For innovative pharmaceutical companies, that question can determine the fate of a blockbuster.

In T 0592/24, the EPO Boards of Appeal provided one of their clearest applications of G 2/21 to date—and delivered encouraging news for the life sciences sector.

Novartis successfully defended its patent covering the ribociclib/letrozole (Kisqali) combination for HR+/HER2- breast cancer against five opponents.

The central issue was whether post-published clinical evidence could be relied upon to support inventive step.

The answer: Yes—provided the later evidence confirms what the application already teaches.

Novartis’ application contained:

• preclinical in vitro synergy data;
• in vivo xenograft data; and
• a clear preference for the ribociclib/letrozole combination.

Years later, clinical analyses demonstrated significantly improved overall survival compared with the prior-art combination using palbociclib.

The opponents argued that this clinical benefit could not be relied upon because overall survival was never mentioned in the application as filed.

The Board disagreed.

It held that the later clinical data did not introduce a new technical effect. Instead, it merely confirmed an effect that was already encompassed by the original technical teaching. The fact that definitive clinical evidence only became available years after filing was recognised as an inherent feature of pharmaceutical development—not a reason to disregard the evidence.

This is one of the most practical G 2/21 decisions for pharmaceutical and biotech patent drafting.

Some practical lessons:

• Build your case early. Strong preclinical data can provide the foundation for relying on later clinical evidence.
• Identify preferred embodiments. A disclosed preference may later support evidence demonstrating superior clinical performance.
• Scientific methodology matters. Comparative experiments should be carefully designed and capable of withstanding scrutiny.
• Post-published evidence should confirm—not create—the inventive concept.
• “Try-and-see” remains a demanding obviousness standard. Knowing a class of compounds is not the same as having a reason to select the successful one.
• Bring your best evidence early. The Board refused to admit several late-filed documents from both sides, reminding parties that appeal proceedings are not a second opportunity to build their case.

T 0592/24 is another important piece in the post-G 2/21 landscape. It demonstrates that well-supported preclinical data can keep the door open for later clinical evidence—provided the application already contains the seeds of the technical effect ultimately relied upon.