This is one of the most famous—and sobering—statistics in oncology drug development.
Based on decades of translational research and meta-analyses, the widely cited figure is that approximately 90% to 95% of oncology drugs that show efficacy in mouse tumor models fail to be effective in humans.
This means the success rate (the percentage that do work) is typically cited as being between 5% and 10%.
Here is the breakdown of why this happens and the specific data behind it:
1. The “5% to 10%” Rule of Thumb
The most frequently cited study on this topic comes from a 2001 analysis by Iain W. Johnson and colleagues. They looked at the historical correlation between mouse tumor models (primarily xenografts where human tumors are implanted into immunodeficient mice) and Phase II clinical trial results in humans.
- True positive rate: They found that if a drug worked in mice, the probability it would work in humans was only about 5%.
- The catch (False positives): Mice are actually very good at predicting toxicity, but very bad at predicting efficacy.
More recent data from the National Institutes of Health (NIH) and Tufts Center for the Study of Drug Development paint a slightly more nuanced picture:
- The overall clinical approval rate for oncology drugs from Phase I to approval is about 5.3% to 6%.
- However, of the drugs that fail specifically because of lack of efficacy in Phase II or Phase III trials, a massive portion had previously shown strong anti-tumor responses in mouse models.
2. Why is “Ovation” (Promising Mouse Data) So Misleading? (The Scientific Reasons)
For a startup like XYZ Company, its KD061 program showing “>70% tumor growth inhibition in mice” is a necessary first step, but it is far from a guarantee of human efficacy. The 90-95% failure rate is due to three major biological disparities:
- The “Mickey Mouse” Tumor Problem: Standard xenograft mice have no functional immune system (they are “nude” or “SCID” mice). This means the mouse immune system doesn’t react to the human tumor, which completely ignores the role of the tumor microenvironment and immune evasion—both of which are key to solid tumor (like renal cell carcinoma) resistance in humans.
- Pharmacokinetics vs. Pharmacodynamics: A mouse liver metabolizes drugs completely differently than a human liver. A drug might achieve a high enough concentration in the mouse bloodstream to kill the tumor, but the human liver might rapidly metabolize the drug into an inactive form before it ever reaches the tumor.
- Tumor Heterogeneity: The cancer cells implanted into a mouse are typically a homogenous cell line grown in a lab for decades. Human tumors are highly heterogeneous—meaning different cells within the same tumor have different mutations. Drugs that hit one specific pathway in a mouse tumor often miss the redundant escape pathways that human tumors activate.
3. The “New Hope”
While the historical rate is 5-10%, that number is climbing for drugs with validated biomarkers. The FDA’s recent oncology approvals increasingly use a “biomarker-driven” approach.
For XYZ Company specifically, its target (HIF-2alpha and ferroptosis) is highly relevant. Notably, the FDA-approved drug belzutifan (Welireg), which targets HIF-2alpha, did translate well from mice to humans for renal cell carcinoma. If XYZ Company’s KD061 indeed induces ferroptosis via a novel mechanism that bypasses standard GPX4 resistance, it may fall into a higher success bracket than the historical 5%—but it is still operating within a statistical reality where 90% of drugs fail in humans despite working in mice.
Bottom Line for Investors/Stakeholders
When you evaluate a preclinical company like XYZ Company, “efficacy in mice” is a ticket to the game, not a win. The real “Ovation” moments for XYZ Company will occur when it hits the 2026/2027 milestones:
- IND filing (Q4 2026): Proving the drug is manufacturable and safe in toxicology studies.
- Phase 1 human dosing (2027): Proving it is tolerable in humans.
- Phase 2 efficacy: Showing actual tumor shrinkage in patients.
Until then, the historical data suggests a ~90-95% probability of failure that they must overcome with smart trial design and biomarker selection.
Posted in News
