Atezolizumab/Bevacizumab Demonstrates Durable Efficacy in Advanced Malignant Peritoneal Mesothelioma

The combination of atezolizumab (Tecentriq) plus bevacizumab (Avastin) elicited encouraging, durable responses in patients with advanced malignant peritoneal mesothelioma, irrespective of tumor mutational burden (TMB) and PD-L1 expression status, according to data from a phase 2 trial (NCT03074513) published in Cancer Discovery.¹

The doublet resulted in an objective response rate (ORR) of 40% (95% CI, 19.1%-64.0%; n = 8/20) per independent radiology review (IRR) and RECIST v1.1 criteria. Moreover, the median duration of response (DOR) with the regimen was 12.8 months. Notably, 75% of responses (n = 6) persisted for longer than 10 months.

Additionally, the progression-free survival (PFS) with the doublet was estimated to be 17.6 months (95% CI, 9.1–not reached), and the PFS rate achieved with the combination at 1 year was 61% (95% CI, 35%-80%). The overall survival (OS) had not yet been reached, but the 1-year OS rate was 85% (95% CI, 60%-95%).

“Atezolizumab/bevacizumab was well tolerated and led to robust and durable responses in patients with malignant peritoneal mesothelioma who had progressed on, or were intolerant to, prior platinum-pemetrexed chemotherapy with meaningful prolongation of survival,” lead study author Kanwal P.S. Raghav, MBBS, MD, of The University of Texas MD Anderson Cancer Center, and colleagues, wrote in the paper. “This study establishes a promising treatment option for our patients who suffer from this morbid cancer and represents an unprecedented effort to bridge the gap of dedicated research in this orphan disease.”

Many key studies that have examined immune checkpoint inhibitors in mesothelioma, such as CheckMate-743 or PROMISE-meso, were designed to examine the agents in those with malignant peritoneal disease; they excluded patients with malignant pleural disease. As such, most of the data acquired, and the subsequent approvals that have been granted, are specifically for patients with that malignant peritoneal disease.

However, findings from in vivo studies have indicated that the VEGF pathway is function in malignant peritoneal mesothelioma and inhibition of this pathway can lead to decreased proliferation and metastasis. As such, it was hypothesized that the combination of immune checkpoint inhibitors and antiangiogenic therapy could result in synergistic activity in patients with this disease.

In the phase 2 trial, investigators set out to evaluate the safety and efficacy of atezolizumab, a humanized monoclonal antibody that targets PD-L1, plus bevacizumab, a monoclonal antibody against VEGF-A, in previously treated patients with advanced malignant peritoneal mesothelioma, and to determine biomarkers of response to treatment.

To be eligible for enrollment, patients needed to be at least 18 years or older, have histologically confirmed advanced malignant peritoneal mesothelioma that was not amenable to definitive cytoreductive surgery (CRS), and they needed to have progressed on or have been intolerant to at least 1 line of systemic chemotherapy that involved a platinum-pemetrexed doublet. Patients were able to have previously received bevacizumab.

Moreover, patients were required to have measurable disease, an ECOG performance status of 0 or 1, and acceptable organ and bone marrow function. Patients who had extraperitoneal metastases were permitted.

If they had previously received immunotherapy, had an active diagnosis of autoimmune disease or immunodeficiency, had a concurrent malignancy, a known history of active or untreated central nervous system metastases, or were undergoing systemic immunosuppressive therapy at the time of enrollment, they were excluded.

In the open-label, single-stage, multicohort basket trial, investigators evaluated the doublet in several rare cancers, including malignant peritoneal mesothelioma. An individual analysis was planned for each 20-patient cohort.

In the malignant peritoneal mesothelioma cohort, atezolizumab was given at a fixed dose of 1200 mg in combination with intravenous bevacizumab, which was administered at a dose of 15 mg/kg every 21 days. Patients received treatment until they experienced progressive disease or intolerable toxicity.

The primary end point of the study was confirmed ORR per RECIST v1.1 criteria by IRR. Important secondary end points comprised safety, disease control rate (DCR), DOR, PFS, and OS. ORR, DCR, DOR, and PFS were also evaluated per immune-modified RECIST criteria. A key exploratory end point focused on evaluating tissue correlates for clinical activity.

A total of 20 patients were enrolled to the study between March 30, 2017, and February 12, 2019. All patients received treatment with the combination regimen. The median age of study participants was 63 years (range, 33-87), 60% were women, and 75% self-reported that they had not previously been exposed to asbestos. Ten percent of patients had biphasic histology and the remaining 90% had epithelioid disease.

Moreover, 60% of patients previously had CRS and hyperthermic intraoperative peritoneal perfusion with chemotherapy in addition to systemic chemotherapy. All participants previously received platinum-pemetrexed therapy, and 1 patient had prior bevacizumab. Forty percent of patients had previously received 2 or more lines of therapy prior to study enrollment. The median time from last systemic treatment to enrollment was 1.5 months. Additionally, 85% of patients had documented disease progression and 15% were intolerant to previous platinum-pemetrexed therapy.

At a median follow-up of 23.5 months (range, 10.1-36.1), patients had received a median of 15 cycles of treatment (range, 4-38) and 30% were still receiving treatment at the time of the data cutoff. Fifty percent of patients discontinued treatment because of progressive disease, 10% did so because of toxicity, 5% because of death, and 1% due to withdrawn consent.

Data from a post-hoc analysis indicated that the ORRs reported in several key subgroups that reflected patient, disease, and pre-study treatment factors, were similar. At 12 weeks, the DCR per RECIST v1.1 criteria was 95%; at 18 months, this rate was 85%.

No pseudo-progression was observed in this cohort of patients. At the time of the data cutoff, 50% of patients experienced progressive events and 7 patients had died. Most patients experienced progression within the peritoneal cavity, and 20% of patients experienced extraperitoneal progression.

An exploratory analysis was done to evaluate the impact of selection bias regarding indolent tumor biology in confounding outcomes by assessing patient and population dynamics of the cohort before trial enrollment.

Disease regression with chemotherapy was observed in 35% of patients, which proved to be comparable with response rates with this kind of combination in prior studies. The median time to next treatment on platinum-pemetrexed before enrollment was 8.3 months (95% CI, 6.3-10.3) vs 17.6 months with atezolizumab/bevacizumab. Notably, durable responses to the novel doublet were observed irrespective of response characteristics on previous chemotherapy.

Investigators also examined whether established biomarkers of response to immune checkpoint inhibition, like microsatellite instability, PD-L1 expression, and TMB, would be predictive of response to the doublet in this patient population. TMB status was determined in 70% of patients. Thirty-one percent of patients had PD-L1 expression of 0%, 46% had expression ranging from 1% to 50%, and 23% had expression of 50% or greater.

Notably, responses with the doublet were reported in patients with PD-L1 positivity and negativity, at 44.5% and 25.0%, respectively (odds ratio, 2.4; 95% CI, 0.2-38.0; P > .99). Moreover, the median PD-L1 expression was not found to differ between responders and those who did not respond.

Additionally, the median TMB for all participants was 0.8 mutations/megabase and the median TMB was comparable between responders and non-responders, at 0.8 mutations/megabase vs 0.8 mutations/megabase, respectively (P = .83). By using whole-exome sequencing, no link between response and specific gene alterations was reported.

Another analysis revealed that responders were observed to have lower epithelial-mesenchymal transition (EMT) gene-signature scores vs non-responders and this was found to correlate with magnitude of response to the doublet (spearman r, 0.8; 95% CI, 0.5-0.9; P = .0010). A high EMT gene expression was linked with worse ORR, at 14% vs 86% (odds ratio, 0.03; 95% CI, 0.0-0.6; P = .029).

Regarding safety, any-grade treatment-emergent adverse effects (TEAEs) were experienced by 85% of patients (n = 20); these effects were grade 3 in 50% of patients. The most frequently experienced TEAEs included hypertension (40%) and anemia (10%). Notably, no grade 4 or 5 toxicities were observed.

Additionally, 10% of patients experienced immune-related toxicities that were grade 3 in severity, and they included pancreatitis and thrombocytopenia. These patients were managed with corticosteroids, and they ended up discontinuing treatment with the regimen. Proteinuria was the only toxicity to result in dose interruption of bevacizumab, and this was experienced by 25% of patients following a median of 6 cycles of treatment.

Reference

1. Raghav K, Liu S, Overman MJ, et al. Efficacy, safety and biomarker analysis of combined PD-L1 (atezolizumab) and VEGF (bevacizumab) blockade in advanced malignant peritoneal mesothelioma. Cancer Discov. Published online July 14, 2021. doi:10.1158/2159-8290.CD-21-0331