SAN ANTONIO – A variety of biomarkers failed to improve on the results of tumor HER2 status for predicting the benefits of therapy targeting HER2 signaling, based on an analysis of data from the CLEOPATRA trial.
In the trial, 808 patients with metastatic HER2-positive disease were given trastuzumab (Herceptin) and docetaxel (Taxotere) and evenly randomized to either pertuzumab (Perjeta) or placebo.
Dr. Jose Baselga
The main results, previously reported, showed that pertuzumab was significantly superior to placebo in terms of progression-free survival and overall survival.
In the new analysis, reported at the San Antonio Breast Cancer Symposium, a variety of protocol-predefined biomarkers of HER2 signaling in tumor tissue and in serum did not identify any patient subgroups that did better or worse on pertuzumab.
"Our analysis confirms that HER2 is the only marker for selecting patients for HER2-targeted therapy. This was despite a comprehensive exploration of a broad panel of candidate biomarkers," commented lead investigator Dr. Jose Baselga of the Memorial Sloan-Kettering Cancer Center in New York. "It is also consistent with the TRYPHAENA and NeoSphere studies, also presented here in San Antonio."
Discussing study limitations, Dr. Baselga said that "the lack of a HER2-naive treatment control arm may have resulted in the absence of a signal for other biomarkers in CLEOPATRA."
In additional findings, mutation of the gene for the catalytic subunit alpha of PI3 kinase (PIK3CA) in the tumor was associated with a poorer prognosis in the trial population as a whole. Patients whose tumors had the alternate, wildtype form of the gene were about one-fourth less likely to experience cancer progression or death.
"Mutations in PI3 kinase were not associated with resistance to pertuzumab, as patients derived similar additional benefit independent of PI3 kinase mutational status. However, the PI3 kinase mutational status may identify patients with poorer prognosis and particular unmet medical needs," Dr. Baselga commented. Prior studies have similarly shown mutant PI3 kinase to be associated with resistance to lapatinib (Tykerb), another HER2-targeted therapy and with a poorer prognosis after trastuzumab therapy.
"Clinical trials of HER2-targeted molecules in combination with PI3 kinase pathway–targeted agents may therefore be justified based on our findings," he added.
It would be interesting to see how a subset of untreated HER2-positive tumors behave as a function of PI3 kinase mutational status, Dr. Baselga said. "I think that will be what we need to know."
In the new analysis, the investigators evaluated a dozen biomarkers lying on the HER2 signaling pathway for both predictive and prognostic value. "From the cell surface down, we analyzed ligands, receptor tyrosine kinases, soluble HER2, and key intracellular pathway markers," Dr. Baselga explained. For most, they defined high and low levels as above- and below-median values, respectively.
In analyses of the predictive value of the biomarkers, the improvement in progression-free survival with pertuzumab relative to placebo was similar for patients having high versus low levels of all the biomarkers studied.
But in analyses of the prognostic value of the biomarkers in both trial arms pooled, patients had better progression-free survival if they had low serum levels of soluble HER2 (hazard ratio, 1.23; P = .04); in terms of ligands and receptor tyrosine kinases, if they had high levels of HER2 mRNA (0.77; P = .008) and HER3 mRNA (0.81; P = .03), and a high HER2 modified histoscore (0.83; P = .05); and in terms of tumor intracellular pathway markers, if they had mutation of the PIK3CA gene (0.83 for wild type; P = .0001).
The last "was by far our strongest prognostic marker," Dr. Baselga commented. Further analysis showed median progression-free survival was better with pertuzumab in both the mutant PIK3CA group (12.5 vs. 8.6 months) and in the wildtype PIK3CA group (21.8 vs. 13.8 months).
"Patients who harbor PI3 kinase mutations have clearly a worse prognosis. The treatment benefit, however, is maintained...the risk reduction is similar for mutants, for wild type, and for the whole study population," he noted.
In a final, longitudinal analysis to determine if serum biomarkers were early indicators of disease progression, levels of the biomarkers at baseline, week 9, and the time of disease progression showed no correlation with disease progression and did not differ between treatment arms.
Dr. Baselga disclosed that he is a consultant for Roche/Genentech and Sanofi. The trial was sponsored by Genentech.