Bladder Cancer Agent Fails to Pass Muster With ODAC

WASHINGTON--Citing inadequate data to prove efficacy, ODAC (Oncologic Drugs Advisory Committee) voted unanimously against recommending that the FDA approve the new drug application for Anthra Pharmaceuticals’ valrubicin (Valstar).

The drug, an intravesically administered anthracycline, was tested (as AD32) in 90 patients in two phase II nonran-domized trials for the treatment of refractory carcinoma in situ (CIS) of the urinary bladder.

"Valrubicin’s safety is not in question," said ODAC consultant George Sledge, MD, professor of medicine and pathology, Indiana Cancer Pavilion, Indianapolis. "However, efficacy of the drug has not been demonstrated in these trials."

Part of Anthra’s argument for approval of Valstar was based on the lack of any approved drug treatment options for CIS of the urinary bladder once a patient has failed several courses of BCG therapy.

"Once BCG treatment fails," said H. Barton Grossman, MD, of M.D. Anderson Cancer Center, speaking on behalf of Anthra, "salvage strategies must be considered." Further BCG treatment pro-duces more risk than benefits, he said, and radiotherapy "doesn’t work." Mito-mycin (Mutamycin) shows only a 7% complete response rate, and interferon produces low, nondurable responses. Radical cystectomy entails 30% morbidity and poor quality of life, and may be refused by some patients. "BCG is the prime treatment," Dr. Grossman said, "and BCG-refractory cancer is the primary problem. An alternative is needed."

The ODAC panel’s objections involved two efficacy issues: (1) Did complete response correlate with patient benefit in this nonrandomized trial? (2) Did the treatment actually change the patients’ disease or can the observed benefits be ascribed simply to patient heterogeneity?

The company cited complete response rates of 44% at 3 months, 21% at 6 months, 14% at 12 months, and 12% at 15 months. However, the FDA felt that only 7 of the 20 complete responses claimed by the company were verifiable because these were the only patients with documented multifocal disease at baseline.

Some panel members argued that, in any case, complete response was not an appropriate endpoint in this trial.

Time to cystectomy was delayed in responders (median, 23 months vs 8.3 months in nonre-sponders). However, "a significant proportion of patients who failed did not go on to cystectomy," said panelist Howard I. Scher, MD, of Memorial Sloan-Kettering Cancer Center. Only 37 of 60 patients who failed on Valstar therapy went on to cystectomy.

"Is a complete response in fact a clinical benefit?" Dr. Sledge asked. "Does this benefit improve the time to cystectomy?" And, he continued, is a delay in cystectomy a real benefit or simply due to a different natural history of the disease between responders and nonre-sponders? "I am concerned that this indicates selection bias in a nonrandomized trial," he said.

Robert Temple, MD, of the FDA, called this "a classic phase II vs phase III problem. You can document efficacy, but can you conclude there is sufficient benefit without a concurrent control group?"

Some panel members suggested that further data from Anthra might help support the company’s efficacy claims.