Tapping into the regulatory goldmine: Interview with Ruxandra Teslo

In the past few months, the U.S. Food and Drug Administration (FDA) has taken steps toward greater transparency by publishing nearly 300 Complete Response Letters (CRLs), the detailed rejection notices sent to drug developers when an application falls short. For an industry used to working behind closed doors, these disclosures represent an important shift. It even recently allowed for the public reaction to the FDA’s reasoning on the rejection of Lykos’ therapy. Yet, while this new transparency is an improvement, some still think it is limited: the letters are selective, often redacted, and far from the exhaustive dossiers that could truly help innovators learn from past failures. 

That gap is at the heart of Biotech’s Lost Archive, an essay by genomics researcher Ruxandra Teslo. In it, she argues that the FDA is sitting on an unparalleled reservoir of regulatory knowledge locked away in Common Technical Documents (CTDs), the submissions that cover everything from manufacturing processes to clinical trial design. Teslo proposes a novel way to unlock part of this archive by using U.S. bankruptcy law to recover the regulatory filings of failed biotech companies and make them accessible for training artificial intelligence (AI) tools and guiding future developers.  

Labiotech wanted to learn more about it, so I had a conversation with Ruxandra Teslo. 

Regulatory opacity: the CTD goldmine  

Could you explain the core idea of the proposal you made in your essay? What does it aim to accomplish? 

Whenever a company submits an investigational new drug (IND) application to the FDA, it has to include in its documents, details from manufacturing processes to clinical trial protocols. And as the application progresses, this document is going to be enriched with the conversations between the company and the FDA. This is called the Common Technical Documents (CTDs), and while they seem like a series of boring details from the outside, they’re a potential goldmine for the industry. This is the starting point of my proposal: how to make these CTDs accessible to companies.  

CTDs contain knowledge that could save a lot of time and money. Recently, Ultragenyx’s therapy has been delayed due to manufacturing issues, and a six-month delay can cost hundreds of millions for companies.  

The thing is, these documents are not for the FDA to give away to the public, although it would be beneficial for public health. You don’t want to release some companies’ trade secrets. 

The FDA recently released about 300 CRLs in the name of transparency. How do you perceive this initiative, and how do CRLs compare to CTDs, the object of your proposal? 

I think it is an important step in the right direction, and I think the FDA is doing its best, considering it has to deal with companies’ trade secrets. The FDA’s hands are tied there.  

A CRL is significantly shorter than a CTD; it’s ten pages versus a thousand for a CTD. It is a much smaller and less detailed data set.  

In your essay, you argue that regulatory opacity creates friction within the industry. Could you maybe expand on how this plays out? 

I compare this opacity to living in a country where you would have law but wouldn’t have direct access to case law, so you wouldn’t know for sure how the law is applied. You could only have access to the reasoning of the decision by the intermediary of hired professionals – you would have to pay lawyers only to have access to case law.  

To me, that’s what the biopharma industry is today. I am focused on CTDs, and I only saw one of these documents only because a founder agreed to share it with me. These documents are hard to get your hands on.  

The result of that is that biotechs don’t really know the intricacies of the regulatory process. So, when they move on to clinical development, they have to hire these third-party vendors: manufacturing consultants, contract research organizations (CRO), contract development and manufacturing organizations (CDMO). And from having talked to biotech, these organizations will often tell you different things about FDA standards.  

Of course, more established biopharmas have the advantage over startups, as they have been through the process multiple times. A study from 2004 found that between 30 to 55% of the advantage enjoyed by large firms in approval timelines can be attributed to familiarity with the regulatory processes alone. 

In 2023, there was a discussion between the FDA and cell and gene therapy developers because they needed more clarity on the standards. The FDA said it couldn’t give too much guidance as it shouldn’t become prescriptive. Interestingly, what came out is that although the developers wanted more guidance, they were incentivized to keep their manufacturing methods to themselves, as it could represent a competitive advantage even if they didn’t compete directly with each other in this area. 

The proposal: How to make CTDs available 

Your proposal relies on acquiring orphaned FDA documents through bankruptcy law. Could you elaborate? Is it scalable? 

Biotechs don’t usually go through traditional bankruptcy, when you take on actual debt, but they go through various ways of dissolution. That means you can catch them in court and negotiate on the acquisition of their CTDs on a higher bidder basis. If there is a higher bidder, it probably means they want to pursue the development of the drug concerned by the CTD, which is great, but that is not what we want. We simply want to acquire the data and share it, but if someone wants to keep the development of the drug going, there’s no reason to stand in the middle. 

Since biotechs don’t go through traditional bankruptcy, you can acquire the CTDs during any proceeding of dissolution, and all you need is an asset transfer agreement. 

I have an example of a founder and CEO, who coincidentally worked as a lawyer for 10 years, and decided to do just that, bid on a CTD he needed for his own company. Around 50% of biotechs fail within five years, so I think there’s room for scalability in terms of the number of CTDs to acquire. It’s about finding a way to keep track of CTDs on the “market” in a smart way. 

Building a library of CTDs comes at a cost: scanning CTDs that are up for acquisition, actually acquiring them, and so on. Who should pay? 

Ideally, there would be some governmental involvement, and I think it should revolve around a philanthropic organization. There is no way to have a definite plan about who should found and contribute to the project yet.  

Also, I think the immediate cost is relative. The CEO I mentioned earlier, who acquired a CTD, paid $25,000 for it. So, if I am being generous, let’s say a CTD costs on average $50,000 a piece, and we were to acquire and make public 20 CTDs, that would amount to $1 million. Add another million to maintain servers, and your immediate costs are $2 million. To me, it seems a low price considering the value of the data and how it would benefit the entire industry. Of course, if the project takes shape and gains momentum, you can imagine some companies could donate CTDs. 

It’s not necessarily because it’s expensive that it hasn’t been done yet; I think it’s a question of awareness. 

Beyond the funding, I see no other option than a non-profit organization to manage the project if it becomes a reality. 

You mentioned CTDs were long documents, above 1000 pages, while it may be valuable data, it’s also hard to digest. How do you think the CTD library should be managed? 

I think AI would be an ideal solution to make the database as efficient as possible. The priority would be to allow an AI-assisted query to browse through and within CTDs to allow users to get the info they need as efficiently as possible.  

Then, depending on the size of the data set, you can imagine different functions. Auto-drafting some part of the IND submission is an example. Novo Nordisk said that using an AI tool to auto-draft its submission allows it to get the process done in 10 minutes instead of a few weeks – of course, it’s probably not 10 minutes since people will check the draft, but still. 

As the database increases and the AI receives queries, it could be trained to recognize patterns in regulatory practice  

Biotech’s lost archives: an idea with potential, with the concrete steps to figure out 

It’s hard not to root for the idea of a non-profit making extensive regulatory data available to biotechs, no matter their size. CTDs are indeed an untapped source that could help rebalance the playing field between companies that know the process inside out and newcomers forced to play a guessing game. A global database of CTDs would be the equivalent of opening up case law in a legal system where judgments are normally hidden, giving even the smallest biotech the ability to challenge the advice of consultants and CDMOs with hard evidence. 

But the path forward is far from straightforward. Who pays for such an archive, who governs it, and how much sensitive information can realistically be shared are all questions that remain unanswered. What Teslo’s proposal does achieve, however, is to reframe the debate: transparency is not just about publishing selective rejection letters, but about whether the regulatory knowledge of past failures should remain locked away or be repurposed to speed up new treatments to patients. The global idea is now out there, along with several keys to implement it; it’s up to the industry to pick it up and refine it.  

Ruxandra Teslo told me she had already received some positive feedback from founders following the publication of her essay, and that it will help her refine it, so the chances are this conversation isn’t over.