Mixed Survival Data for First-Line Combo in Advanced HCC

Cabozantinib (Cabometyx) plus atezolizumab (Tecentriq) significantly improved progression-free survival (PFS) versus sorafenib (Nexavar) alone for patients with previously untreated advanced hepatocellular carcinoma (HCC), an interim analysis of a phase III study showed.

The COSMIC-312 trial met one of its primary endpoints, with the first-line tyrosine kinase inhibitor (TKI) plus PD-L1 checkpoint inhibitor combination reducing the risk of disease progression or death by 37% compared with single-agent sorafenib (median PFS 6.8 vs 4.2 months, respectively; HR 0.63, 99% CI 0.44-0.91, P=0.0012), reported Robin Kate Kelley, MD, of the University of California San Francisco.

However, an interim overall survival (OS) analysis, the trial’s second primary endpoint, showed no significant difference between the groups treated with atezolizumab-cabozantinib or sorafenib (median 15.4 vs 15.5 months, respectively; HR 0.90, 96% CI 0.69-1.18, P=0.438), according to a European Society for Medical Oncology (ESMO) virtual plenary presentation at ESMO Asia Virtual Oncology Week 2021.

A third arm of the study showed a numerical PFS advantage for the TKI cabozantinib alone compared with sorafenib (5.8 vs 4.3 months; HR 0.70, 99% CI 0.51-1.01), but “did not meet the stringent P-value required for significance in the interim,” Kelley said.

ESMO discussant Ian Chau, MD, of the Institute of Cancer Research and Royal Marsden Hospital in England, noted that current American Society of Clinical Oncology (ASCO) guidelines recommend offering the combination of atezolizumab and bevacizumab (Avastin) as first-line treatment for most patients with advanced HCC, based on results from the IMbrave150 trial, which compared that combination with sorafenib.

“How do we choose between atezolizumab and bevacizumab versus atezolizumab and cabozantinib, both of which have shown a highly significant improvement in PFS so far?” Chau said.

While updated results from IMbrave150 have shown a continued OS improvement with atezolizumab-bevacizumab, “we have not yet observed an overall survival [benefit] with atezolizumab and cabozantinib in the interim analysis,” he said, adding that response rates were also higher with the atezolizumab-bevacizumab combination.

In the current study, objective response rates were 11% for cabozantinib-atezolizumab, 3.7% for sorafenib, and 6.4% for cabozantinib alone.

“If you go with the ESMO-Magnitude of Clinical Benefit Scale, the [atezolizumab-cabozantinib combination] is currently at 3,” Chau said. “So, it is not deemed to be a highly significant intervention yet, but we of course are waiting for the final OS results.”

Cabozantinib received FDA approval in 2019 for patients with HCC who have been previously treated with sorafenib. In 2020, atezolizumab-bevacizumab got the agency’s green light for use in patients with treatment-naive unresectable or metastatic HCC.

COSMIC-312 included 837 patients at 281 centers globally. Patients were randomized 2:1:1 to receive cabozantinib-atezolizumab (n=432), sorafenib alone (n=217), or single-agent cabozantinib (n=188).

Kelley shared results for two primary endpoints in the interim analysis: PFS as assessed in the first 372 patients randomized to cabozantinib-atezolizumab or sorafenib, known as the PFS intention-to-treat (ITT) population (median follow-up of 15.8 months), and OS in the full-study ITT population (median follow-up of 13.6 months). PFS in the monotherapy comparison between cabozantinib and sorafenib was a secondary endpoint.

Disease control rates (response plus stable disease) were 78% for cabozantinib-atezolizumab, 65% for sorafenib, and 84% for single-agent cabozantinib.

“These comparisons were included to better characterize the contribution of cabozantinib in this combination arm, given the lack of single-agent cabozantinib data in the first-line HCC setting,” Chau pointed out.

As for safety, Kelley reported that adverse events (AEs) with cabozantinib-atezolizumab were manageable, with a safety profile that was consistent with that of each single agent.

The most common grade 3 or higher AEs for the combination were palmar-plantar erythrodysesthesia (7.9% vs 8.2% and 8.5% for the sorafenib and cabozantinib monotherapies, respectively), hypertension (7.0%, 6.3%, 11.0%), increased aspartate aminotransferase (6.5%, 2.4%, 5.3%), and increased alanine aminotransferase (6.3%, 1.9%, 5.9%).

Rates of grade 3/4 treatment-related AEs were 51% for cabozantinib-atezolizumab, 30% for sorafenib, and 52% for cabozantinib monotherapy. Treatment-related deaths occurred in 1.9% of those on the cabozantinib-atezolizumab arm, and in 0.5% of those on each of the monotherapy arms.