How Does Biomarker Stratification Affect Drug Development Cost? It Depends. #BIO2010

On the San Diego Biotechnology Network LinkedIn group, we asked you what you wanted to hear about from the BIO Convention in Chicago, and you requested to hear about Biomarkers. Tuesday I attended a breakout session titled ‘Impact of Biomarkers on Drug Development Complexity and Cost,’ and it described a study done by panel members Federico Goodsaid of the FDA, Michael Palmer of Adaptive Pharmacogenomics, Mark Trusheim of MIT, Steven Averbuch of BMS, Theresa Long of the Van Andel Institute. The study modeled different scenarios utilizing biomarker information and the financial impact on estimated net present value (eNPV) of developed drugs. The group was truly interdisciplinary, and it was clear from the quality of the presentations that they likely worked well together due to their communication skills.
The session described a workshop held in October of 2009 in which case studies on oncology and alzheimers were discussed (featured in Nature Reviews Drug Discovery as well–requires login).
MIT economist Mark Trusheim began by describing the ‘pharmaeconomics’ of choosing patient populations for clinical trials. There are obvious benefits to using biomarkers to enrich populations, but there are many factors which need to be considered, and ultimately it depends on the science behind the drug and the market. The group determined three factors to be the most important: variability in the therapeutic effectiveness of the drug, prevalence of the biomarker, and the quality of companion diagnostics. Trusheim also described modeling different long term scenarios ranging from Phase II extending through to the end of market exclusivity, from the perfect ‘Nirvana’ situation, where everything goes perfectly, to ‘pharmageddon’ where it goes bad at every turn. Trusheim indicated that in both the oncology and Alzheimer’s study, a very significant increase in eNPV could be achieved by utilizing biomarker information.
Steven Averbuch, VP of Oncology at BMS, covered the study results for Herceptin and Vectibix for oncology indications. In the case of Herceptin, it had a large effect on a small population, saving money and adding to the eNPV by allowing a smaller clinical trial, but perhaps precluding the discovery of other biomarkers or indications for the drug, as was found to be important for drugs such as Gleevec. Vectibix had a large effect on a large population, obviating the need for biomarker stratification but giving the drug a higher eNPV because of a larger market size. Averbuch reiterated Trusheim’s three important factors, and the need for increased communication between all the stakeholders to utilize biomarker information to help move from the ‘pharmageddon’ to nirvana drug development situation.
Theresa Long, presented the study on the Alzheimer’s drug Bapineuzumab, and the effect of biomarker stratification on a chronic condition. They used the ApoE4 biomarker and started with data past Phase 2. Three different scenarios were modeled, from an all-inclusive to stratified, with an 80% increase in eNPV in the biomarker study. Long stressed the importance of knowing the prevalence of the biomarker in the population, even for chronic conditions such as Alzheimer’s which has blockbuster potential.
This session was inspiring as it showed how science can drive drug development and how biomarker stratification could lead pharma and healthcare towards a path of increased communication resulting in lower costs. The study has been submitted for publication and the modeling tools that were used will be available. Federico Goodsaid indicated that the tool could be made available to those who contact him.