companies in many cases just aren't launching the drugs in those countries as quickly as they'll launch them in The US or Japan as well.

这很遗憾。如果某种细胞疗法定价百万美元，而主要成本来自实验室制备流程本身——不，绝不可能这么便宜。当然很大部分是用来回收巨额研发投入，有时高达数十亿。

It's sad because you could also so if a drug is therapy, if this cell therapy is a million dollars, and you're telling me that's largely because of the actual, I guess, laboratory process that is required to even do it. Oh, no way. It's inexpensive. And of course, lot of it is paying back their investment, which was Proficient. Sometimes in sometimes in the billions.

特别是考虑到我们之前讨论的临床试验类型。如果市场规模更大，人均成本本可降低——这是基本经济学原理。但现状是药企要么大幅折价进入欧洲市场，要么干脆放弃——对于细胞疗法这种非标准化产品，连折价销售都难以实现。

I mean, you know, especially for the kind of clinical trials we were describing earlier. The overall price per patient could be lowered if the market was bigger. Right? That's I think that's the general economics of it. But, you know, we're we're stuck in this position where they they either way undersell it to European countries or or underprice it to European countries just just to get it out there, which sounds like it can't even be done with cell therapies because it's not like a pill.

制备过程本身成本极高导致无法规模化生产。我的观点是：无法规模化就难以降价，我们陷入两难。但至少我们拥有这些疗法，至于如何支付——这完全是另一个议题了。我笔记里有相关内容，不确定是否你写过或我的团队想请教你。

The actual process of doing it is very expensive, and so you can't scale. I guess the point I'm making is when you can't scale, you can't lower price, and so we're stuck, but hey, at least we're at at least we have something. And, you know, figuring out how to pay for it is is a whole other issue. I I have it in my notes here. Don't know if you wrote about this or my staff just wanted to ask you about it.

但说到保修，这是否是一种创新方式来支付这些极其昂贵的疗法？

But on warranty, is that an innovative way to think about paying for some of these extremely expensive therapies?

是的。我认为尤其是当你处理基因疗法时，在很多情况下这都是一次性治疗，希望能治愈疾病。但如果你仔细想想，你不是在十年或一生中持续提供治疗，所以必须通过那一次治疗收回所有投资和回报，这就是为什么这些疗法的成本如此之高。在这些情况下，你可能想做的是尝试将成本分摊到一段时间内。所以类似于保修的概念，如果治疗效果不持久，如果基因疗法的预期效果没有持续三年、五年或十年，没有像最初预期那样提供持久的疾病治愈，公司可能需要退还部分费用。

Yeah. I think when you especially when you're dealing with gene therapy where it's sort of a in many cases, it's gonna be a one time administration and you're hoping to cure a disease. But if you think about it, you're not you're not gonna be delivering the therapy over the course of ten years or the lifetime, so you have to recoup all the investment, all the returns in that single administration, and that's why the cost of these therapies are so high. In those cases, what you might wanna do is try to amortize that cost out over time. And so it's something like a warranty where, you know, if the treatment isn't durable, if the gene therapy, if the effect that it's intended to have doesn't last three years or five years or ten years, it doesn't offer a durable cure for the disease like was anticipated at the outset, the company might owe back some of the money.

因此，与其将所有成本都集中在一次治疗中，公司仍需对产品长期效果负责。因为在许多临床试验中，当这些基因疗法进入市场时，部分不确定性在于你不知道效果是否会持久。临床试验可能持续两年，患者在试验中的临床反应可能持续两年，但五年后，这种效果还会存在吗？事实上，我们已经看到其中一些疗法的效果会随时间减弱。因此，保修或找到分摊成本的方法，是应对这些高昂前期费用的一种方式，同时也与产品预期带来的疗效不确定性相协调。

So rather than load it all into that single administration, the company's still on the hook for the performance of that product over time. Because in many of these clinical trials, when these gene therapies are coming onto the market, part of the uncertainty is you don't know whether or not the effect's gonna be durable. The clinical trial might last two years and the patient's clinical response might last two years in the clinical trial, but then five years out, is that effect still gonna be there? And in fact, we've seen with some of these some of these therapies that the benefit wanes over time. And so a warranty or finding a way to amortize the cost over time is one way to deal with those very high upfront costs that aligns with some of the uncertainty around benefits that the products intend to deliver.

那会有什么效果？我的意思是，它不会改变前期成本，对吧？或者说它如何改变改善现状的动机？我可能没太理解。

What would that do? I mean, would it it wouldn't make the upfront cost different, would it? Or is it does it how does it change the incentives to improve the situation? I'm not maybe I'm not following it.

嗯，最简单的形式可能是公司在五年内分期获得付款，比如保险公司第一年支付X金额，第二年支付X金额，持续五到六年。他们会说，如果疗效在第二、第三或第四年开始下降，我们就不需要支付后续款项。这样就不是一次性支付所有费用，而是将费用分摊到一段时间内，同时公司要对药物在这段时间内的效果负责。

Yeah well you know in its simplest form you might think of something where the company gets paid a certain amount over five years and they'll let's say the insurance company will pay x amount in year one, x amount in year two, out to five or six years. And they'll say, well if the benefit starts to decline in year two or year three or year four then we won't owe those other payments. That way instead of paying all the money upfront, they spread it out over time and the company was on the hook for the performance of the drug over that period of time.

哦，我明白了，明白了。

Oh I see, I see.

那是最简单的方式。当然还有其他融资方案可以考虑，但这是最简单的。不过这会变得很麻烦，因为保险公司会说，跟踪患者五年会很困难。我们会失去这个患者，他们可能会更换保险计划。

That would be the simplest way, yeah. There's other kinds of financing schemes that you could think about, but that would be the simplest. You know, but it becomes cumbersome because what the insurance company will say is, you know, it's gonna be hard to follow the patient for five years. We're gonna lose that patient. They're gonna switch insurance plans.

我们如何持续追踪他们？因此必须考虑如何在市场中建立中介机制，以便持续收集数据并让制药公司对结果负责。从实际操作来看这些措施更难实施，但从政策层面考量则非常合理。

How do we continue to follow them? So then you have to think of how do you create intermediaries in the market where you can continue to collect that data and hold the drug company sort of accountable for the results. So they're harder to implement these things as a practical matter. They make a lot of sense as a policy matter.

这就是价值医疗的核心所在——与按服务收费模式相对。理论上听起来很棒，实践中却极难落实，因为'价值'本质上是个主观概念。对事物进行价值评估通常涉及某种主观性，我认为这正是医疗系统往往默认采用按服务收费编码模式的原因——我做这个手术，就能获得相应报酬，事先就能明确收益等等。

Well that's the whole thing about value based medicine, right, as opposed to fee for service. It in theory sounds great in practice. That is very difficult to implement because value is inherently a word that subjective in nature. To place value on something is usually some kind of There's usually some kind of subjectivity involved, and I think that's why the system often defaults to this sort of fee for service coding where, okay, I do this procedure, I get this much money, I know that ahead of time, etcetera, etcetera.

关注投入而非结果，确实如此，因为衡量投入比衡量结果更容易。

Focusing on the inputs rather than the outcomes, yeah, because it's easier to measure the inputs rather than the outcomes.

确实如此。我不确定我们是否已找到实施价值医疗的方法。或者你认为我们已经找到了？还是某些保险公司在这方面遥遥领先？是否存在可供借鉴的优秀模式？我知道我们话题转得有点突然。

Very much so, and I don't know if we've figured out how to do that value based care. Or or do you think we have? Or or or some insurance companies well ahead on that? Is there a good model that we would look to? I know we're changing subject pretty dramatically here.

我认为联邦医疗保险和其他私营保险公司已开始更有效地推进这项工作，特别是观察当前医疗交付体系时，越来越多服务提供方采用按人头付费模式——他们收取固定总金额并承担风险。他们正开始投资于测量大规模人群临床结果和临床表现的方法。因此我认为，用于关联支付的测量工具已大幅改进，基于这些结果指标的保险合约也越来越多。

I look. I think I think Medicare and other private insurers have started to do it more effectively, and particularly when you look at the delivery system now where more providers are capitated, meaning they're they're taking they're getting paid a lump sum and taking risk. They're starting to invest in ways to measure their clinical outcomes and their clinical performance over large populations. And so I think that the tools for measuring the things that you would tie the payment to have gotten a lot better, and you're starting to see more insurance contracts get written based on those outcomes, the measures of those outcomes.

最后一个问题——我们已讨论一小时了——您认为人工智能将如何影响这场对话，特别是在FDA监管方面？委员会已就此举行过几次听证会。我对AI讨论的担忧在于，很多时候我们讨论的只是高级软件而非真正AI。AI介入医疗最有趣的部分或许是诊断领域——我们是否会发展到与AI医生对话获取诊断的阶段？这种诊断能否获得FDA批准？

What about, last question before we go, it's been an hour. How do you think AI starts to play into this into this entire conversation, especially with FDA regulation? We've had a few hearings on it, a few discussions on it in the committee. My problem with the discussion of AI is always that it's oftentimes that we're just talking about advanced software, not really AI. I think the real interesting part about how AI could play into healthcare is when we start talking about a diagnosis, and can you, are we gonna get to a point where you you have a conversation with an AI doctor and they diagnose you, and will that ever be an FDA approved decision?

谁知道呢？我甚至不知道...当然测试方法总是有的，但您怎么看这个问题？

Who knows? I how do you even I guess there's ways to test it, of course, but I mean, it's what do you think of it?

是的，我认为它将会实现，也能够实现。将AI与药物研发分开来看，这些源自自然语言处理和大语言模型的药物，我认为将会出现由AI增强流程辅助开发的药物。这些药物仍需经过传统监管流程，必须完成临床试验。而且我认为AI工具在该场景中能帮助解释药物的生物学合理性，说明为何该药物应对特定疾病有效。

Yeah, look, I think it will be and I think it can be. So separating out AI and the drug development side, the drugs that come out of some of these natural language processing, these large language models, I think there will be drugs that are where where an AI augmented process are helping to develop drugs. Those drugs are gonna be put through a traditional regulatory scheme. They're still gonna have to go through clinical trials. And if anything, I think the AI tool in that scenario is going to help explain the biological plausibility of the drug, why the drug should work in a certain disease.

因为AI工具解决的是合理性问题。若能更好地理解药物为何对特定疾病状态有效，实际上有助于简化监管流程。FDA越了解药物在特定情况下应如何发挥作用，并通过临床数据加以验证，就越能增强监管机构的信心。在医疗器械方面，使用AI工具或大语言模型直接与患者互动辅助诊断，甚至在某些情况下做出诊断，我认为当前技术已足以支持在低风险、常规场景中应用这些工具。

Because what AI tools solve was plausibility. And so to the extent that there's a better understanding of why the drug would work in that certain disease state, that actually could be helpful in making the regulatory process more streamlined. Because the more that FDA understands about why a drug should be working in a particular circumstance and you go on and demonstrate that with clinical data, that gives the agency more confidence. On the medical device side, so using an AI tool, a large language model to actually interact with a patient and help facilitate diagnosis, in many cases even potentially make the diagnosis. I think the technology is sufficient right now probably to be using these tools in certain lower risk, more routine circumstances.

挑战在于没人真正愿意让大语言模型走完监管流程，因为人们不清楚FDA会如何对待。因此目前这些大语言模型被定位为临床决策支持软件。他们声称这种AI模型旨在为医疗提供者提供信息支持，给予更多洞察。比如你登录远程医疗平台时，初始交互可能是AI工具询问症状并形成初步诊断，再将信息转交给医生进行后续诊疗。这种情况下，该工具就是用于临床决策支持。

The challenge is nobody really wants to take an LLM through the regulatory process because people are unclear about how FDA would approach it. And so they're positioning these large language models right now as what they call clinical decision support software. So they're saying, well, this this AI model is to help inform the provider and give the provider more insights. And so you might go on to a telehealth platform and your initial interaction might be with an AI tool that's asking you questions and formulating a presumptive diagnosis, and then it now passes that information on to the doctor and then the doctor interacts with you or the provider interacts with you. So the tool in that circumstance is for clinical decision support.

我相信在特定低风险场景中，AI工具不仅能与患者互动，甚至可开具治疗方案。想想人们去看初级保健医师的主诉——排尿疼痛、背痛、腹痛等常规问题。虽然部分可能很严重，但AI工具可能已足够智能，能根据问题判断何时转交医生处理，何时按常规流程处置。这类应用如何获得监管批准？FDA无法拆解AI模型的工作机制。

There is the capability I believe in certain lower risk settings with an AI tool to not only interact with a patient, but you know, even prescribe therapy. If you think about what the chief complaints are that most people go to a primary care physician with, it's you know, pain on urination, back pain, belly aches, it's routine things. Some of them could be serious, but the AI tools probably are sophisticated enough to know when to default to a provider and when it's something that's more routine based on the questions that they have, where it can be dealt with more in a more routine fashion. How do you get regulatory approval for that? FDA is not gonna be able to rip apart the AI model.

这是个黑箱系统，他们无法弄清其运作原理。因此监管需要从前后两端入手：FDA需要规范训练数据集，确保用于训练AI工具的临床数据足够可靠、具有代表性；同时要用标准化问题集测试AI工具，确保其回答始终准确或至少达到医疗提供者的准确度。这套监管框架尚未完全建立。

It's black box, then they're not gonna be able to figure out how it's working. And so the regulation needs to be put on the front end and the back end. What I mean by that is FDA needs to figure out how do we regulate the training sets, the clinical data training sets that these AI tools are trained on to make sure that the data itself is reliable enough, is representative enough, so the AI tool is being trained on an appropriate body of data. And then how do we query the AI tool with a sort of standard battery of questions to make sure that the answers that it's giving are consistently accurate or accurate enough, as accurate as a provider would be. And that regulatory construct isn't isn't fully created.

我认为首批通过监管审批的企业很可能是资金雄厚的远程医疗提供商。他们需要逐个适应症攻克——先针对这种主诉获批，再那种主诉，积累足够案例后，才可能获得初级保健场景的通用审批，这还有很长的路要走。

I think the first companies that bring these tools through regulatory approval, what they're gonna have to do, it's probably gonna be some telehealth provider who's well funded. And what they'll have to do is bite off indication by indication. So they'll have to get approval for this chief complaint, and that chief complaint, and that chief complaint, and couple enough of these together where then maybe they can get approval for a general use case in a primary care setting, which would be a long way off.

这很难，但我还是要问。如果最终审批权始终在人类医生手中，为何还需要FDA批准？因为从概念上说，这与医生自己进行大量网络研究得出结论，或者让助手（比如住院医师）做同样工作后再复核，并没有本质区别。

It's hard, but I still have to ask. I mean, so if if if the final approval is always still a human physician, why do you need any FDA approval? Because there's really conceptually not much difference between that physician doing a lot of Google research and just doing a lot of Internet research themselves and coming up with an explanation, Or having their assistant do it let's say. Another human resident do the same thing and then they check the work to make sure it makes sense.

对，所以当你...哦抱歉。我是指在医生作为最终接触点的情况下，那不会获得FDA批准。

Right, so when you're Oh sorry. I mean in the circumstance where the physician is the final touch point, that wouldn't be FDA approved. That

那么目前就是这样运作的吗？我们

would And so is currently how it is? Are we

正在使用这些工具...现在这样是可以的。这部分是被排除在外的。FDA确实监管某些临床决策支持软件，但通常不监管这些工具，只要最终决定权在医生手中。它的作用是询问患者、准备资料、提出初步诊断。

using And that's these tools for okay right now. That's okay. So that's carved out. FDA does regulate some clinical decision support software, but typically isn't regulating these tools, as long as the physician, as long as the tool itself ultimately defaults to the physician. So what it's doing is interviewing the patient, preparing, coming up with a presumptive diagnosis.

现在我明白了。

I understand now.

但那不是真正的价值所在，对吧？你一开始就说明了。

But that's not the real value, right? You laid it out at the outset.

不，是的，最终...比如在太空飞船上，可能是企业号之类的，你需要一个全息医生，因为它不消耗氧气。

No, yeah, eventually, yeah, eventually, no, if you're on a spaceship for, maybe it's the Enterprise or something like that and you need to have a holographic doctor on board because it doesn't take up any oxygen.

甚至不需要食物。即使是诊断流感或链球菌性咽炎这类简单病症，你也可以将诊断测试与经过相关训练的大语言模型交互结合。我认为现有工具已足够处理这些低风险干预，并知道何时转交给专业医生，但没人开发这类工具，因为他们担心监管合规的成本问题。

But even Doesn't need food. Even diagnosing you know flu or strep throat or something simple where you can couple a diagnostic test with an interaction with a large language model that's trained on that. I think that the tools now are capable enough to handle those lower risk interventions and know when to default to a provider, but nobody's developing the tools for those purposes because they're worried about, you know, they're worried about how they're gonna be regulated and the cost of complying.

好的。也许在这些听证会上，老实说，对话效果并不理想，因为我认为并非所有人都理解人工智能，但您已经非常清晰地为我们阐述了需要提出的问题。也就是说，需要FDA介入，需要为这个未来设定参数，这样技术开发者才有动力去研发这项技术。不过，如果您讨论的是辅助医生，我的意思是，这项技术真的有那么不同吗？显然它没那么先进，但也不该有太大差异，对吧？因为即使是临床辅助技术，其目标也应该是提供最终答案，只是碰巧需要医生的批准。

Okay. So maybe that's the in these hearings honestly, the the conversations are not good because I don't think everyone understands AI, but I think you've laid it out very clearly for us on on the questions we need to be asking, which is, you know, you need to FDA, you need to start setting parameters for this future so that the technology actually innovators have an incentive to develop the technology in the first place. Although, if you're talking about assisting a doctor, I mean, why is the technology all that different? Mean, it's clearly not as advanced, but it shouldn't be that different, right? Because the goal for, even if it's clinician assisted technology, should be to provide a final answer, it just happens to need approval from a doctor.

顺便问一句，就这类软件而言，目前有什么特别令人印象深刻的产品吗？

There something out there that's really impressive, by the way, as far as that kind of software is concerned?

其实已有工具能实现这些功能，它们整合到电子病历系统中，可以查询医生或查看化验结果，做出某种推定分析或诊断，并向医生提出询问。这类机器学习工具已存在一段时间了。想想放射科医生用来诊断X光结果的工具，或病理学家用来观察玻片的工具。这些并非我们通常认为的像ChatGPT那样的大语言模型，而是通过模式识别进行工作的机器学习形式。

Well look, there's tools already that have been doing that that layer on to electronic medical records and query physicians or look at lab results and make sort of a presumptive analysis or diagnosis and inquiries of the physicians. These kinds of tools, these sort of machine learning tools have existed for a while. And if you think about tools that radiologists are using to diagnose findings on x rays, or pathologists are using to look up at the slides. Those aren't true, those aren't the large language models that we think about with like ChatGPT. Are forms of machine learning where they're basically engaging in forms of pattern recognition.

它们基于封闭数据集训练，通过大量样本学习识别特定特征。严格来说，我不会称其为...

They're trained on closed data sets and taught how to identify certain findings by being shown a lot of Right, I wouldn't call

人工智能。对，确实不会

it AI. Yeah, would

他们称之为人工智能的一种形式。机器学习属于人工智能范畴。您所说的真正机遇在于大语言模型。我正在做一项分析，将医师执照考试题目输入不同大语言模型，看看哪个能成为最优秀的医生。根据我目前的初步体验和研究结果，Anthropic公司的Claude在回答医学问题上可能表现最佳，优于谷歌的Gemini Advanced和ChatGPT4.0。

They call it are form of AI. Machine learning is a form of AI. The real opportunities that I think you're talking about is the large language models. So I'm doing an analysis right now where we're feeding the medical licensing exam to the different large language models to see which is the best doctor. But I think right now based on my kind of cursory experience with these and doing some of this work, it looks like Claude from Anthropic may be the best at answering medical questions better than Google's Gemini Advanced and better than ChatGPT four point o.

是啊，这确实会很有意思

Yeah. Well, it'd be interesting to see

如果我们走出去，把它当作医生来使用，它看起来确实，你知道，目前是可靠的。

if we'll go out and use it as a doctor, it does look, you know, it's reliable right now.

是的。医生斯科特·戈特利布对大家说，你不需要看医生，谷歌一下就行，没问题的。这就是

Yep. Doctor. Scott Gottlieb, everyone, says, you don't need to see a doctor, just Google it, it'll be fine. That's what

他们说的。

they said.

那会很有趣。我们期待继续与您合作，探索改进FDA政策的途径，因为我认为，最终能让人们获得他们所需的医疗保健创新。所以，斯科特·戈利医生，再次感谢您的参与。非常感谢。

That would be interesting. We look forward to continuing to work with you on ways we can improve FDA policy because, and and I I think, in in the end, get people the access to innovations in health care that they need. So Doctor. Scott Golley, thanks for being on again. Appreciate you.

非常感谢。

Thanks a lot.