Surviving the Phase II Cliff: Why So Many Promising Drugs Fail — and What We Can Do About It
- evadanielson55
- 6 hours ago
- 2 min read
If you work in translational medicine, you already know the story.
A promising antibody candidate advances through discovery. Preclinical data looks strong. Phase I demonstrates acceptable safety. The team moves forward with cautious optimism.
And then Phase II happens.
Across the industry, only about 34% of drugs successfully progress from Phase II to Phase III. For many organizations, this moment has become what some quietly call the "Phase II Cliff"—the point where years of work and hundreds of millions of dollars can vanish.
For leaders responsible for translational strategy, these failures rarely feel mysterious. They feel frustratingly predictable.
The Translational Gap
The root cause is often not poor science or poor execution. It’s a failure of translation.
Modern drug development still relies heavily on models that simplify biology: animal models, overexpressed cell systems, and controlled in vitro assays. These tools are valuable, but they often fail to capture the complex reality of human disease.
The result is a familiar pattern:
The target exists, but it may not be causally relevant in human pathology.
The drug reaches the tissue, but never truly engages its target.
The mechanism works in mice, but not in patients.
This disconnect—the Translational Gap—is responsible for a massive proportion of clinical attrition.
Moving Beyond Target Presence
At Offspring Biosciences, we believe the industry must move beyond asking whether a target exists and instead ask:
Does this target actually drive disease in the patient's specific tissue architecture?
This shift requires validating therapeutics directly in human disease tissue, where the true complexity of pathology, the microenvironment, and cellular interactions can be observed in spatial context.
The Tissue Insights™ Approach
Our Tissue Insights™ platform applies the core principles of the AstraZeneca 5R framework—Right Target, Right Drug, Right Mechanism, Right Safety, and Right Patient—directly within human disease tissue.
Using advanced molecular pathology approaches, including:
High-order multiplex immunofluorescence (mIF)
In situ proximity ligation assays (isPLA) for sub-40nm target engagement
AI-assisted quantitative digital pathology
...we generate spatially resolved statistical datasets that help teams answer critical translational questions much earlier.
Questions like:
Is the target present in the right cells within the diseased tissue?
Does the therapeutic physically engage the target in situ and drive a downstream response?
Could off-target binding patterns signal a toxicity risk before expensive GLP studies?
Answering these questions early can dramatically reduce uncertainty before entering expensive clinical phases.
From Pixels to P-Values
For translational leaders, the goal isn’t just generating beautiful images. It’s generating 'Deep Insights'—transforming subjective pathology into objective statistical data ("Pixels to P-Values"). It is the kind of decision-grade data that allows you to confidently answer whether a program should advance, pivot, or stop.
When we shift from merely proving target presence to validating biological mechanisms in actual human tissue, we give drug programs their best chance to succeed in the clinic.
📄 Want the full framework? Find the full white paper: “Surviving the Phase II Cliff: The Tissue Insights™ Framework for De-Risking Antibody Development” and more on our Platforms Page.
