Eric J. Small, MD, of the University of California at San Francisco, discussed what he thought went wrong for cabozantinib (Cometriq) in the ­COMET-1 trial, yet another example of a phase III trial that failed to deliver on the promise of phase II results.

“Although this was a negative study, cabozantinib is an active agent. The activity observed was in intermediate endpoints, however, and not in the primary endpoint of the study, survival. In my opinion, the toxicity of the drug may have contributed to the negative results. Although overall negative, the greater impact on survival of cabozantinib in men with visceral metastases bears further study,” Dr. Small said.

Furthermore, Dr. Small found the improved survival in patients with visceral metastasis of interest. “Here is an agent that might be suitable for these patients, who typically have a worse response. Developing predictive and prognostic markers in this group of patients is a priority,” he stated.

“Most of us were astonished by the striking improvement in bone scans on cabozantinib in the phase II trials,” Dr. Small said. “But red flags were raised by the early data.”

Considerable Toxicity

“Cabozantinib came on the scene in the setting of extraordinary responses, but with considerable toxicity. Dose reductions were needed in 70% of patients in the phase II experience,” he continued.

At that time, 40 mg was determined as the lowest effective dose. A multicenter nonrandomized trial compared 40 mg with 100 mg, and the 100-mg dose was deemed too toxic, since 84% of patients required dose reductions.

“COMET-1 dosed cabozantinib at 60 mg, thereby allowing patients to undergo a dose reduction if needed and still be dosed at the lowest effective dose. About 70% of patients had a grade 3 or higher adverse event, and one-third discontinued treatment due to toxicity,” he said.

The fact that more patients had dose reductions of cabozantinib on 60 mg in COMET-1 than on 40 mg in the phase II trials (60% vs 31%, respectively) argues that there may be a toxicity threshold just above 40 mg, Dr. Small suggested. 

Prior experience with cabozantinib as well as cabazitaxel (Jevtana) may be an additional factor influencing the phase III results. “The phase II trials were conducted at 14 sites experienced in giving cabozantinib, whereas this was not true in the phase III trial. It is also possible that prior use of cabazitaxel influenced results, because 24% of the phase II patients vs 38% of the ­COMET-1 patients were treated with prior cabazitaxel,” he elaborated. ■

Disclosure: Dr. Small reported no potential conflicts of interest.