Cancer biology informs ORR
Gleevec, Herceptin and Tagrisso succeed because they inhibit solo drivers. Solo drivers don't have many co-occurring bypass mutations, but more importantly, the bypass mutations do not cluster together.
Investment lessons · HER2
HER2, the soloist
Trastuzumab has generated $100B+ since 1998, although HER2 is neither the most frequently activated oncogene nor the most highly expressed target. Patient data explains it: HER2 is a solo driver in breast cancer, while in stomach cancer bypass alterations are more frequent and cluster together. Ask which mutations cluster with your target before the trial.
The best target I have ever seen · by Pawel Zawadzki, PhD
$100B+
Generated by trastuzumab since approval in 1998
The case
An unexpected success story
HER2 is not the most frequently activated oncogene, nor is it the most highly expressed drug target in cancer.
So what went right?
It is the same factor that made Gleevec so wildly successful. HER2 is a solo driver, particularly so in breast cancer.
Finding 1
What the patient data revealed
Patient data uncovered that HER2 is a good drug target.
- HER2 is frequently activated, particularly in breast, stomach, bladder and esophageal tumors.
- Importantly, beyond activating Gain-of-Function mutations, high-level amplifications (>6 additional gene copies) are frequent. These parameters characterize good drug targets.
HER2 activations across tissues
Finding 2
What the patient data revealed
More importantly, HER2 is a solo driver in breast cancer, but much less so in stomach cancer.
- Clinical response rates to trastuzumab are very high, but why are they stronger in breast cancer than in other tissues?
- Bypass alterations (mutations that can replace HER2 signaling) were less frequent in breast cancer than in tissues with lower response rates.
- Critically important, the bypass alterations clustered together in tissues with lower response rates but not in breast cancer.
Example of alterations that can bypass HER2 in primary tumors
| Gene | Frequency | Clustering |
|---|---|---|
| Breast, HER2-positive | ||
| EGFR | 5% | none |
| PTEN | 20% | |
| Stomach, HER2-positive | ||
| EGFR | 15% | strong, found together 4× more likely than expected (p = 0.01) |
| PTEN | 27% | |
The take-home message
Solo-drivers are rare in solid tumors
Patient data opportunity
Ask a critical question: In my target patient population, what other mutations cluster together? This can inform ORR before clinical trials and, more importantly, guide drug combinations that are key to high ORR and durability.
Recommendations
Consult patient data for reliable outcomes
01
A soloist—or part of a band
Use patient data to determine whether your target is an independent driver or is supported by another player.
02
Primary tumors predict the ORR and durability of clinical trials
Measure the clustering of mutations in treatment naive tumors.
03
Patient data reveals which drug combinations to pursue
Clustered mutations indicate specific combination partners that will dramatically improve the ORR and durability of your asset.
Medicines intended to work in patients should be developed with patient data in mind.
Investment lessons · FGFR2Solo driver vs a “team player”Amgen paid $1.9B+ for an FGFR2 asset; later trials did not repeat the early signal.Read online + PDFOpen →Investment lessons · PIK3CAWhen $2 billion buys you a seat – not certaintyNovartis paid $2B+ for a mutant-selective PI3Kα inhibitor.Read online + PDFOpen →