Predicting Success: A Pipeline Report

Much has been said about pharma's R&D lag, but a closer look at the candidates currently in Phase III shows that what the pipeline lacks in quantity it makes up for in quality. This Pharmaceutical Executive pipeline report identifies some of the top candidates in five categories: women's health, oncology, central nervous system (CNS), cardiovascular disease (CVD), and metabolic/ endocrine diseases.

Data on drug candidates were provided by the Pharma Solutions division of the medical publishing giant Wolters Kluwer Health, which gave Pharm Exec access to its extensive Adis RD&I database, containing a wide array of information on thousands of candidates from pre-clinical through Phase III, and by the investment bank Lehman Brothers, which tracks and evaluates numerous products in its Pharma-pipelines database. In the listings that follow, drugs are ranked by a proprietary 1-100 score, developed by Adis, that assesses clinical impact.

Adis's ratings are based on 20 parameters and are created by editors who monitor medical literature, attend conferences, and use knowledge gained from other proprietary sources. In the Adis system, a compound with no evident advantages compared to existing treatments receives a 45 or below. A compound with modest benefits in safety, efficacy, or both receives a 45-59; a compound with important benefits, 60 or above. A rating of 80 and up suggests substantial clinical advantages; in some cases a rating of 80 means there are no alternative treatments. An Adis score of 55 was the cutoff in these listings.

Women's Health

Wherever possible, the listing for a candidate also includes Lehman Brothers' assessment of the likelihood of approval (see "Chance of Success," page 66) and an assessment of potential peak sales. Without exception, in each therapeutic area, the compound with the highest projected peak sales is among the products with the top Adis ratings.

Key findings include:

• The two highest rated compounds are a drug to treat osteoporosis and a recombinant enzyme replacement therapy in the metabolic/endocrine diseases category.

• Oncology as a class has the most agents that could substantially improve therapeutic options.

• Drugs with the highest projected earnings are always among the top-ranked, but not necessarily at the very top.

• Drugs originated by small biotech firms benefit from out-licensing to larger companies.

• A first-in-class selective progesterone receptor modulator could change the landscape of women's health.

• Genomics-based drugs are among the top-ranked candidates in every therapeutic area analyzed.

• Selective estrogen receptor modulator (SERM) development is prolific in the areas of oncology and osteoporosis.

• Promising new CNS drugs for stroke and head injury may effectively confer neuroprotection after trauma.

For Women Only

Across all five therapeutic categories, two compounds are rated 86-the highest number assigned to any in this analysis. One of those is a Pfizer compound for postmenopausal osteoporosis. (See "Women's Health." )

Lasofoxifene. This selective estrogen receptor modulator (SERM) inhibits bone loss and lowers cholesterol and is entering Phase III trials to evaluate its use in breast cancer prevention. Its projected peak sales are $1 billion a year, one of the highest in this report. Currently, the only SERM available for the prevention and treatment of osteoporosis is Eli Lilly's Evista (raloxifene), a first-in-class therapy.

According to researchers, lasofoxifene "promises to be more potent than any other SERM." In a one-year Phase II trial, it was studied head-to-head with raloxifene. Not only did lasofoxifene increase bone mineral density (BMD) compared with placebo, but lasofoxifene .25mg was superior to raloxifene 60mg in increasing lumbar spine BMD. Moreover, the drug decreased LDL-cholesterol by 25 percent compared with a 17 percent reduction by raloxifene. Pfizer is compiling more data with Phase III clinical trials, enrolling more than 10,000 postmenopausal women to evaluate lasofoxifene's efficacy in increasing BMD and treating fracture.

Oncology

Felicia Cosman, MD, clinical director for the National Osteoporosis Foundation says, "It would be good if the SERMs under development could show increased efficacy in reducing fractures at sites other than the spine. That's the type of clinical trial data that matters. Also, it would be better if a SERM did not cause hot flashes or increase the risk of venous thromboembolism."

Arzoxifene. Eli Lilly is harnessing its therapeutic experience with raloxifene in its efforts to launch another SERM, arzoxifene (rated 72), which has estrogen-agonist activity in bone and serum lipids and estrogen-antagonist activity in breast and uterine tissue. Arzoxifene is in development for osteoporosis, breast cancer prevention, and endometrial cancer and has peak sales potential of $600 million.

S12911. French pharma company Servier, developer of this stromium-containing osteoporosis drug, claims that it both stimulates bone formation and inhibits bone reduction. In a three-year clinical trial with 1,649 postmenopausal women, the drug reduced the risk of new vertebral fractures by 41 percent compared with placebo.

In a separate 24-month study of 160 patients, women treated with S12911 experienced increased BMD at the spine (5.53 percent), femoral neck (2.46 percent), and hip (3.21 percent). Based on these studies, the safety and tolerability profile of S12911 was acceptable. The candidate has a rating of 80, a 60 percent chance of being approved in the European market, and projected peak sales of $300 million.

Preos. The third osteoporosis treatment on the list is a recombinant human parathyroid hormone, which differs from currently available antiresorptive treatments because it is anabolic. That means it has the potential to reverse bone loss rather than just reduce bone resorption. Despite its subcutaneous injection administration, it has a rating of 75, a 75 percent chance of approval, and projected sales of $650 million.

Its developer, Canadian-based NPS Pharmaceuticals, is ushering the compound through a series of safety studies to rule out carcinogenicity. In efficacy studies, patients experienced dose-dependent BMD increases of 2.4-6.9 percent at the lumbar spine. In September 2003, Phase II/III data showed that although hip BMD was unchanged after 12 months of treatment, when Preos was combined with alendronate, there were significant increases in hip BMD. Since then, however, additional data have been negative.

Asoprisnil. Perhaps the most exciting compound in this cluster, asoprisnil is the first selective progesterone receptor modulator (SPRM) to enter Phase III trials. Currently very few clinical data are available, but Takeda's early studies show that asoprisnil may have the potential to treat a variety of disorders, including endometriosis, fibroids, excessive uterine bleeding, and, in conjunction with HRT, to protect the breasts and endometrium. Asoprisnil's rating is 69, with a 20 percent chance of success, and a market potential of $700 million.

Central Nervous System

"This is an interesting class," says Whitworth. "If this agent truly does not increase endometrial thickening or the incidence of strokes or blood clots, it could be very useful, particularly in younger women."

HPV vaccine. Another notable compound in this group is the human papillomavirus (HPV) vaccine developed by Stressgen and in-licensed by Roche for the treatment of anal dysplasia. It elicits a CD4+-independent cytotoxic t-cell response and would be suitable for use in immunocompromised patients. Rated 60, the candidate has a 15 percent chance of approval and a potential to earn $500 million annually in sales.

Cancer Compounds

The 395 oncology agents in the pipeline exemplify the groundbreaking research taking place on many fronts. (See "Oncology.") Likewise, the top ten selected for this report represent the breadth and depth of Phase III research and development. Of all of the therapeutic areas analyzed, oncology-with 85 compounds currently in Phase III development-is the most prolific. It also has the highest number of compounds rated 70 or above, which means that their approval would offer treatment benefits comparable with existing therapies in efficacy, safety, or both. Moreover, of all the drugs analyzed, oncology compounds have the highest projected peak sales.

Cardiovascular Disease

Fenretinide. An oral compound for breast cancer being developed by Ortho-McNeil and the National Cancer Institute leads with a rating of 85. Fenretinide is a synthetic retinoid that has demonstrated efficacy as a chemotherapeutic agent in women with early-stage breast cancer. In a five-year study of 2,972 women with stage-one breast cancer, premenopausal women who were treated with fenrenatide experienced a reduced incidence of local recurrence and contralateral development of breast tumors relative to their placebo- treated counterparts. Yet, that finding did not hold true for postmenopausal women.

In a separate analysis of 2,867 women with early breast cancer, women treated with 200mg of fenrenatide once daily for five years had a significantly lower risk of ovarian cancer compared with placebo-treated controls (0 vs. 6 cases). However, the effect only lasted for the duration of the intervention. Fenretanide has also been evaluated in conjunction with tamoxifen therapy, and the combination has shown positive results. In some cases, the regimen resulted in partial or complete remissions and disease stabilization.

Efaproxiral. Rated 80, the lead clinical candidate for Colorado-based Allos Therapeutics (ATI) has a proven efficacy that is attributable to its activity as a hemoglobin modifier. More specifically, the IV-infusion therapy increases the oxygen-carrying capacity of hemoglobin while simultaneously inducing cytotoxic effects and providing neuroprotection. Researchers believe that the adjuvant use of a radiosensitizing agent such as efaproxiral can alter the treatment landscape for those patients, partially because it does not have to cross the blood-brain barrier to be effective.

In a Phase II study of 69 patients with metastatic brain cancer, the first group received efaproxiral and radiotherapy, while the second group received radiotherapy alone. During follow-up, the median survival time for patients who received efaproxiral was 6.9 months compared with 4.1 months for the group receiving radiotherapy alone. The 12-month survival rate for group one was 32 percent compared with 15 percent for group two.

Tipifamib. Developed by Janssen Pharmaceutica (a J&J company), this drug is the first farnesol-transferase inhibitor to enter clinical trials. As a class, farnesol-transferase inhibitors directly affect endothelial cell growth and stimulate apoptosis (programmed cell death). Along with a good tolerability profile, tipifamib has the advantage of oral dosing, but researchers warn that cancer-cell resistance to agents in this class could be a treatment-limiting problem.

Phase I and III clinical trials demonstrated moderate response from patients with various types of cancer. In 27 patients with previously treated advanced breast cancer, for example, tipifamib produced a partial response rate of 12 percent in patients, with a resulting stable disease rate of 35 percent. In another study, tipifamib had modest activity in patients with chronic myeloid leukemia (CML) and myelofibrosis but was inactive in patients with multiple myeloma. The drug is rated 80, has a 30 percent likelihood of approval, and projected annual peak sales of $500 million.

Vatalanib. Another mechanism of action being refined in the research and development process is vascular endothelial growth factor (VEGF) receptor tyrosine kinase inhibition. A notable example of this class is vatalanib, the compound with the highest projected peak sales ($1.3 billion).

Metatbolic and Endocrine Disorders

Novartis and collaborator Schering AG have combined forces to develop this novel agent to be used in combination with a chemotherapy regimen as part of first- and second-line treatment for metastatic colorectal cancer. Its likelihood of approval is estimated at 25 percent. Phase III trials started in April 2003, and so far vatalanib has demonstrated linear kinetics and good tolerability.

Early clinical trials involving more than 700 patients showed a significant reduction in tumor blood supply based on MRI imagery, and tumor vascular response was seen within 36 hours of treatment. According to Joerg Reinhardt, PhD, head of pharma development at Novartis Pharma AG, "Vatalanib is the first VEGF inhibitor, and during treatment it provides a higher level of specificity to cancer cells and less toxicity."

There are currently two Phase III trials testing the efficacy and safety of vatalanib. One is evaluating 1,090 patients with previously untreated colorectal cancer and the other is evaluating 830 patients with previously irinotecan-treated colorectal cancer that has progressed.

Still, vatalanib, along with the other top-ranking cancer agents, is intended for use as part of a combination regimen. "From an oncology perspective, we are interested in new therapies that make chemotherapy more tolerable," says Reinhardt, adding, "In the long-term, the hope is that you won't need the less specific therapy at all, but at this point that's still ten years away."

Oregovomab. This is one of a few vaccines in development for cancer that are thought to elicit their response on the epitope of the antigen. Early clinical trials were promising, because patients with ovarian cancer who generated immune responses to the targeted vaccination survived longer.

Rated 73, Oregovomab is a monoclonal antibody that targets the ovarian tumor-associated antigen CA-125. AltaRex has partnered with Unither Pharma to extend marketing reach, and presumably, to drive the development process forward. United Therapeutics recently initiated two Phase III trials to evaluate oregovomab for the treatment of advanced ovarian cancer following successful completion of first-line treatment.

Chance of Success

AE 941. Administered orally, this is the first anticancer agent derived from shark cartilage to demonstrate efficacy without serious adverse events in patients with solid tumors. Rated 71, AE 941 is yet another angiogenesis inhibitor that induces cell apoptosis.

Miproxifene. This SERM (rate at 71) shows higher selectivity for breast cancer cell lines than tamoxifen, with less hematological toxicity. Preclinical safety data are positive, but more Phase III clinical data are needed and the outlook is still uncertain. "I think that miproxifene will have a tough time competing against the aromatase inhibitors," says Pat Whitworth, MD, director of the Nashville Breast Center in Tennessee. He adds that, although it may not compete well in the market for treating postmenopausal women, there may be potential for using miproxifene to treat premenopausal women.

Atrasentan. Developed by Abbott Labs in collaboration with Eisai, this endothelin A receptor antagonist is intended for use in men with hormone-refractory prostate cancer. The likelihood of market approval for Atrasentan is 30 percent, estimated peak sales are $500 million, and its rating is 70.

MS-209. Schering AG is seeking approval of this oral cancer treatment in Europe. It currently has a rating of 66, a 15-30 percent chance of approval, and a peak sales potential of $50 million annually. MS-209 has shown promising results in clinical trials for treating multidrug resistant cancer when used with paclitaxel.

Decitabine. An IV formulation with a rating of 65, decitabine is being developed by SuperGen for the treatment of myelodysplastic syndromes. In clinical trials, it has been useful as a single agent for the treatment of blast phase myelogenous leukemia. However, like oncology candidates, it may have more utility as an adjuvant therapy.

CNS Candidates

Top-ranked compounds for central nervous system disorders address a broad range of conditions, including brain cancer, head injuries, stroke, sleep disorders, sexual dysfunction, multiple sclerosis, and fibromyalgia. Surprisingly, none of the drugs in development for Alzheimer's disease, depression, or schizophrenia made the cut. (See "Central Nervous System.")

PN 246. The inclusion of a standardized extract of Ginkgo biloba leaf is surprising. Its rating of 80 is even more remarkable. Denmark-based Pharma Nord is seeking approval of PN 246 in the United Kingdom for the treatment of fibromyalgia. There is no projection of peak sales or likelihood of approval. PN 246 is currently available in Denmark as an other the counter product to improve well-being and is being tested in clinical trials in combination with co-enzyme Q-10.

Initial clinical results have shown "improved quality of life" with some reduction in the symptoms of fibromyalgia, a disorder of unknown etiology characterized by musculoskeletal pain and fatigue.

Dapoxetine. This selective serotonin reuptake inhibitor (SSRI) is being evaluated for the treatment of premature ejaculation and is slated to be the top earner in this group, earning a potential $1 billion in peak sales. That figure is justified in light of the potential market for premature ejaculation, which affects 30-70 percent of the male population and has not been attributed to an organic cause.

Yet, although Eli Lilly and J&J are the co-developers, its chance of gaining market approval is only 25 percent.

The safety profile is good; dapoxetine is structurally similar to Prozac (fluoxetine). Furthermore, Phase II clinical data assessing dapoxetine's effects on ejaculatory latency time (ELT) in 155 men showed a significant increase in ELT relative to placebo.

Repinotan and disufenton. Both candidates are currently being tested for their use in treating stroke. Nevertheless, although each agent can be broadly defined as neuroprotective, there are differences. Repinotan (rated 70 with a 30 percent chance of success) is the first selective, high-affinity serotonin 1A receptor agonist to demonstrate potent neuroprotective activity in preclinical studies. Developed by Bayer, it has projected peak sales of $500 million and may change the treatment paradigm for ischemic stroke and traumatic brain injury.

A Phase II study evaluated its effect on recovery in stroke patients. In each of the 240 subjects, repinotan was administered within six hours post-stroke. A daily dose of 1.25mg resulted in measurable improvement in 78 percent of patients based on a four-point decrease in the National Institutes of Health Stroke Scale, compared with improvement in 67.3 percent of placebo-treated patients. Yet, despite encouraging data, Bayer is approaching Phase III cautiously. "There's a discouraging collection of medical literature documenting stroke drugs that have failed in clinical trials," says an NIH official.

Disufenton, on the other hand, has a rating of 60. Although animal models show that disufenton reduces infarct volume and functional deficits following middle cerebral artery occlusion, there have not been any well designed trials to assess efficacy in humans. According to Lehman estimates. it could gross $250 million annually, but there is only a 25 percent chance of approval. Within the category of stroke and head injuries, the approval rate is historically low, so that probability of success could be viewed optimistically.

Dexanabinol. Like the stroke market, there are no reliable first-line agents for head injuries. As a synthetic nonpsychotropic cannabinoid that combines NMDA receptor antagonism, free radical scavenging, and anti-inflammatory activity, this agent has a broad range of activities. The candidate is being developed by Pharmos and has a rating of 69.

Natalizumab. Elan is developing this humanized anti-alpha4-integrin monocolonal antibody for treating multiple sclerosis. It's rated 67, has a 40 percent chance of approval, and has a market potential of $300 million annually.

Insomnia therapies. Sleep disorder candidates comprise a third of the top-ranked CNS agents, but their likelihood of approval is relatively low, and projected peak sales are under $250 million for each agent. Rated 65 and 60, respectively, agomelatine (Servier) and ramelteon (Takeda) are melatonin agonists, and Lundbeck's gaboxadol (rated 57) is a potent selective GABA-A agonist.

LAX-101. San Francisco-based Amarin is applying its expertise in neurodegenerative diseases to develop this compound to treat Huntington's disease, a genetic neurodegenerative brain disorder for which there is no effective treatment. So far, clinical trials have yielded no data that demonstrate a significant benefit, but trials continue.

Cures for the Heart

The pipeline for cardiovascular candidates is small compared with other therapeutic categories, and, for the most part, moderately rated. (See "Cardiovascular Disease.") According to Jeffrey Borer, MD, chief of the division of cardiovascular pathophysiology at Weill Medical College of Cornell University and chairman of FDA's Cardio-Renal Advisory Committee, there is a logical explanation: "In oncology if you have an effective drug for a disease that is uniformly fatal over a five-year period, you don't have to study too many patients to get approval, but in CVD, very few diseases are uniformly fatal over five years. To show more than a symptomatic effect, you need to enroll large study populations, which costs a lot of money. Companies are understandably careful about how they invest their money."

Conivaptan. Four of the nine top-rated cardiovascular agents are being developed to treat heart failure. Heading the list is conivaptan, an oral arginine vasopressin V1/V2 receptor antagonist. Although it is rated 74 and has a 50 percent success rate, its projected peak sales are only $38.4 million. Essentially, conivaptin is another option for enhancing the diuretic process, and numerous options are out there now.

Still, there is an important niche for the drug. "In people with severe congestive heart failure, the diurectics we have are not effective at relieving fluid," says Borer. "So one would like to have additional diuretics in the armamentarium."

OPC-6535. Another drug in development for heart failure, this Otsuka compound (rated 70) relies on leukocyte activation inhibition to prevent the inflammatory processes associated with heart failure by blocking cytokine production in the myocardium.

Ecadotril. Rated 69, this Bayer candidate (S-enatiomer of acerphan) increases elimination of fractional sodium and urinary cyclic monophosphate. In an open study, plasma atrial natriuretic factor (ANF) levels increased two to three times in ecadotril-treated patients and subsequently resulted in a significant decrease in pulmonary capillary wedge pressure.

Tedisamil. Solvay developed this potassium channel blocker independently and, until recently, it did fairly well in clinical trials. Tedisamil, rated 68, shows dose-dependent efficacy for rapidly converting recent onset atrial fibrillation. Unfortunately, safety concerns about its possible proarrythmic effects have led to the temporary interruption of clinical trials. Lehman still predicts a 60 percent chance of approval and estimates peak sales of $300 million. Solvay's current strategy is to develop tedisamil as an oral agent for the short-term treatment of atrial fibrillation.

Tezosentan. This candidate is designed to relieve inflammation and vasoconstriction through endothelin receptorA/B (ETA/ETB) antagonism. The presence of ETA and ETB receptors on the outer walls of blood vessels makes them logical targets, because of their tendency to induce inflammatory responses when the heart becomes diseased. Developed by Roche and Genentech, tezosentan is rated 57 and has a 25 percent chance of success and projected sales of $150 million.

Amediplase. Menarini, an Italian pharma company, is testing this compound for myocardial infarction. It combines tissue plasminogen activation and urokinase plasminogen activation to induce thrombolysis. According to the company, "amediplase is more potent and long lasting than alteplase," which has been used as a thrombolytic agent in patients with myocardial infarction since 1987. One major benefit could be that amlediplase (rated 66) is administered by bolus injection.

Roxifiban. Rated 55, the lowest threshold for inclusion in this analysis, this cardiovascular drug has one of the highest projected peak sales: $1 billion annually. Bristol-Myers Squibb is developing this potent selective, long-lasting glycoprotein IIb/IIIa receptor antagonist independently for the treatment of thrombotic disorders.

Likelihood of approval is only 15 percent, but that can partially be attributed to a late NDA filing resulting from dosing issues. The focus will now be on twice-daily dosing rather than once-daily. In Phase II trials, tolerability data was positive in six-months-plus studies involving more than 1,000 patients.

Metabolic/Endocrine Disorders

IMS Health predicts that the market for diabetes treatment alone will be $7.3 billion by 2009, based on a 23 percent annualized growth rate. Yet, despite that huge market, the analysis yielded a total of five agents for inclusion in the top-ranked category. (See "Metabolic/Endocrine Disorders.") In many ways, diabetes management is state-of-the-art, but incremental progress is still underway, especially in the areas of drug delivery and controlling patients blood sugar levels.

David Brillen, MD, director of the Cornell Diabetes Center at the Weill Medical College of Cornell University sees a future in which more people are candidates for treatment with agents developed for diabetes. Soon, targets for treatment may include individuals who present with dysmetabolic syndrome, a cluster of metabolic abnormalities, including high blood pressure, dyslipidemia, glucose abnormalities, and inappropriate fat distribution or outright obesity.

Aryplase. Though diabetes treatments dominate the metabolic/endocrine category, the top-rated drug-which shares an 86 rating with lasofoxifene-is an enzyme replacement therapy for a metabolic disorder that affects only 1,100 people worldwide. Aryplase, a recombinant human arylsulfatase B, is being developed by BioMarin for children and adolescents with mucopolysaccharidoisis VI. Arylsulfatase B is essential for the breakdown of certain complex carbohydrates, and its deficiency leads to cellular lysosomal build-up throughout the body, resulting in slow growth, upper airway obstruction, enlarged liver, and skeletal deformities.

Aryplase's efficacy and safety data are positive, and enduring results have been demonstrated in clinical trials. Lehman financial data are unavailable, but clearly, this drug will have limited sales.

Inhaled insulin. Novo Nordisk has in-licensed Aradigm's insulin-inhalation delivery device, which has demonstrated safety, efficacy, and dose reproducibility comparable with an injectible insulin regimen. The system uses liquid insulin, which is converted into an aerosol containing small particles, and an electronic device, which can be used for rapid transfer of insulin into the bloodstream or localized pulmonary delivery. Brillen compliments the inhaler's design but has concerns about customizing therapy for different patients. "It's difficult to inhale smaller units of insulin," he explains, "so even if this works for someone with Type 2 diabetes, it may not work for adolescents, children, or extremely slender individuals who need smaller units of insulin." The product, rated 78, is expected to gross $500 million when it peaks and has a 50 percent chance of approval.

Exenatide. Lilly's glucagon-like peptide 1 (GLP-1) analog, in-licensed from Amylin, could provide an effective treatment for patients who are not using insulin and are not controlling their blood sugar with available oral antidiabetics. The excitement about this drug-which is rated 77 and has a 70 percent chance of approval in the United States and projected sales of up to $1 billion-is justified.

According to Brillen, GLP-1 does two important things: it increases endogenous pancreatic production of insulin and suppresses glucagons, which means it could become a first-in-class drug. "I think this would be a good supplement to existing therapies," Brillen says. "The only drawback I see is that it's injectable."

In a Phase III trial, after 28 days, 43 percent of exenatide-treated patients had glycosolated hemoglobin (HbA1c) levels less than 8 percent, compared with 5 percent of patients treated with a placebo.

Tesaglitazar. AstraZeneca's independently developed perisome proliferator-activated receptor (PPAR) antagonist could be a source of interest based simply on its profile. This compound is an oral insulin sensitizer with superior potency to the glitazones Avandia (rosiglitazone) and Actos (pioglitazone), in improving glycemic control and glucose tolerance.

According to financial projections, this compound could gross $750 million, and its likelihood of approval has recently been upgraded from 15 percent to 40 percent. Brillen says, "There is the possibility that this drug may be useful in a broader range of people, including those at risk of diabetes, dysmetabolic syndrome, or polycystic ovarian syndrome." He also believes PPARs are important targets, because of "the wealth of evidence showing their role in fat metabolism and the ability of PPAR antagonists to reduce free fatty acids."

The Power of Partners

The Pharmaceutical Research and Manufacturers of America estimates there are more than 1,000 compounds in active clinical development. Predicting which drugs are likely to succeed is a complicated undertaking. Researchers at the Wharton School at the University of Pennsylvania developed a system to determine which of the 960 compounds they analyze were likely to be approved by FDA.

The Wharton model evaluated several critical factors to predict success:

• a company's overall experience based on the number of compounds it was involved with between 1988 and 2000

• experience in the relevant therapeutic category

• diversification of experience

• alliances with small and large com-panies: small (three compounds), medium (4-24 compounds), or large (25 or more compounds).

Using regression analysis, researchers found that compounds co-developed in alliances are more likely to be approved than those developed independently. That is generally true for drugs in Phase II or III trials that are in-licensed by a large company. They also found that companies with considerable experience are more likely to advance in clinical trials. Specifically, they found that 55 percent of Phase II candidates are co-developed, compared with 62 percent of Phase III therapies.