Attending a life science conference recently, I was suddenly struck by the dramatic change in perception that has occurred regarding the relevance of biomarkers to drug discovery and development.
It was heartening to see, not least because I was part of the industrialisation of genomics research at its peak in 2000. I witnessed the biotech bubble bursting spectacularly in the form of tumbling share prices and the disintegration and transformation of the bioinformatics giants such as Incyte Genomics and Celera which spat out many erstwhile biomarker researchers onto the streets to look for other forms of employment.
Genomic and Proteomic Biomarkers: Nothing New, But the Prevailing Wind Has Changed
By the year 2000, Incyte had accumulated a vast repository of knowledge and research tools which would have come in very handy today. Gene databases and expression microarrays containing thousands of liver-specific cDNA variants to monitor the expression of potential toxicity indicators; similar microarrays containing diabetes-relevant pathway genes that would be useful today for efficacy markers, clinical trial stratification or early diagnostics; gene polymorphisms in candidate genes which could explain clinical variation to treatment response; lists and lists of novel genes and proteins that were co-expressed in disease or normal tissues that would form the starting point for identifying surrogate markers of clinical endpoints, not to mention disease intervention targets, and so on.
Discovery didn’t seem to be hard and indeed it would be even easier today, with the advent of faster and cheaper technology for genomic sequencing and analysis. The issue was finding a suitable application, proving clinical relevance and gaining industry acceptance.
A decade ago, the Blockbuster Drug Model was very much in prevalence and the pharmaceutical industry hadn’t yet woken up to the fact that their increasing R&D spend wasn’t yielding greater productivity in the shape of increasing drug approval numbers. The industry, for the most part, wasn’t ready to consider pre-selecting patients for clinical trials or treatment in order to target therapies more effectively, partly because it would reduce the potential market for their new drugs but also (and more importantly) because the scientific knowledge (biological processes and molecular pathways; pharmacology; metabolic pathways etc) needed to support biomarker use in the clinic wasn’t yet there .
The Opportunity: Better Prediction of Clinical Outcomes is Vital to Improve Drug Approval Numbers
Whilst it seems that nearly every review paper on the state of the Pharma industry quotes the $802M cost of developing a new prescription drug (1), stemming from the high rates of attrition particularly in Phase II which continues today (2), the answer appears to lie in gaining at an earlier stage, better quality data on which to make decisions.
Clear financial benefits of the adoption of biomarkers have encouraged GSK, Roche, Merck, J&J, Pfizer, AZ and others to adopt biomarker discovery alongside their drug discovery programs (3).
“Biomarkers have certainly impacted on our internal decision making on whether to move forward to the next phase of clinical development” James Weatherall, AstraZeneca
We’re now seeing some prominent examples of substantial cost and time savings from having biomarkers to help better define and diagnose diseases, predict adverse drug events, identify patient groups who would benefit from certain treatments, and give an early indication of efficacy (4). The best example among many remains Roche’s Herceptin, using the Her 2 gene expression marker to indicate the 20% of patients who would benefit (5). The 8-year acceleration of Herceptin’s approval saved and estimated $35M in clinical trial costs and generated $2.5 billion of income, and crucially, gave 120,000 patients access to the drug before they might otherwise have done.
“Every drug discovery project must now have a biomarker” Dr Louise Jones, Cancer Research UK (Genesis Conference 2010)
This all looks good for the newly (or is it re-newly?) emerging biomarker industry, but naturally the future is by no means rosy or risk-free (6).
Challenge Number 1: Choosing The Applications That Will Benefit
The first challenge for any newcomer to the biomarker discovery industry will be to identify an application for a biomarker that will benefit patients, pharma companies and payers.
Oncology has led the way in the adoption of biomarkers and will continue to do so because of the toxic nature of the drugs which at best benefit 30-50% of the patients they are prescribed to. Biomarkers present a way to cut through the trial-and-error nature of drug prescribing in order to get to the treatment that works for the patient and time is of the essence in this disease.
A key challenge will be to identify other diseases, development challenges and treatments for which a predictive biomarker would add significant advantages. Biomarker discovery companies will need to be extremely careful here – the wrong move will have a significant damaging impact on their future success.
Challenge Number 2: Proving New Markers Increase Predictability in Development and Treatment.
Like new drugs, biomarkers must be approved by the regulatory authorities to be used in the clinic, although this may not necessarily be the case for biomarkers that assist early drug discovery. Regulatory approval currently requires a wealth of documentary evidence that the biomarker correlates directly with the primary clinical endpoint in question. Which is why successful biomarkers in use today (cholesterol testing for cardiovascular disease risk; prostate serum antigen levels for prostate cancer; Her2 for Herceptin administration, for example) have been known for a long time.
Some companies (eg GSK) have seen this coming and for a number of years have actively sought clinical trial participant approval to test their samples retrospectively for biomarkers. At the same time, the FDA is developing guidance for the development and implementation of biomarker studies on the critical path to new medical products. However, if you’re a new company entering the biomarker discovery field you’ll need prospective clinical studies and they are currently expensive and difficult to fund using current models.
Challenge Number 3: Gaining Timely Adoption By Industry and Regulatory Authorities
Any of us who have worked with pharma on new technology will be familiar with the unbelievable slowness with which they accept and adopt new approaches over existing ones. Biomarkers are unlikely to be the exception, no matter how much they’re needed. Timely regulatory authority acceptance of new biomarkers will be vital to help this along.
In the worst case scenario we’ll see new biomarkers adopted alongside the old ones – with the net result of an increase in R&D costs and development timelines; completely the opposite of the intention.
Challenge number 4: Get Paid!
Diagnostics have typically been undervalued in comparison to medical products. If a biomarker is needed in order to gain approval for a new medical product then its value will need to reflect this – not least because the biomarker industry will need to survive. Whilst the risk of developing a biomarker to full clinical implementation will not be as high as developing a new treatment, nevertheless the costs of doing so will be significant and so the rewards must reflect this in order to attract investment for new companies entering the industry.
The Future Could Be Brighter
If we do find the right biomarkers, for the right application, prove their clinical relevance, and get funded then it could be great for patients, this industry and society generally. I really hope we get it right this time.
“The developer will have enabled clinical differentiation of its product; the regulator will have better science from which to base decisions, making the process clearer and faster; and patients and physicians will have better information to use when considering individual risk and benefit. Payers will have information to support providing appropriate patient access to the drug.” Austin and Babiss, Roche
For the next post on the issue of biomarkers, click here.
(1) Tufts Center for the Study of Drug Development October 14, 2005.
(2) Drug Discovery Today, Volume 11, No. 17/18 Sept 2006, “Translational research in the pharmaceutical industry: from bench to bedside”, D. O’Connell and D. Roblin
(4) Thomson Reuters (2010) White Paper: Biomarkers: an indispensible addition to the drug development toolkit
(5) Thomson Reuters (2010) White Paper: Establishing the standards in biomarker research
(6) M. J. Finley Austin and Lee Babiss (2006). Commentary: Where and How Could Biomarkers Be Used In 2016? The AAPS Journal 2006; 8 (1) Article 22