Anatomy of a Bulletproof IND — The Nonclinical Sections Explained

Your IND is rejected before a single human being ever reads it. Not because the science is wrong — because a 10-year-old legacy file was missing a digital tag. In this episode, we break down exactly what goes into the nonclinical sections of an IND, how the 5-module eCTD structure works, and the SEND dataset rules that silently kill submissions before they ever reach a reviewer.

Key takeaways:

• The IND is organized into 5 modules — and the nonclinical program lives primarily in Module 2 (summaries) and Module 4 (study reports and SEND datasets)
• Module 2.4 (Nonclinical Overview) is written last — after the detailed 2.6 summaries are complete — because it summarizes them
• Every toxicology study listed in Module 2 must have an associated study report in Module 4, and vice versa — no orphans allowed
• SEND datasets are required for almost all tox studies, including nonGLP studies — and a missing SEND dataset triggers automatic rejection before a human reviewer ever sees your data
• Even studies run 10 years ago still need at minimum a TS domain to pass validation — age of the study is not an exemption

Links:

• My course: The Complete Guide to Nonclinical Development, https://www.nonclinical.academy/
• Work with Dessi: toxistrategy.com
• Read the full newsletter issue on LinkedIn: https://the-nonclinical.com/

The Nonclinical is hosted by Dessi McEntee, MS, DABT — board-certified toxicologist and Fractional Head of Toxicology. Subscribe to the newsletter on LinkedIn, take the course at nonclinical.academy, or work with Dessi at toxistrategy.com.

Transcript

SPEAKER_01 0:00

Imagine the scenario. You're a biotech leader and you've spent like five years and maybe 40 million dollars developing a compound that just performs miraculously in your early preclinical models.

SPEAKER_00 0:12

Oh, yeah. The absolute dream scenario.

SPEAKER_01 0:14

Right. It's the breakthrough your startup has been betting its entire runway on. So your team has been working late nights for months, compiling all this paperwork to get clearance for human trials. You finally submit your massive application to the FDA, and then it is immediately rejected. Just devastating. And it's not even rejected by, you know, a panel of seasoned scientific reviewers who found some flaw in your mechanism of action. It's rejected by a computer algorithm before a human being ever even lays eyes on your data.

SPEAKER_00 0:43

Yeah, and all because like a 10-year-old legacy file was missing a single digital tag.

SPEAKER_01 0:49

Exactly. Welcome to the high-stakes reality of the investigational new drug application, or IND.

SPEAKER_00 0:55

It is genuinely the ultimate nightmare scenario for any biotech leader. Because that transition from pure scientific discovery into the realm of strict regulatory compliance is arguably the most jarring shift a development team will ever go through.

SPEAKER_01 1:10

Yeah, it's a totally different world.

SPEAKER_00 1:12

It really is. You are leaving behind this unbound exploration of biology and stepping into a remarkably rigid, unforgiving architecture where, you know, a single missing data set can literally halt your entire pipeline.

SPEAKER_01 1:25

So if you are listening to this right now as a biotech leader managing early stage drug development programs, we know you carry an immense amount of pressure. Your fundamental goal is getting that life-saving compound to first in human or FIH clinical trials. Aaron Powell Right.

SPEAKER_00 1:39

That's a whole ballgame.

SPEAKER_01 1:40

Right. And today we are doing a deep dive into a brilliant set of source notes on the anatomy of an IND provided by Desi McInty. She is a head of toxicology, a biotech board member, and she actually specializes in building what she calls bulletproof toxicology programs.

SPEAKER_00 1:55

Aaron Powell Her insights are just absolutely critical for anyone steering a drug through this phase. Yeah. Because she understands that the IND is well, it's much more than just a scientific dossier. It is a highly engineered compliance document.

SPEAKER_01 2:08

Aaron Powell Yeah. The regulators don't just care about the science. They care about how the science is organized, how it's linked, and how it's presented.

SPEAKER_00 2:13

Aaron Powell Exactly.

SPEAKER_01 2:14

So the mission for this deep dive is to demystify this colossal regulatory puzzle, specifically drilling down into the extremely complex non-clinical sections. We want to give you the blueprint so you can guide your team to a seamless clearance without facing those devastating automated validation rejections.

SPEAKER_00 2:34

It's all about avoiding the algorithm's wrath.

SPEAKER_01 2:36

For sure. Okay, let's unpack this. What exactly is the structural foundation we are dealing with when we talk about a modern IND?

SPEAKER_00 2:43

Well, if we connect this to the bigger picture, the IND is your formal legal request to the FDA to administer an experimental drug to humans. And decades ago, this was a monumental physical undertaking. Like sponsors would wheel in literal shopping carts full of massive three-ring paper binders to the FDA offices.

SPEAKER_01 3:02

Oh wow, literal shopping carts.

SPEAKER_00 3:04

Literal carts. But today we operate entirely within a digital system known as the ECTD. That's the electronic common technical document.

SPEAKER_01 3:13

Which, I mean, that sounds great on the surface. Moving from paper to digital usually implies things get easier or like at least more flexible.

SPEAKER_00 3:20

You would think so, right. But the digital format actually demands far more precision. The ECTD is built upon five mandatory pillars or modules.

SPEAKER_01 3:30

Okay. Lay them out for us.

SPEAKER_00 3:32

So first you need the administrative nuts and bolts, things like investigator information and cover letters. That's module one.

SPEAKER_01 3:38

Makes sense.

SPEAKER_00 3:38

Then the FDA wants a high-level summary of absolutely everything, and that lives in module two.

SPEAKER_01 3:44

So module two is basically the executive briefing package.

SPEAKER_00 3:47

Precisely. It's the big picture. After that, we get into the raw data. Module three is all about your compound itself. So the quality, which is grounded heavily in your chemistry, manufacturing, and controls data, or CMC.

SPEAKER_01 3:59

Right, the actual physical drug.

SPEAKER_00 4:00

Exactly. Then module four contains all your animal and bench data, meaning your non-clinical study reports. And finally, module five holds whatever human clinical data you might have, complete with their SDTM datasets, which is simply the standard data tabulation model for clinical info.

SPEAKER_01 4:18

Okay, I really like visualizing this five-module ECTD structure, like building a highly regulated skyscraper. You are handed the architectural blueprint by the FDA and you just cannot deviate from it.

SPEAKER_00 4:31

You really can't.

SPEAKER_01 4:32

Like you can't just decide to put a window where the FDA demanded a load-bearing wall simply because you think it creates a better narrative flow for your drug. In fact, Desi's notes emphasize that even if a room in this skyscraper is completely empty because you didn't need to run a certain type of study, you still have to build the door to that room. And you have to hang a sign on the door explaining exactly why the room is empty.

SPEAKER_00 4:52

That analogy perfectly captures the strictness of the system. You are required to include the exact headers and sections as outlined by the regulators. No exceptions. No exceptions. If you skip a study, let's say a certain type of carcinogenicity study just wasn't applicable to your compound stage yet. You still must keep the specific header in the document and explicitly write a scientific justification for its absence.

SPEAKER_01 5:18

Because if you just delete the header.

SPEAKER_00 5:20

Right. If you just delete the header, the FDA reviewer doesn't know if you strategically chose not to run the study or if you simply forgot to include it.

SPEAKER_01 5:27

Uh, okay. So the rigidity isn't just bureaucratic red tape meant to make a regulatory affairs team miserable.

SPEAKER_00 5:34

None at all. I mean, it ensures that FDA reviewers can universally navigate complex scientific data across thousands of submissions. When a reviewer opens module three, they need to know exactly where the manufacturing data is without having to decipher a brand new organizational system for every single drug sponsor that comes across their desk. Standardization is the only way the agency can function at scale.

SPEAKER_01 5:55

That makes perfect sense. So if I am a biotech leader pushing for first in human trials, where is my team living inside this skyscraper?

SPEAKER_00 6:02

Prior to human trials, your team is spending almost all their time and resources in the non-clinical engine room, which consists of modules two and four. Because this is where the early stage safety profile of your drug is actually proven and documented.

SPEAKER_01 6:16

Yeah, I noticed in the source material that there is a very strict symbiotic relationship between module two and module four. Is it just a matter of ensuring the summaries match the reports?

SPEAKER_00 6:26

It's more than that. It's a literal one-to-one mapping. Module four holds the actual full-length non-clinical study reports. Module two contains the summaries of those exact reports. Got it. So for every single study you list in module two, there absolutely must be a corresponding full report sitting in module four. And the reverse is true as well. You cannot have a rogue study report floating around module four that isn't summarized and indexed in module two.

SPEAKER_01 6:51

It's a mirror image. The FDA wants to be able to read the summary, click a link, and immediately land on the raw data, backing up that summary.

SPEAKER_00 6:58

That traceability is paramount. Yeah. And you know, the FDA wants to see your data presented conceptually. So within module two, there is a very specific hierarchy, primarily living in section 2.6, which contains the highly detailed written summaries. First, they want to know what the drug does to the body. That's your pharmacology.

SPEAKER_01 7:18

And according to the notes, that isn't just the primary mechanism of action. That includes all your secondary pharmacology, like the off-target effects, plus your safety pharmacology, so cardiovascular, respiratory, and central nervous system studies.

SPEAKER_00 7:31

Exactly. Next, the FDA wants to know what the body does to the drug. That's pharmacokinetics. This section summarizes your PK studies in the standard ADME order.

SPEAKER_01 7:41

Which is absorption, distribution, metabolism, and excretion.

SPEAKER_00 7:44

Right. And finally, they want to know how the drug might harm the body, which brings us to the toxicology written summary.

SPEAKER_01 7:50

Which is the heavyweight section, right? Single dose, repeat dose, genotoxicity, reproductive taukases. All of that incredibly dense data gets summarized here.

SPEAKER_00 7:58

It's massive.

SPEAKER_01 7:59

But Desi's notes highlight a very specific authoring workflow regarding this detailed 2.6 section and the broader executive summary, which is section 2.4. The rule is that section 2.6 must be written before section 2.4. So wait, what does this all mean? Shouldn't a writer outline the broad executive overview first and then fill in the granular details underneath it?

SPEAKER_00 8:21

If you approach this like writing a novel or a standard business proposal, that workflow seems logical. But in biotech regulatory writing, the granular details must dictate the summary. Think about the massive risk involved here.

SPEAKER_01 8:35

Okay, what kind of risk?

SPEAKER_00 8:36

If you write a sweeping executive overview in section 2.4 first, you're essentially making promises about your drug's safety and efficacy based on your overarching hopes or maybe just preliminary impressions from the lab.

SPEAKER_01 8:48

Oh, I see. It's like writing the glowing press release for your clinical trial results before you've even unblinded the data. You are boxing yourself into a narrative that the ground level data might not actually support.

SPEAKER_00 8:58

That is the perfect way to look at it. When your medical writers finally get down to synthesizing the raw data from Module 4 into the detailed 2.6 summaries, they might uncover nuances, like, say, a slight elevation in liver enzymes at a certain dose, or a minor off-target binding affinity.

SPEAKER_01 9:16

And those nuances could contradict the broad claims you just made in your 2.4 overview.

SPEAKER_00 9:21

Exactly. And if the FDA catches a contradiction between your broad summary and your detailed data, your credibility is instantly shot.

SPEAKER_01 9:29

Entirely. So you have to synthesize all the raw data in 2.6 first to establish the absolute ground-level truth of your compound.

SPEAKER_00 9:37

Aaron Powell Right. Only after that foundation is set in stone can you zoom out and write the overarching narrative in 2.4. It ensures your executive summary is perfectly accurate and doesn't accidentally introduce unverified claims.

SPEAKER_01 9:49

Okay, so you build the argument from the raw data up, never top down. That is a critical operational shift for a leadership team to enforce. But as we established at the beginning of this deep dive, the logic of your arguments won't really matter if your application gets bounced by a computer program before a human even reads it.

SPEAKER_00 10:05

Yeah, and this is where we have to talk about send datasets, standard for exchange of non-clinical data. This is the hidden tripwire in module four that can blow up an entire IND instantly.

SPEAKER_01 10:16

Here's where it gets really interesting. Let's walk through this because this feels like the area where biotech executives are most vulnerable if they aren't intimately involved in the IT side of their submissions.

SPEAKER_00 10:26

They are incredibly vulnerable here.

SPEAKER_01 10:28

The general rule outlined in the notes is that almost every main toxicology study you list in module four must have an associated send data set.

SPEAKER_00 10:37

Yes, and that scope includes single dose, repeat dose, carcinogenicity, and safety pharmacology studies. But the trap here is that the send requirements are a constantly moving target. They depend on the specific type of study, the date the study was initiated, and even which specific FDA center you were submitting to.

SPEAKER_01 10:56

And one of the biggest traps teams fall into is the assumption about GLP versus non-GLP studies. Good laboratory practice? Biotechs frequently run non-GLP studies early on because they are faster and cheaper to execute when you're just looking for early safety signals, right? And there is this assumption that because it isn't a formal GLP study, it might fly under the send radar.

SPEAKER_00 11:16

It is a very common and very dangerous assumption. The FDA demands standardized data regardless of the GLP status. If you decide that a single dose non-GLP study is important enough to include in your toxicology package for the IND, it absolutely unequivocally needs a send data set.

SPEAKER_01 11:34

Wow. Which brings us to an even more insidious problem, relying on historical data. A lot of early stage biotechs acquire compounds that have been sitting on a shelf at a larger pharma company, or maybe they rely on foundational animal work done a decade ago at a university.

SPEAKER_00 11:49

Very common practice.

SPEAKER_01 11:50

DESI notes that even for a 10-year-old non-GLP study, you still need, at a minimum, what's called a TS domain to pass the digital validation. So you're telling me a 10-year-old non-GLP study can trigger an automatic rejection of a multimillion dollar modern drug application just because it's missing a specific data domain.

SPEAKER_00 12:09

Yes, absolutely. And we should clarify what that means. A TS domain simply stands for trial summary. It is essentially a standardized digital tag, a metadata wrapper that tells the FDA's computer systems what it is looking at. But if you don't have that simple digital tag on a 10-year-old file, what's fascinating here is how ruthless the automated validation process is. Let's paint that Friday night scenario we talked about earlier. Your team has been running on fumes for six months. You've ordered the pizzas, everyone is gathered around the conference table, and you finally hit submit on the FDA portal.

SPEAKER_01 12:42

I can feel the tension.

SPEAKER_00 12:43

Right. The portal's algorithm immediately scans the ECTD architecture. If it detects that a legacy toxicology study is listed in module four, but it lacks that required TS domain, the portal doesn't politely flag it and route it to a human reviewer with a warning note.

SPEAKER_01 12:59

It categorically rejects the entire multimillion dollar IND submission.

SPEAKER_00 13:04

Instantly, you get an automated failure notice. The clock on your first in-human trials completely stops until your team can retroactively build and validate a digital data domain for an experiment that was conducted like during a different presidential administration.

SPEAKER_01 13:18

That is staggering. It highlights a profound shift in the role of a biotech leader. You aren't just managing brilliant toxicologists and pharmacologists to ensure the science is sound. You are fundamentally managing IT compliance.

SPEAKER_00 13:32

You really are. Your data architecture has to be just as flawless as your molecular architecture, or the science literally doesn't matter.

SPEAKER_01 13:39

That is the harsh reality of the modern ECTD format. The data standards are the gatekeepers. So with these incredibly strict structural rules, the top-down offering risks, and the looming threat of automated validation rejections, how does a leader practically guide their team through drafting thousands of pages of documents? You are racing against a funding clock. You can't just sit around waiting for every single study to be perfectly finalized before you start writing.

SPEAKER_00 14:04

You don't have to wait, but you have to operate with extreme discipline. The source material outlines how you can begin authoring the IND using audited drafts.

SPEAKER_01 14:13

And an audited draft is a study report that has been through quality assurance, the data is essentially locked, but it hasn't received the final formal ink signature from the study director, is that right?

SPEAKER_00 14:23

Exactly. Using audited drafts is a huge time saver for teams trying to conserve their cash runway. You can start drafting your 2.6 summaries based on these drafts, but it requires absolute militant alignment upon finalization.

SPEAKER_01 14:38

Militant alignment, I like that phrasing.

SPEAKER_00 14:41

It has to be. If the final signed report changes a single finding, say it modifies the description of an adverse effect from mild to moderate, or it slightly tweaks the NOAL, the no observed adverse effect level, you have to go back and update every single mention of it throughout the entire IND.

SPEAKER_01 14:57

Because of the cascade effect. Whatever gets written in those Model 2 summaries sets the bedrock foundation for literally every future regulatory and clinical document you will produce.

SPEAKER_00 15:06

This is particularly true for the investigators brochure, or IB. The IB is the comprehensive document given to the clinical investigators, the doctors, who will actually be running your human trials.

SPEAKER_01 15:18

Right. They need to know what they are dealing with.

SPEAKER_00 15:20

Exactly. They rely on it to understand the safety profile of the drug they are giving to patients. The source material notes that the non-clinical information of the IB should typically be copied and pasted directly from the IND summaries.

SPEAKER_01 15:34

I love this detail because it runs counter to everything we are taught about writing. You think back to grade school, and copy pasting was the ultimate sin. It was plagiarism. You were constantly told to put things in your own words. But in biotech leadership, copy pasting isn't plagiarism, it's regulatory version control.

SPEAKER_00 15:51

It is the golden rule of this process. Desse McIntyre explicitly states your cock study summaries are your guiding light. You actively want the text to be identical. You do not want a rogue medical writer getting creative with a thesaurus when describing a hepatic adverse event in the investigator's brochure.

SPEAKER_01 16:09

Because if a reviewer sees hepatocellular hypertrophy in the IND, but then the IBE calls it liver enlargement, they are going to wonder what else is lost in translation.

SPEAKER_00 16:20

It goes much deeper than just translation. In the eyes of the FDA, consistency is a direct proxy for scientific reliability.

SPEAKER_01 16:28

Oh, that's a great point.

SPEAKER_00 16:29

Yeah, if your narrative shifts between documents, even just rhetorically, it implies to the reviewer that your understanding of the drug safety profile is unstable. They will immediately lose trust in your entire application. They'll start questioning the raw data, wondering if you're trying to hide something or if your team is simply disorganized.

SPEAKER_01 16:47

Consistency is a direct proxy for scientific reliability. That is a brilliant way to frame it for a medical writing team. It transforms the act of copy-pasting from a lazy shortcut into a critical compliance strategy.

SPEAKER_00 16:59

It forces the team to rely on the ground level data rather than their own narrative flair.

SPEAKER_01 17:04

Okay, let's step back and look at the gauntlet we've run today for the biotech leader. We started by exploring the rigid ECTD architecture, that five-module skyscraper, where you can't alter the blueprints, and you must hang a sign on every empty room.

SPEAKER_00 17:18

Right. We then ventured into the non-clinical engine room, mapping the raw data in module four directly to the dense summaries in module two. We established why you must build your arguments from the raw data upward, writing the detailed 2.6 sections before the broad 2.4 executive overview.

SPEAKER_01 17:37

Just to ensure you never write a narrative your data can't support.

SPEAKER_00 17:40

Exactly.

SPEAKER_01 17:40

Then we navigated the terrifying automated gates of CEN datasets, learning that even a decade-old non-GLP study can trigger a catastrophic validation rejection if it is missing a simple TS domain tag. And finally, we embrace the strategic power of verbatim copy pasting, redefining it as regulatory version control to ensure absolute bulletproof consistency across your investigators' brochure.

SPEAKER_00 18:03

It is an immense amount of pressure, without a doubt. But as the source notes point out, once you understand the architecture and surrender to the strict flow of the hierarchy, the process actually becomes highly methodical and predictable.

SPEAKER_01 18:15

And predictability is exactly what you want when millions of dollars and patient lives are on the line. But you know, mapping out all these algorithmic tripwires and rigid hierarchies leaves me with a lingering thought that I want to pose to you, the biotech leader listening right now.

SPEAKER_00 18:29

It's a tension every leader feels in this space.

SPEAKER_01 18:31

Aaron Powell Exactly. Since the FDA demands such rigid hierarchy, automated data compliance, and verbatim copy-pasting to avoid even the slightest hint of contradiction, how do you maintain the compelling scientific story of why your early stage drug actually matters when you are forced to communicate within such an algorithmic, inflexible box? How do you keep the vital spark of that original biological discovery alive when you are consumed by the mathematics of the paperwork?

SPEAKER_00 18:59

It's the ultimate balancing act.

SPEAKER_01 19:01

It really is a puzzle every successful biotech founder has to solve. Thank you for joining us on this deep dive into the anatomy of an IND. May your toxicological summaries always be perfectly consistent, and may your Sunday data sets always pass validation on the very first try. We'll catch you next time.