The Changing Dynamics of Cancer Drug Development

The growing number and size of oncology players has made the field far more competitive, eroding--or at least, shifting--some of the advantages that continue to attract smaller firms to the field.

Oncology is probably the most popular therapeutic area for drug development. About a third of all drug candidates in development are for oncology indications. There are over 150 candidates in Phase III trials, including both new compounds and marketed products being developed for new indications.

The rationale is simple. The medical need is high. Compared to primary care drugs, oncology medicines look cheap to develop and market; generous pricing yields large markets—and billion dollar products.

The growing number and size of oncology players, however, has made the field far more competitive, eroding--or at least, shifting--some of its advantages, particularly for smaller firms. As more products reach the market, accelerated approval and orphan drug status are becoming rarer. Securing high prices is likely to become harder, too, as government cost burdens rise, and as existing treatments improve. New drugs must increasingly be positioned relative to marketed products, increasing trial size and complexity. Getting access to patients is more competitive.

Larger firms with an established oncology franchise are best placed to survive the changing dynamics. They're fast expanding their compounds' use across as many subsegments as possible, and in as many different cancer types as possible—the only way to create blockbusters in such a fragmented field. So for example, a product might rapidly progress from an indication for previously treated patients to first-line therapy. And its use might broaden out from colorectal cancer to breast, lung and subsequently other tumors, as is the case for Genentech Inc. 's bevacizumab (Avastin).

Such moves accelerate changes in standard of care, raising the bar for new agents to demonstrate efficacy, and hence creating entry barriers for smaller newcomers. Colorectal cancer is a good example of a disease whose treatment has been transformed in the space of just two or three years. Response rates to drugs have risen from about 15% to 50% with new combinations, making it far harder, and more expensive, for smaller entrants to demonstrate their candidate's benefits over existing treatments. Picking the reference product in clinical trials and the optimal trial design also becomes trickier as more drugs reach the market.

An analysis of Phase III cancer trials suggests that large pharma and Genentech are developing their own candidates for most of the common, and thus the most lucrative, cancers: breast, lung, prostate, and colorectal. (See Exhibit 1.) The recent registration of drugs in such areas (for example, docetaxel (Taxotere) in prostate cancer, oxaliplatin (Eloxatin) as an adjuvant treatment in colorectal cancer or Avastin in colorectal, and shortly in advanced lung and breast cancers) and the number of ongoing Phase III trials has made the development of drugs for those diseases difficult for small players.

In contrast, biotechs and smaller companies do not appear to be selecting their programs based on disease incidence. (See Exhibit 2.) Such companies' choice of indication is probably driven in part by science, and in part by their perception of the likelihood of gaining orphan drug status and being able to market the drug alone.

The trouble is, the new oncology market dynamics call into question the viability of these assumptions. Even markets for smaller indications may become more challenging as the number of successful products rises. Rituximab (Rituxan) from Biogen Idec Inc. and Genentech, for instance, has significantly raised the bar for new products in non-Hodgkin's lymphoma. Such a trend will force development towards tricky indications, increasing the Phase III failure rate, already high in cancer because of the poor predictive value of Phase II trials. Even candidates that do make it to market may be difficult to position; they may be part of a specific combination cocktail, or relegated to third- or fourth-line therapy.

That said, the success of some cancer drugs is carving out new niche opportunities, too. Gleevec-resistant chronic myeloid leukemia (CML) has become a new target—for now, though, being addressed by Novartis AG and Bristol-Myers Squibb Co. Still, the development of resistance to new treatments—a hallmark of cancer cells—will in theory allow smaller companies, too, to position compounds downstream. Therapies that increase progression-free survival without curing the disease will not shut out the successive use of different drugs. These drugs will be increasingly targeted, however, selected on the basis of pharmacogenomic and drug-resistance tests. Novartis also recently filed an MAA for Gleevec in Philadelphia chromosome-positive acute lymphoblastic leukemia patients and is studying the drug in other genetically defined patient subpopulations.

Although large companies' in-house programs are mostly aimed at the larger subsegments of cancer, their choices of in-licensed programs aren't. The success of a handful of high-profile drugs targeting apparently smaller markets—Rituxan in lymphoma, Gleevec in leukemia, or bortezomib (Velcade) in myeloma—have shown that high price, and the transformation of rapidly lethal cancers into more chronic disorders, allowing for longer treatment, will compensate for the size of the incident targeted population. (Gleevec's activity in CML, for instance, has significantly increased the diagnosis of this rare leukemia.) This in turn has emboldened many larger firms to take on projects of their own, but mostly from outside.

Hematological malignancies have proved an important subsegment in oncology. As well as illustrating how blockbusters can arise from niche markets, it also is where new concepts have appeared both in chemotherapy (where cures now exist for some acute leukemias and lymphomas) and in targeted therapies.

And it's targeted therapies--like Gleevec or Genentech's breast cancer drug trastuzumab (Herceptin)—in turn, which have driven much of the increase in oncology development programs, and now represent the largest segment of drugs in development. (See Exhibit 3.) Most marketed cancer drugs, in contrast, are cytotoxics, given at their maximal tolerated doses, and often responsible for toxic side-effects such as decrease in white blood cell counts and nausea.

Targeted drugs are also one of pharma's most popular in-licensing choices. Although some companies still have a significant number of cytotoxics in Phase III (mainly to broaden the scope of indications of already marketed products), they are rapidly embracing targeted therapies, taking advantage of their small-molecule chemistry and screening capabilities. (See "Early Rewards for Careful Targeting," START-UP, January 2006 (Also see "Early Rewards for Careful Targeting in Oncology" - Scrip, 1 Jan, 2006.).)

Pharma firms appear willing to in-license monoclonal antibodies and targeted therapies even at an early stage of their development. (See Exhibit 4.) They're less likely to buy in vaccines or immunostimulants without at least Phase II data. The late 2004 deal between Bristol and Medarex Inc. covered an immunostimulant and a vaccine, both beyond Phase II [See Deal]; likewise Coley Pharmaceutical Group Inc. 's toll-like receptor 9 agonist ProMune was in Phase II trials when Pfizer bought rights in 2005. [See Deal] Cytotoxics, particularly in new formulations, are out for most pharma—they're seen as me too's. On the other hand, they see gene- and cell-therapy candidates as too risky. Thus biotech companies developing these types of therapy will likely have to progress them to Phase II or Phase III trials before finding a partner.

Targeted therapies may throw up new, potentially lucrative subsegments and opportunities for smaller firms chasing the advantages associated with niche drugs, but they will also present more complex development challenges than classical cytotoxics. The efficacy of cytotoxics is measured by tumor shrinkage, but many targeted products slow tumor progression without reducing tumor volume. And they typically only slow progression in a subset of patients. If predictive markers aren't identified for that subset, trials may not be sufficiently strong-powered to show efficacy. Clinical results for EGFR inhibitors erlotinib (Tarceva) (developed by OSI Pharmaceuticals Inc. and sold with Roche and Genentech [See Deal]) and AstraZeneca PLC 's gefitinib (Iressa) in lung cancer, for example, were mixed, since one obvious marker, receptor expression, turned out to be only a weak predictor of responsiveness. (See "Problems with Biomarkers: A Tale of Three Drugs," IN VIVO, April 2005 (Also see "Problems With Biomarkers: A Tale of Three Drugs" - In Vivo, 1 Apr, 2005.).) It took many large trials to determine more useful markers (e.g., receptor gene mutation or amplification), through post- rather than pre-selection. Gleevec's an exception in this regard, since in the case of GIST tumors, the gene abnormality that determines response is part of the disease definition. (About 95% of those with GIST tumors will respond to Gleevec.) Herceptin remains the only example of a drug whose use is determined by an identified marker—the Her-2/neu gene--that's not part of the definition of the disease.

Targeted therapies will thus require, and drive, the development of markers for the prediction of individual responses. This will represent a big change from the probabilistic drug selection method used when only a few products were available, and when tumor shrinkage was the best measure of efficacy. The development of those markers may significantly reduce the target population, and add development time and cost, since drug labeling is by disease, and not by marker, so registration in other cancers that express the same target will require a full development plan. Many companies with Phase I programs will struggle to bring their compounds through to successful proof of concept in the absence of predictive markers of activity.

Oncology will remain an attractive area for biotech and pharma. SG Cowen forecasts that the market (including adjuvants) will grow from $42 billion in 2004 to $70 billion in 2009. But the field is changing, and the odds appear to be shifting against biotechs going it alone. Product positioning will become more important in a larger market with multiple products and a few blockbusters, and where combinations are becoming more important. Some indications that look attractive based on the present product offering may become less attractive in five years' time, in particular if epidemiological data do not support the entrance of multiple products. High prices and short approvals are no longer guaranteed, and the advent of targeted therapies increases development complexity—and cost.

Fortunately for biotechs, Big Pharma is in-licensing, but it's choosing carefully. So must biotechs in deciding when to out-license, and how to broaden their portfolio.

François Thomas, MD, is a medical oncologist and principal at Bioserve Ltd, a consultancy dedicated to biotechnology and oncology. He is also a venture partner at VC firm Atlas Venture.