NEW YORK – An analysis of archival tissue and fresh biopsies from patients with advanced hepatocellular carcinoma (HCC) treated with a combination of the immune checkpoint inhibitor atezolizumab (Genentech's Tecentriq) and the VEGF inhibitor bevacizumab (Genentech's Avastin) has shed light on the genomic markers associated with response to the combined drugs.
The analysis comes as the US Food and Drug Administration is currently reviewing a new drug application for the therapy combination, the benefit of which was shown in the Phase III IMbrave150 study comparing the combination to the current standard-of-care, sorafenib (Bayer's Nexavar). Genentech's parent firm Roche, however, is seeking approval for the combination in all previously untreated HCC patients with unresectable tumors and not in a biomarker-defined indication.
This exploratory analysis, presented during the American Association for Cancer Research's Virtual Annual Meeting this week, illuminated several genomic correlations that may be useful in predicting response to the atezolizumab/bevacizumab combination. The analysis also validated the mechanisms by which these two agents synergistically work together — with atezolizumab reactivating the immune response to tumor cells and bevacizumab cutting off blood supply to tumors via angiogenesis and limiting immune suppression.
To conduct the analysis, presenter Andrew Zhu of Massachusetts General Hospital and his colleagues explored characteristics within two separate arms of the Phase Ib GO30140 study. Within Arm A, all patients received the atezolizumab/bevacizumab combination and the researchers sought to identify the molecular features associated with response or lack of response to the regimen. The researchers performed statistical analysis to determine the association between biomarker expression and clinical response, or progression-free survival, to the drug combination.
In Arm F, patients were randomized to either receive atezolizumab and bevacizumab, or atezolizumab alone. Here, the researchers looked into the molecular features associated with improved progression-free survival when patients were treated with the combination versus the checkpoint inhibitor monotherapy.
In Arm A, the researchers performed whole-exome sequencing for tumor mutational burden analysis and RNA sequencing for gene signature analysis. They determined that tumor mutational burden was not associated with progression-free survival with the bevacizumab plus atezolizumab combination. Patients with a strong pre-existing immune response as indicated by high expression of PD-L1 (CD274), as well as patients who had high T effector signature — that is high expressions of G2MB, PRF1, and CXL9 — were more likely to experience tumor shrinkage and have longer progression-free survival with the combination treatment.
On the other hand, expression of genes involved in NOTCH pathway activation, including high expression of HES1, was associated with a lack of response to the atezolizumab and bevacizumab combination and a shorter progression-free survival.
In a discussion following the presented results, Zev Wainberg of the University of California Los Angeles noted that these associations were consistent with findings in past studies. However, he added that the correlations identified in Arm A were somewhat limited in their usefulness because "there was no way to know whether it's the combination or single-agent checkpoint inhibition that is inducing these responses, and in fact, has the biomarker correlation."
The design of the Arm F portion of the study, meanwhile, randomized patients to just the immune checkpoint agent atezolizumab or the atezolizumab/bevacizumab combination and allowed researchers to identify biomarkers that would differentiate which patients would respond better to the combination versus just the immunotherapy. Through RNA sequencing, the study demonstrated that high VEGFR2 correlated with improved benefit with the combination. The same went for high T-regulatory (T-reg) signature, which was also associated with longer progression-free survival in the patients treated with atezolizumab and bevacizumab compared to those receiving just atezolizumab.
This suggests that high VEGFR2 and high T-reg signatures are both needed to identify patients in whom the combination works especially well, with bevacizumab promoting angiogenesis and overcoming immune suppression, and atezolizumab reactivating the immune system's response to tumor cells.
"This is consistent with prior studies showing that VEGF-A/VEGFR-2 pathway blockade may directly inhibit tumor-induced T-reg proliferation," Zhu said.
Zhu also noted that, after treatment with the atezolizumab and bevacizumab combination, decreased VEGFR-expression was observed in tumor biopsies. This decreased expression was noted across the board, regardless of treatment response or PFS, in patients treated with the combination. The decreased VEGFR expression was not seen in the biopsies of the tumors treated only with atezolizumab. T-reg signature expression, on the other hand, was decreased in both the combination and atezolizumab-alone groups.
This post-treatment biopsy analysis was another way of demonstrating that anti-VEGF treatment together with the immune checkpoint blockade treatment could work together to inhibit cytotoxic T cell proliferation.
Finally, Zhu and his colleagues explored the association between a myeloid inflammation signature and the combination of atezolizumab and bevacizumab. This signature was previously demonstrated to be predictive of response with the same combination of drugs used to treat renal cell carcinoma and was associated with improved progression-free survival. As the researchers expected, this signature was also predictive in HCC.
"What they were able to do in this study was borrow a previously validated gene signature set from renal cell carcinoma from a prior atezolizumab/ bevacizumab study, which showed that a myeloid-high subgroup had a particularly favorable effect with the combination of atezolizumab/bevacizumab over atezolizumab alone, and they validated that in that renal cell carcinoma core and then tested the same signature here in hepatocellular carcinoma," explained Wainberg in his discussion.
He further highlighted that the myeloid-low group did not experience any increased benefit with the combination over atezolizumab alone, suggesting that the signature could be used to identify patients who would derive no benefit from the combination treatment. Wainberg did point out the caveat, though, that there was no exploration of this signature's predictive association to just bevacizumab.
"We’re starting to look at more and more signatures to see how they correlate with clinical outcome [in HCC]," Wainberg said, concluding, "PD-L1 and VEGF combinations are here to stay in this disease, and this is certainly true not just with the combination of atezolizumab with bevacizumab, but also with the early data that's emerging with pembrolizumab [Merck's Keytruda] and lenvatinib (Eisai's Lenvima], amongst others."
Of course, Wainberg and Zhu alike pointed out that these findings need to be validated in larger data sets, since the arms of the Phase I GO30140 trial only consisted of about 90 to 100 patients each. Going forward, Wainberg said, if further research starts to suggest that VEGF inhibition via bevacizumab is mechanistically working with PD-L1 by inhibiting immune suppression via T regs, "one wonders whether the sequence of these drugs should be studied differently and if continuous PD-L1 VEGF inhibition is altogether necessary."