Case Note
How TIGIT Combinations Looked Stronger Than Their Escalation Boundary.
A pre-trial editorial note on why supportive TIGIT combination signals looked more settled than the class-level contradiction structure actually allowed.
The issue was not whether the signal existed. The issue was whether it had earned the right to scale.
The Support Trap
TIGIT combinations looked strong because the field had enough local support to tell a compelling story. The mistake began when that support was treated as if it had already survived contradiction. It had not. The class signal was still conditional, and the escalation language kept outrunning that fact.
Why the Story Became Persuasive
The combination logic sounded coherent, the checkpoint framework felt familiar, and selective encouraging updates gave the impression of convergence. That is how a class thesis begins to feel stable before it has actually been forced through a rigorous contradiction audit.
Where the Boundary Sat
The real boundary sat at class portability. Supportive local signals did not yet prove that TIGIT combinations would remain stable across contexts, populations, and late-stage design pressure. The issue was not the absence of mechanism. The issue was unresolved breadth.
Support is not stability, and class excitement is not the same thing as a resolved escalation boundary.
Why This Matters Now
This pattern repeats whenever oncology treats encouraging support as if it had already become robust, portable confidence. That is exactly why evidence lock matters. Once hindsight is removed, the unresolved part of the story becomes much easier to see.
What this changes for a live program
The value of the read is not only diagnostic. It changes the next decision. If the confidence claim depends on a boundary that still looks unstable, escalation should pause, narrow, or be re-tested before the claim expands further.
Public evidence can show where the instability sits. High-confidence program-specific resolution still requires internal data, but the decision logic already becomes clearer: do not spend broad confidence where only local support exists.
References
Representative pre-trial references
- Roche reports interim results for phase III SKYSCRAPER-01 study in PD-L1-high metastatic NSCLCMajor contradiction event for the TIGIT class thesis.
- Roche provides update on Phase III SKYSCRAPER-01 study in PD-L1-high metastatic NSCLCFurther class-relevant update showing unresolved breadth and durability.
- Gilead and Arcus announce domvanalimab plus zimberelimab and chemotherapy exceeded one year of median PFS in first-line upper GI cancersRepresentative supportive signal showing why the TIGIT combination story became persuasive.