When New Drugs Fail: An Analysis of the SU5416 Trial Experience

SANTA MONICA, California-Vascular endothelial growth factor (VEGF) offers a number of sites for therapeutic targeting and is thought to play an important role in human cancers. Several angiogenesis inhibitors that work through effects on VEGF are under development, but one that failed in a phase III study may offer important insights for investigators developing this new area of anticancer therapy according to Lee S. Rosen, MD. Dr. Rosen, who was involved in research on the tyrosine kinase inhibitor SU5416, is director of developmental therapeutics, Cancer Institute Medical Group, Santa Monica, California.

"SU5416 is a tyrosine kinase inhibitor that made it all the way to phase III clinical trials in advanced, untreated colorectal cancer, but patients given regimens of fluorouracil (5-FU)/leucovorin with or without SU5416 showed absolutely no difference in their survival," Dr. Rosen said. "These were the results that halted the development of this compound. The key question is ‘What happened?’ Can we use some Monday-morning quarterbacking of our own data to help other drugs that are in clinical development?"

Push for Rapid Testing

Dr. Rosen said that SU5416 went into clinical trials partly because it was effective in several animal models of metastases. There was also additive activity with 5-FU in preclinical studies. At the time the drug was pulled from clinical development, almost 2,000 patients had been treated. More than 1,500 had taken the recommended phase II doses, many in combination with various cytotoxic chemotherapy regimens. Many had been treated for more than 6 months, some longer than 1 year. The manufacturer had sponsored several phase I, phase II, and phase III trials, and the National Cancer Institute had commissioned several more.

No classically defined responses were seen in the original phase I trial, which was a study of about 60 patients. Many had progressed very quickly on cytotoxic chemotherapy, and then the disease remained stable for a year or a year and a half, something very important, but no responses were seen," Dr. Rosen said.

He attributed the push for rapid testing of SU5416 to two factors: hints that the drug may be active in colorectal cancer, and the size of the colorectal cancer market. ‘‘There was a hint of activity in early trials in a woman who was very heavily pretreated and had shrinkage of one of her tumors, although others grew," Dr. Rosen said.

No Easy Answer

SU5416 was pulled from development for colorectal cancer following results from the phase III trial. Questions about dose adequacy emerged when a new assay, which became available only after the trial had begun, showed that SU5416 inhibited VEGF receptor activity well after 4 hours but then declined quickly. Phase II dosing was twice weekly, on the assumption that the drug would remain active for 72 hours.

Dr. Rosen said that this trial adds weight to the argument that selection of proper endpoints for clinical trials should be reconsidered and that there is a need for better surrogate markers of efficacy. "However, with SU5416, an extensive preclinical program delineated the target, showed that the drug hit the target, and used some surrogate markers that, although imperfect, were the best available in 1997 when the clinical trial started," Dr. Rosen said.

"In the SU5416 trial, the surrogate markers we thought would be significant just weren’t. That tells us that there is no easy answer in designing the perfect clinical trial to validate these targets."

Dr. Rosen proposed using phase I studies to test hypotheses about surrogate markers. "Reasonable use of biopsies and imaging to try to validate proposed surrogate markers is important so that when we are trying to get to phase II or phase III studies, we can use those surrogate markers more knowledgeably and perhaps make better decisions about whether a particular molecule is worth taking forward into further studies," he said.

Toxicity/Efficacy Balance

Clinical investigators also need better ways to look at drug-specific, class-specific, and tumor-specific adverse events. "With the VEGF inhibitors there seem to be a group of common toxicities that can be managed. We should recall that drugs such as cisplatin can cause severe toxicity but can also be extremely effective,’’ Dr. Rosen noted.

"Where a drug is extremely effective, we can accept toxicity if it can be managed," he continued. "In a drug with marginal or no clinical benefit, toxicity becomes a more important issue." Exploring toxicity issues might also help investigators understand the pathways of these new drugs.

Dr. Rosen said that investigators focus on which patients respond and often ignore those who do not. "Looking at differences between paths to success and paths to failure can help us understand the biology and activity of a particular drug," he said.

With regard to toxicity, Dr. Rosen said that SU5416 was extremely difficult to administer because it is cremophor-based and requires a paclitaxel-like premedication regimen. "The dose-limiting toxicities were severe headaches and projectile nausea and vomiting refractory to aggressive care. Toxicities were reversible, however, and tended to decrease over time."

Dr. Rosen said that even if SU5416 does not get developed further, there are data on over 1,200 patients and samples stored in many research laboratories that might provide useful information about why the drug failed. "This might help us understand how to prevent the same thing from happening with other drugs in this class or with other classes of drugs."

Other VEGF Inhibitors

Other VEGF inhibitors under development include PTK787, an orally bioavailable VEGF receptor tyrosine kinase inhibitor. Phase I studies with oxaliplatin (Eloxatin), 5-FU, and leucovorin reported no apparent pharmacokinetic interactions, Dr. Rosen noted, so phase II and phase III studies are planned.

ZD6474 is also an orally bioavailable VEGF receptor tyrosine kinase inhibitor. Phase I studies suggested that the compound inhibits angiogenesis, and might cross-react with the epidermal growth factor (EGF) receptor. Toxicities included skin rash and diarrhea, which Dr. Rosen said were probably due to crosstalk with the EGF receptor.

"Finally, we also need to understand better the regulatory approval process and, as scientists, continue to participate. Faced with direct doctor-patient interactions, we can help communicate what is important to those patients," Dr. Rosen said. "In the design of clinical studies using novel agents, maybe we are setting the bar too high when we are looking at patients with advanced disease."