Personalized Medicine: Next Big Hope or Next Big Hype?

March 2, 2005
Oncology drugs are driving the marketplace closer to the vision of personalized medicine, but major infrastructure change is still needed.
An age of personalized medicineis dawning. Photo from the coverof "Medicines for You: Studying How Your Genes Can Make a Difference," a consumer educationbrochure from NIH/NIGMS.By Angelo De Palma, Ph.D., Contributing EditorRecent breakthroughs in genomics have made it possible to predict, more accurately than ever before, how each patient will respond to any given drug. As a result, we hear that an age of personalized medicine is dawning, in which therapies will be tailored to groups of patients, or even individuals.This brave new world would mean the end of the blockbuster drug model and would transform the way that medicine is practiced, and drugs are made and marketed. Armed with new testing methodologies, manufacturers would nimbly change product and production schedules in response to changing patient needs.Are we there yet? Definitely not. Blockbusters coexist nicely with niche products today, and probably will for the next 10 to 20 years, according to Adam Bianchi, COO at Cutting Edge Information (Durham, N.C.) and companies will continue to research and develop mid-sized products and blockbusters for the near and mid-term. “Today’s pipeline drugs are the medicines that companies will be launching and marketing for the next 10 to 20 years,” he says. Others see the blockbuster model as outdated, since even the most successful drugs don’t work in all the patients all the time. “Ninety percent of all drugs are only 50% effective,” says John Funkhouser, president of PharmaNetics, (Morrisville, N.C.), “and the top 25 drugs are anywhere from 40 to 70% effective.”Determining the most responsive patient groups will change the structure of modern healthcare and bring sweeping changes in the way society views medicine. However, the blockbuster model will not die easily, Funkhouser believes. “Pharma is facing wide-ranging issues from consolidation, weak pipelines, longer regulatory review, fewer years of commercial intellectual property protection, generics, price pressures and political heat,” he says.Oncology is one therapeutic area that’s moving closer to the vision of personalized medicine. A small but growing number of companies are applying a technique called “theranostics,” identifying potential responders a priori, and bundling appropriate drugs with diagnostics.Consider Genentech’s Herceptin, for example. This breast cancer treatment originally provided only modest benefits, and some troubling side effects, for a broad patient population. However, once patients were singled out and treated who expressed the HER2/neu gene, linked to growth of more aggressive tumors, the drug’s efficacy shot up, justifying the adverse events. Bayer Diagnostics, which developed the HER2/neu diagnostic, predicts that the theranostics markets will grow by 20% per year through 2009.As Herceptin shows, personalized medicine will create opportunities for some molecules that fail to meet traditional safety and efficacy criteria. Selecting a subgroup based on genotype can help sponsors boost efficacy into the statistically meaningful range, albeit for a smaller — sometimes much smaller — population than originally anticipated.Individual differences justifying personalized therapies may be dramatic or subtle. In extreme cases, patients may express a gene (or protein) or not, or the gene may exist but be inactivated.Usually, expression differences are a matter of degree. Genes make proteins, which make up the complex cascades of disease and metabolic pathways. Improperly expressed proteins can cause a drug to be completely ineffective but innocuous in one patient, ineffective but toxic in another, or safe and effective in yet another. Metabolic enzymes (and the genes coding for them), which determine a drug’s toxicity and side effects, could work so well that a drug is chopped up and excreted before it gets to work.So personalized medicine, if it has any chance of success, relies on treating genotypes rather than phenotypes, and having the means to test for the underlying characteristic.The manufacturing responseIn response to the promises and threats of personalized therapies, pharmaceutical companies must accelerate retooling and streamlining of production lines, just as heavy and light manufacturers in other industries have done over the last twenty years. Personalized medicine means smaller batches of a wider variety of drugs. “Eventually, production managers will have to think in terms of 100 different batches of one million pills rather than one batch of 100 million pills,” notes Bianchi. Isolating process suites will become a growth industry as manufacturers will abhor overcapacity even more than they do today.“Leaning out” operations will also be critical, as pharmaceuticals shift from the blockbuster model to personalized gene therapies and gene-specific small-molecule treatments, says Richard Chua, executive VP at the Juran Institute (Southbury, Conn.). “Companies competing in this arena will need to be lean in every sense of the word,” he notes, “including the ability to switch smoothly from one product to another.” Leaning out such processes will offer the usual benefits of streamlined, lower-variability processes, benefits which will multiply with the plant’s complexity.At the same time pharmaceutical manufacturers will need to sharpen quality control, an activity Bianchi calls a “battleground” in the ongoing healthcare cost-control war.It is unclear whether genomics and personalized medicine will be a centralizing or decentralizing force in pharmaceutical manufacturing. Companies will want the cost savings of mega-facilities, but also the flexibility that smaller facilities bring.The reality for production managers is that they will have to accommodate both cost-consciousness and rapid change. While large drug firms will probably always rely on blockbusters to some extent, the eventual emergence of medicines tailored for niche groups will require manufacturing facilities that are adept at handling smaller fill orders and that can switch rapidly between brands, says Dr. Frederick Cohen, president and founder of Crownstone (Bensalem, Pa.), a healthcare consultancy group.Manufacturing flexibility can take many forms. For example, an antibody-toxin conjugate, where the toxin is switched depending on the tumor type, can be manufactured in a dedicated bioprocess facility (for the antibody) and in multiple, smaller suites (toxin). Conjugation could occur in a small, separate facility.Similarly, modern bioprocess facility planning involves the multiple-product approach. This is in stark contrast to biotech’s beginnings, when yields were so low that plants were constructed around single products. “Today, processes are designed to fit inside facilities,” observes Dr. Michael Kamarck, senior VP at Wyeth BioPharma (Andover, Mass.). This approach is already enabling a larger number of products to be made within the walls of a single facility.Nevertheless, the current manufacturing infrastructure was built upon, and in turn supports, the blockbuster paradigm. Personalized medicine will probably never take hold without changes to the pharmaceutical cost structure, except for very rare situations where a treatment cures a previously fatal illness. “We need a better way to ensure that, as treatments get better, we have the flexibility to make these complex products,” comments Dr. Ajaz Hussain, deputy director in FDA’s Office of Pharmaceutical Science. “The only way to resolve our current challenges is through more innovation, new thinking.”John Funkhouser puts it less diplomatically: “Personalized medicine will happen because the present model is bankrupting the Western world.”FDA: ready and willingMapping the genome and uncovering associations between genes and diseases are relatively straightforward exercises given today’s genomic tools. Developing an in vitro diagnostic based on the gene is somewhat more difficult. But as genomics companies are learning, the hardest thing of all is to produce a therapy based on genetic information.Linking drugs and diagnostics raises regulatory issues that are larger than one would expect by simple addition. FDA treats these products as combinations, and regulates them somewhat differently than single-component products.The immediate conundrum: Should the combination product be regulated and licensed as a drug, a device (as are most diagnostics), or both? This conundrum proved significant for Herceptin, which experienced a very long approval time, in part because FDA wasn’t experienced with that type of product.FDA made significant progress towards streamlining approval through its de facto combination of CBER (Center for Biologics Evaluation and Research) and CDER (Center for Drug Evaluation and Research). As a result, most drugs and biologics are now reviewed by CDER. Diagnostics, however, fall under the auspices of CDRH (Center for Devices and Radiologic Health), and are foreign territory for both drug companies and drug regulators.Realizing the need, FDA created the Office of Combination Products. Today, the Office serves as the gatekeeper for therapy-diagnostic combinations, guiding combination products to the proper review Centers.For straightforward combinations, like drug-eluting stents, products are usually assigned to either CDER or CDRH, depending on their principal mode of action (drug or device, respectively). For therapy-diagnostic combinations, which contain two or more products with the capability of standalone status, the Office of Combination Products assigns a primary Center and facilitates cooperation between appropriate Centers.Approval of diagnostic-therapeutic combinations will likely occur in stages, according to Marc Scheineson, a former FDA commissioner for legislative affairs and now head of the food and drug practice at the Washington, D.C. law firm Alston & Bird. “Diagnostics will be approved first as devices, then be used to screen patients for clinical trials,” he notes. “After the drug shows efficacy the drug would then be approved.”Personalized medicine presents clinical development issues for sponsors as well as regulators. First there is the non-trivial matter of designing a diagnostic and therapeutic that go hand in hand, and timing their market release so that one is not orphaned by the other’s delay. “Targeting something that’s expressed in a subset of patients requires that you be able to test for that marker,” says Dr. Janet Woodcock, FDA’s deputy commissioner for operations. If tested on the general population, a good percentage of which are known to be non-responders, the number of subjects required to achieve statistical significance is huge. “If you can identify responders prospectively you can spare those the drug won’t help, and may not even need a control group for certain diseases,” notes Woodcock.For its part, FDA says it is up to the task of approving an avalanche, should it occur, of personalized medicine products, since its user fees would rise with increasing numbers of applications. “But I’m afraid applications probably won’t come in a rush,” Woodcock observes.The question arises of how FDA would regulate a personalized therapy that involved removing tissue from a patient, treating it outside the body with either a drug or device, and re-implanting it in the same individual. In those cases, FDA would regulate platform-based personalized treatments as it does blood banking or gene therapy today, by licensing the process and processing rather than taking a narrow view of tissue differences.“We wouldn’t treat cells as different products because they came from different individuals,” says Woodcock. “But of course if something was added to the cells before returning them to the patient we would regulate that added component as we would any other drug.”Demand for novel molecular diagnostics, for both new and existing drugs, will naturally follow the trend towards more targeted therapies. Today’s wealth of “me-too” drugs may turn out to be a blessing, as protein- and gene-profiling studies identify not just high patient responders but rediscover long-discarded drugs as ideally suited to patient subgroups.Segmenting markets: death knell for blockbusters?Personalized medicine raises economic issues related to market segmentation. Given the industry’s reliance on blockbusters and claims of billion dollar-plus development costs, justifying development of a personalized treatment would seem impossible. Under the current development paradigm it seems that any approach that fractionates a market is more likely to generate orphan drugs than blockbusters. With product markets smaller and more segmented, manufacturers must charge more for their products to recoup development and manufacturing costs. Higher prices are justified, at least in theory, because non-responders or those likely to experience serious side effects have been weeded out and those who remain will experience even greater benefit.In any event, the marketplace for niche products—orphan drugs are an extreme case—while not exactly thriving, is at least active. According to Scheineson, smaller pharmaceutical and biotech companies, funded by established firms, are leading the charge in personalized medicine. “Smaller companies don’t need billion-dollar drugs to survive,” Scheineson observes.One could argue that, at some point, the trajectories of personalized medicines and orphan pharmaceuticals will intersect. For example, Genzyme’s protein replacement for Gaucher’s disease enjoys just 5,000 potential customers worldwide.Wyeth manufactures two drugs for the two major forms of hemophilia: a factor 8 product serving 30,000 patients worldwide, and a factor 9 replacement with only 5,000 patients. If hemophilia itself were not so rare, these products would be considered personalized medicines rather than orphan drugs. Another Wyeth product, Enbrel, is perhaps more illustrative in that it is appropriate for treating perhaps 100,000 out of the one million Americans with severe rheumatoid arthritis.If these examples mean anything, it’s that market fractionation is not always a bad thing and may not matter ultimately. Besides, markets for chronic diseases are huge, even as treatment space crowds with me-too drugs and incremental improvements in formulation, delivery, or dosage form. In an ideal scenario the relative market sizes of individual treatments — say, for congestive heart failure — may remain the same even when specific drugs are prescribed only to “appropriate” patients, e.g. strong responders. Today drugs compete for all patients within a disease category. In the future, this may not be desirable or even necessary if manufacturers can make a strong case for superior efficacy in patient subgroups (assuming, of course, that superior efficacy is real).According to Wyeth’s Kamarck, biotech may be thought of as personalized medicine’s first wave, since therapeutic proteins treat specific, well-defined diseases in limited patient populations. His company’s hemophilia treatments certainly qualify by that definition.Oncology will provide a fertile development arena for personalized medicine products, Kamarck believes. However, to justify the expense of these drugs, they will need to be significantly more effective than existing treatments. For example, Iressa, AstraZeneca’s lung cancer treatment, costs about $60 per pill.Manufacturers will charge appropriately for personalized medicines, based on the market size and superiority of the treatment relative to the standard of care. “It helps if you can create a ‘perfect medicine,’” says Kamarck, “Even then [personalized medicine] won’t compete with the Lipitors of the world, but it will be a good business.”Spurred by scientific advances and the desire for cost containment, personalized medicine will continue picking up steam. Pharmacogenomic studies will continue to unveil responder populations, perhaps reviving a lot of older medicines in the process. “There is good stuff in the bone yard,” Funkhouser notes.Will production facilities be ready for personalized medicine? FDA’s Hussain worries that the current manufacturing paradigm may not be well suited for the emerging personalized treatment model or for small-volume and orphan drugs. “Our development environment and manufacturing paradigms need to be so flexible and cost-effective that companies will not hesitate to develop such drugs. We have to move in that direction.”If companies tackle problems of technical innovation and efficiency today, proactively, Hussain believes the paradigm will evolve, step by step, to where it should be.