BlackCat Bio — Case Study · Phase 3 / Rare Disease / Pulmonary
A novel Pulmonary trial with no methodological precedent, cascading protocol revisions, and a last-minute interim analysis — finished within hours of protocol lock.
Background
A sponsor approached BlackCat Bio to develop the Statistical Analysis Plan for a Phase 3 pulmonary trial — simultaneously with the protocol, on a compressed timeline, with exacting FDA expectations. The indication was rare, the endpoint structure was novel, and the agency had already signaled specific statistical requirements with no established methodological roadmap to follow. Protocol and SAP had to be developed in lockstep, written concurrently from a moving target.
The challenges
No precedent, high FDA scrutiny
FDA had specific statistical expectations but no analogous approved programs to cite. Every methodological decision required justification from first principles or adjacent evidence.
Secondary objectives in flux every 24 hours
Protocol decisions on secondary endpoints changed on a near-daily basis. Each change threatened SAP consistency — a misalignment caught only at final review could have forced a complete SAP rewrite.
Interim analysis added at the final hour
An interim analysis with O'Brien-Fleming boundaries and associated operational, regulatory, and multiplicity implications was introduced late — requiring rapid design, documentation, and integration into an already-complex SAP architecture.
Protocol and SAP written simultaneously
There was no completed protocol to write from. The SAP had to track a living document while building out the statistical framework independently — requiring continuous cross-referencing to ensure alignment throughout.
How BlackCat Bio's AI agents made the difference
Precedent discovery across indication-adjacent areas
BlackCat Bio's therapeutic-area AI agents scanned FDA guidance documents, Complete Response Letters, and approved SAPs from analogous indications — surfacing methodological justifications that human literature review would have taken days to assemble, if it found them at all.
Real-time protocol-to-SAP consistency tracking
As protocol sections changed, BlackCat Bio's workflow flagged downstream SAP sections requiring update — keeping statistical language, estimands, and analysis populations synchronized with the evolving protocol without manual reconciliation.
Rapid interim analysis integration
When the interim analysis requirement emerged late, AI-assisted synthesis of precedent designs and boundary methodologies produced a defensible, fully integrated interim analysis framework — without extending the delivery timeline.
Regulatory language calibrated to FDA expectations
AI agents trained on regulatory biostatistics drafted and refined statistical language to meet FDA standards for the specific indication area — ensuring precision and internal consistency from the first draft rather than through iterative manual review cycles.
"There was no playbook for this study. Our AI agents found the adjacent evidence, built the justification, and kept the SAP synchronized with a protocol that was changing underneath us — every single day."
BlackCat Bio statistical team
Outcomes
Final SAP delivered within hours of protocol lock
No lag between protocol finalization and SAP delivery — ready for submission without delay.
Zero consistency gaps at final review
Protocol and SAP were fully aligned at lock despite ~24-hour revision cycles throughout development.
Methodologically defensible without direct precedent
Every statistical decision was grounded in relevant guidance and adjacent approvals — documented and traceable.
Interim analysis integrated at the last minute
O'Brien-Fleming design, operational procedures, and multiplicity adjustments incorporated on a compressed timeline without rework.
Why it matters
Phase 3 SAP development is rarely linear. Protocol changes, late FDA feedback, and novel endpoints create conditions where traditional consulting workflows — built around sequential drafting and manual cross-referencing — break down. BlackCat Bio's AI-augmented approach is designed for exactly this environment: high uncertainty, tight timelines, and unforgiving regulatory standards. This engagement demonstrates what becomes possible when therapeutic-area intelligence and real-time consistency checking are built into the statistical workflow from day one.
Working against the clock on a complex SAP?
Bring us your protocol — novel endpoints, moving targets, and all. We'll show you what's possible.