Nick Mulcahy

June 14, 2012

June 14, 2012 (Chicago, Illinois) — New data on abiraterone acetate (Zytiga, Janssen/Johnson & Johnson) "merit consideration" as a "new standard approach" in the first-line treatment of metastatic prostate cancer.

So said the lead researcher of a phase 3 study in which the drug was used in patients untreated with standard chemotherapy, the only approved first-line therapy in this setting. The study was presented here at the 2012 Annual Meeting of the American Society of Clinical Oncology (ASCO).

However, a biostatistician who served as study discussant at the meeting cast doubt on some of the data, particularly the primary end points of the trial.

Susan Halabi, PhD, from Duke University in Durham, North Carolina, also criticized the conduct of the trial, which was unblinded early at the recommendation of its data monitoring committee.

The trial should not have been stopped then.

"In my opinion, the trial should not have been stopped then," she told Medscape Medical News in an interview.

The trial in question, known as COU-AA-302, compared daily abiraterone with placebo in men with asymptomatic or minimally symptomatic metastatic castration-resistant prostate cancer (mCRPC) who had not been treated with chemotherapy. All men in the study also received daily prednisone.

Abiraterone, an oral agent, is one of a number of therapies approved in the past couple of years for the treatment of mCRPC, but only as a second-line treatment after chemotherapy.

This trial is important, in part because the various new treatments for mCRPC are "promising" but have not yet been proven effective as an initial therapy, said lead author Charles Ryan, MD, from the Helen Diller Family Comprehensive Cancer Center at the University of California, San Francisco.

The trial was stopped after a second planned interim analysis because of "compelling evidence of clinical benefit," Dr. Ryan said when he relayed the assessment of the independent data monitoring committee.

Dr. Halabi pointed out that after the trial was unblinded, it was no longer a randomized controlled trial; instead, it was an observational trial. If the trial had been further along, the results would have been more definitive, she explained. "You want the evidence to be both statistically and clinically persuasive," she noted.

However, the manufacturers of abiraterone apparently believe that the new data constitute strong evidence as is. Janssen has submitted drug approval applications for abiraterone to the US Food and Drug Administration and to the European Medicines Agency. Both applications are intended to extend the use of abiraterone plus prednisone to include the treatment of patients with mCRPC who are asymptomatic or mildly symptomatic after the failure of androgen-deprivation therapy and before chemotherapy.

Primary End Points in Question

The trial involved 1088 patients who were randomized to abiraterone (1000 mg daily plus prednisone 5 mg twice a day) or placebo (daily plus prednisone). It was conducted at 151 sites in 12 countries. Median age of the men was about 70 years and median prostate-specific antigen (PSA) level was about 40 ng/mL. More than 80% of the men had bone metastases and about half had soft-tissue or nodal lesions.

The trial had 2 primary end points: overall survival and radiographically assessed progression-free survival (rPFS).

At the time of the second interim analysis (median follow-up of 22 months), abiraterone had produced a statistically significant improvement in rPFS, said Dr. Ryan.

Median rPFS was 8.3 months in the placebo group; it has not yet been reached in the abiraterone group. The difference resulted in a 57% decrease in the risk for progression (hazard ratio, 0.43; P < .0001).

However, Dr. Halabi explained that rPFS is not an established end point in metastatic prostate cancer. "This is the first time that rPFS has been used as a coprimary end point in this setting. We do not know if [rPFS superiority] translates into clinical benefit."

The use of rPFS was adapted from "published consensus material" and accommodated the possibility that, early on, patients might have "bone flares" in reaction to treatment that masquerade as new lesions, said Dr. Ryan.

To be considered rPFS, 1 of 3 events had to have transpired: death; soft tissue lesions (seen on computed tomography or magnetic resonance imaging); or progressive disease seen on bone scan (at less than 12 weeks after randomization, at least 2 new bone lesions plus 2 more at the next scan as confirmation; at 12 weeks or more, just 2 newly detected lesions).

In terms of overall survival, Dr. Ryan reported that there was a "strong trend" in favor of abiraterone.

Median overall survival was 27.2 months in the placebo group; it has not yet been reached in the abiraterone group. This represents a 25% decrease in the risk for death in the abiraterone group (hazard ratio, 0.75; P < .0097), Dr. Ryan said. However, the data on overall survival were not statistically significant.

Dr. Halabi said that the 25% reduction in the risk for the hazard was a best-case scenario statistically, and suggested that it would very likely be less. "It is an overestimate of the clinical benefit as measured by the hazard ratio [for mortality]," she noted.

The drug is active.

Dr. Halabi does not doubt that abiraterone is effective. "The drug is active," she said. She also believes that there would have been a significant impact on overall survival if the data had been given a chance to mature. "If the Data and Safety Monitoring Board had waited, there is a high probability that we would have seen statistically significant survival results," she explained.

But now there is doubt, she said.

At the time of the second interim analysis, 333 patients in the study had died, which is less than half the 773 total deaths planned in the trial design. "These 333 patients may have been totally different from the other patients who will die," she explained.

Secondary End Points Show Value of Treatment

Dr. Ryan reported that there was a statistically significant improvement in all secondary end points, which are "clinically important events."

Median Time to Secondary End Points (Months)

Secondary End Point Abiraterone Group Placebo Group Hazard Ratio P Value
Opiate Use Not reached 23.7 0.69 .0001
Chemotherapy 25.2 16.8 0.58 <.0001
ECOG Performance Decline 12.3 10.9 0.82 .0053
PSA Progress 11.1 5.6 0.49 <.0001

These end points are the "hallmarks" of disease progression in mCRPC, said Dr. Ryan.

Dr. Halabi agrees that they translate into clinical benefit, with the exception of time to PSA progression, about which there is uncertainty.

PSA declined by more than 50% in 62% of men treated with abiraterone and in 24% of men treated with placebo (P < .0001).

These data "suggest" that "previous docetaxel is not required for abiraterone to produce clinical activity," said Dr. Ryan. In other words, the drug works in the absence of previous chemotherapy.

Dr. Ryan also indicated that treatment was discontinued in 69% of men in the abiraterone group and in 84% in the placebo group. However, very few patients discontinued because of adverse events. He added that there were no new important safety signals in this trial.

Subsequent treatment was "common," said Dr. Ryan, with 60% of the men in the placebo group and 44% in the abiraterone group seeking more therapy, most of which was docetaxel.

2012 Annual Meeting of the American Society of Clinical Oncology: Abstract LBA4518. Presented June 2, 2012.