n April 2026, Japan’s Pharmaceuticals and Medical Devices Agency (PMDA) became the first regulatory authority in the world to mandate eCTD v4.0 for all new drug submissions. For someone like me who has followed eCTD developments closely since the early days of v4.0 adoption, this milestone is significant. Five months after PMDA began accepting v4.0 submissions on a voluntary basis in April 2022, my team completed what we believe was Japan’s first formal v4.0 submission. This experience highlighted a gap between a feature being available and a feature being genuinely useful in practice.

The Keyword Incentive Problem

Among eCTD v4.0’s most promising capabilities is its structured keyword system. Compared to v3.2.2, keywords enable richer document metadata, supporting faster review, improved cross-document referencing, and the kind of machine-readable structure that could one day underpin AI-assisted regulatory review.

In practice, however, keyword adoption has been inconsistent. Within one organization where I worked, the policy at the time was unambiguous: all optional keywords would be omitted. The reasoning was equally straightforward: there was no direct, tangible benefit to the applicant for including them.

This is not an isolated case. It reflects a structural dynamic that applies across most industries: When a feature is optional, it will be adopted in proportion to the perceived value it creates for the party bearing the cost of implementation. Where that value is unclear or accrues only to the regulator, adoption by industry will remain low.

In this case, three levers currently exist for changing this dynamic:

1. Agencies Demonstrating the Value
   PMDA has taken steps that other agencies would do well to emulate. The agency has publicly disclosed how it uses keywords internally during review (for example, grouping nodes by keyword can make the data easier to read) and has shown how those keywords surface in PMDA’s own eCTD viewer. This transparency signals to industry that the metadata they provide is actually being used—and used in specific, meaningful ways. It would be a welcome development if other authorities adopted a similar approach.
2. Vendors Reducing the Cost
   Submission management systems can play a significant role in shifting the cost-benefit calculation. For example, the departments responsible for compiling the eCTD and preparing its documents are often separate in Japan. The former has a deep understanding of the keywords themselves within the eCTD, while the latter can better determine which keywords should be assigned and verify whether the assigned keywords are correct. Designing a system that avoids this cross-departmental workflow and allows for cross-referencing documents, the eCTD framework, and a pre-organized keyword list to automatically and appropriately assign keywords might help alleviate the reluctance of field staff to assign keywords. When keyword assignment is automated, context-aware, and validated in real time, this barrier to compliance comes down. As vendors build smarter tooling, the threshold at which optional features become worth using changes.
3. Applicants Discovering Their Own Value
   The most powerful lever—and the most underexplored—is the one applicants hold themselves. Keywords are document metadata, and metadata have uses beyond regulatory submission. For instance, the same structured keyword data that support PMDA review could feed internal AI-assisted document classification, retrieval, or quality systems. Organizations that begin to see their submission data as a strategic asset will find incentives that external mandates cannot create.

Document Reuse: Useful in Theory, Costly in Practice

A second signature capability of v4.0 is document reuse: the ability to reference a previously submitted document rather than resubmit it in full. On paper, this reduces redundancy.

In practice, this experience was somewhat disappointing. Document reuse requires confirming, each time, what was submitted, where, and under which submission context. The operational overhead of tracking reuse eligibility outweighed the savings of not resubmitting.

The conclusion was pragmatic: resubmitting a previously submitted document every time was actually simpler.

This is not a failure of v4.0’s design. In Japan, eCTD submissions are divided into fine-grained Applications, making the submissions themselves simple enough that document reuse rarely delivers its full value in Japan. Unless the time-consuming effort of reuse is dramatically reduced, it will be difficult to justify adopting this feature. For example, if you do not plan to reuse the document, you can include the CTD document number in the document title. This makes it easier to create external links. If you want to create a link to the PDF for Section 5.3, you simply need to look at the far left of the title. Conversely, if you plan to reuse the document, you should not include the CTD document number in the title. This is because the same CTD number may not be assigned for the next submission. While the cost of paying attention to details like this is not significant, our team concluded that it outweighs the benefits gained.

In contrast, the EU—where documents are commonly shared across multiple regulatory authorities—stands to benefit more from avoiding resubmission. Similarly, in the US, where a single Application can be large in scope and may involve the same documents multiple times, resubmitting each time may not be the intuitive choice.

A Small but Telling Success

It would be unfair to catalog only these gaps. One episode from the transition period illustrates where v4.0’s architecture genuinely delivers.

In March 2025, PMDA added a Status column to all controlled vocabularies. The change looks trivial, but had it been attempted under v3.2.2, it would have required revising the eCTD schema itself through the ICH consensus process. For vendors of viewers and submission tools, this kind of update would have also meant substantial remediation work with the potential to ripple across entire systems.

Under v4.0, controlled vocabularies are designed from the start as an independently updatable variable region of the standard. Adding the Status column is a more significant change than simply adding a row. Even so, this change remained within the vocabulary layer, and PMDA was able to add the Status column without altering the standard itself. And because vendors build their products on the same premise, they accommodate such changes smoothly. The modular separation between the document specification and the regulatory vocabulary layer is not an abstract design principle; it is a practical capability that is already being exercised.

Deeper Value: Flexibility for an Evolving Landscape

The pharmaceutical modality landscape is expanding faster than any single regulatory framework can track in real time. Low-molecular-weight compounds, monoclonal antibodies, nucleic acid therapeutics, combination products, and modalities that do not yet have established categories are all moving through development pipelines simultaneously.

Under v3.2.2, incorporating a genuinely novel document type or classification requirement could require revising the standard itself—a slow, resource-intensive process with global coordination implications. Under v4.0, the controlled vocabulary and keyword layers can absorb much of that change without touching the core architecture.

This flexibility is v4.0’s most significant long-term contribution. It is also, for now, its most underappreciated one. Industry has not yet encountered enough of the scenarios that v4.0 was built to handle gracefully.

A Practical Note on Code Management

For organizations navigating the operational realities of v4.0 implementation, codelist management deserves specific attention. For internally managed submissions, modern submission systems handle most of the complexity. Where contract research organizations or external partners are involved, the challenge shifts to ensuring that codelist versions are synchronized across organizations, requiring either shared systems or explicit codelist governance protocols.

While not exclusive to v4.0, codelist management matters more than it did with v3.2.2.

From Compliance to Capability

Japan’s mandatory transition to eCTD v4.0 poses a question that mandates alone cannot answer: Now that the format is required, who will design the incentives that make its capabilities worth using?

The answer is not straightforward, because the work belongs to all three constituencies simultaneously:

• Regulators must demonstrate how structured metadata is used in review, making the value of compliance visible to applicants.
• Vendors must reduce the friction of implementation, shifting the cost-benefit equation.
• Applicants must look beyond the submission itself, identifying how v4.0’s richer data architecture creates value within their own organizations.

If v4.0 ends its first mandatory chapter as nothing more than an updated file format for existing workflows, the standard will have fallen short of its purpose.

The transition from mandatory compliance to genuine capability is the next frontier. It will require deliberate effort from regulators, vendors, and industry alike.