preheader BJMO

Adding atezolizumab to standard platinum-based chemotherapy plus bevacizumab does not improve PFS in patients with late relapsing ovarian cancer

At ESMO 2022, the updated results of ATALANTE were presented. This study evaluated the efficacy of adding atezolizumab to standard platinum-based chemotherapy plus bevacizumab in patients with late relapsing ovarian cancer (OC). Unfortunately, the addition did not result in improved progression-free survival, neither in the intention-to-treat nor the PD-L1 positive population. Although immature and non-significant, encouraging overall survival data however warrant further analysis with longer follow-up.

Ovarian cancer (OC) patients relapsing after a platinum-free interval (PFI) of more than six months benefit from platinum-based combination chemotherapy of carboplatin (Cb) with either pegylated liposomal doxorubicin (PLD), gemcitabine or paclitaxel. When added to chemotherapy, bevacizumab increases response rate and prolongs progression-free survival (PFS).  An objective response of 13.6% has been reported in patients with prior complete or partial response to platinum-based chemotherapy when treated with a single-agent checkpoint inhibitor (CPI) targeting either PD-1 or PD-L1. Importantly, the addition of CPI has been reported to increase chemotherapy and/or bevacizumab activity in other tumours. In this light, the ATALANTE trial evaluated the efficacy and safety of adding atezolizumab, a CPI targeting PD-L1,  to standard platinum-based chemotherapy in combination with bevacizumab in patients with late relapsing OC (PFI>6 months).

Study design

The randomised, double-blinded, phase III ATALANTE trial enrolled patients with relapsed non-mucinous epithelial OC, with a platinum-free interval of more than six months, one or two prior chemotherapy lines, and a good performance status (ECOG PS ≤ 1). They were stratified according to PD-L1 status with a threshold of 1% (≥ 1% vs. <1% vs. unknown), type of chemotherapy (Cb-PLD or gemcitabine or paclitaxel), and platinum-free interval of 6-12 vs. >12 months.  After initial biopsy, patients were randomised 2:1 to six cycles of carboplatin-based chemotherapy with either bevacizumab + placebo or bevacizumab + atezolizumab. During the study, an unexpected low rate of PD-L1 positive samples was found, so the protocol was amended in November 2018 to increase the number of patients from 405 to 614 to reach sufficient power.  Co-primary endpoints were progression-free survival (PFS1) in the intention-to-treat (ITT) and PD-L1 positive populations.  Secondary endpoints were time to second subsequent treatment (TSST), OS, safety and health-related quality of life (HRQoL).

Following a hierarchical approach, if either of the co-primary PFS1 comparisons was significant, OS could be analysed in both the ITT and PD-L1-positive populations.


Baseline characteristics were well-balanced between the two arms. In terms of pathological subtypes, serous high-grade ovarian cancer was the most frequent histology, as expected. Two third of the patients in both groups were previously treated with carboplatin PLD. The ATALANTE trial did not meet its co-primary endpoint. In the ITT population, median PFS was 13.5 vs. 11.3 months in the atezolizumab and placebo arms, respectively. A 17% reduction in the hazard ratio (HR) was found in favour of the experimental arm, but this was not significant (p= 0.041), as this was set to 0.025. In the PD-L1 positive population, PFS did not reach significance either (15.2 vs. 13.1 months; HR[95%CI]: 0.86[0.63-1.16; p= 0.30]).  No significant PFS difference with atezolizumab was found in any subgroup either, including patients with CD8+/PD-L1+ tumours on biopsy at study entry. With a HR of 0.82(0.67-1.01), no statistical differences in TSST were found between the two arms.  OS data are not yet mature but in the ITT population the median OS was in favour of atezolizumab (35.5 vs. 30.6 months, HR: 0.81[0.65-1.01]).  Overall safety data showed no unexpected signals from the addition of atezolizumab. Similar rates of grade 3/4 treatment-related adverse events (AEs) and AEs leading to treatment interruption were observed in both arms. More serious AEs (SAEs) were observed in the atezolizumab group, which is not surprising in the setting of a placebo-control arm.  Among all AEs, immune-related infusion reactions and auto-immune disorders (AESI) were considered of special interest. They were, as expected, more frequent in the atezolizumab group, but generally mild or moderate. Treatment-related grade 3/4 AESI occurred in only 6 and 2% in the atezolizumab and placebo groups. Among all the autoimmune disorders, hypothyroidism was the most prominent one, which is surprising considering the safety profile of immune-oncology drugs. HRQoL was assessed using the EORTC score.  No significant mean change from baseline EORTC QLQ-30 global score was observed in either arm across the study.


The addition of atezolizumab to standard platinum-based chemotherapy plus bevacizumab did not result in a PFS improvement in patients with late relapsing OC, neither in the ITT nor PD-L1 positive populations. A non-significant improvement in OS was observed in the atezolizumab vs. placebo groups, but these data are not mature, and longer follow-up is needed.  Safety data were consistent with the individual profile of each drug and their combination.

Further research on ATALANTE biopsy samples is required to better understand the immunological landscape of late relapsing OC.


Kurtz JE, et al. Phase III ATALANTE/ov29 trial: Atezolizumab (Atz) versus placebo with platinum-based chemotherapy (Cx) plus bevacizumab (bev) in patients (pts) with platinum-sensitive relapse (PSR) of epithelial ovarian cancer (OC). Presented at ESMO 2022; Abstract LBA30.

Speaker Jean Emmanual Kurtz

Jean Emmanual Kurtz

Jean Emmanual Kurtz, MD, PhD, Nouvel Hôpital Civil - Hôpitaux Universitaires de Strasbourg, Strasbourg, France


See: Keyslides

Back to Top