Atezolizumab Plus Chemo Did Not Improve Survival in TNBC

Atezolizumab Plus Chemo Did Not Improve Survival in TNBC

Adding atezolizumab (Tecentriq) to chemotherapy did not significantly improve survival outcomes compared with chemotherapy alone in patients with triple-negative breast cancer (TNBC) whose disease relapsed within 12 months of last treatment with curative intent for early disease, according to data from the phase 3 Impassion132 trial (NCT03371017) presented during the 2024 ESMO Breast Cancer Congress.

At a median follow-up of 9.8 months, the median overall survival (OS) was 11.2 months with chemotherapy alone (n = 177) vs 12.1 months (95% CI, 10.1-15.1) with atezolizumab plus chemotherapy (n = 177) in patients with PD-L1–positive TNBC (HR, 0.93; 95% CI, 0.73-1.20; stratified log-rank P = .59).

Although not formally tested according to the hierarchical design, these results were consistent in the modified intention-to-treat (ITT) population, with a median OS of 9.8 months (95% CI, 8.4-12.0) with chemotherapy alone (n = 192) vs 10.4 months (95% CI, 8.9-12.9) with atezolizumab plus chemotherapy (n = 188; stratified HR, 0.94; 95% CI, 0.76-1.18).

“Novel therapies and trial designs are urgently required for this treatment-resistant population and these data highlight the importance of recognizing, as we have before, that TNBC is highly heterogeneous—especially in the first-line setting,” Rebecca A. Dent, MD, MSc, of the National Cancer Center Singapore and Duke-NUS Medical School, in Singapore, said in a presentation of the data. “We now recognize that PD-L1­–positive and ­negative patients have a very different prognosis. We’ve also realized that patients who harbor PIK3CA mutations also have a different prognosis, and we know that patients who have de novo metastatic TNBC also clearly have a different prognosis than those who have received neoadjuvant therapy and were exposed to a number of agents by the time they reach their first-line status. As we are here discussing novel trial designs for early-relapsing TNBC, I really think we need to take this into account.”

About half of patients who develop metastatic TNBC after standard neoadjuvant or adjuvant chemotherapy will relapse within 1 year of chemotherapy completion. Early-relapsing TNBC represents a biologically and clinically distinct entity in that it is intrinsically resistant to standard options and is more prevalent in younger patients who have large primary tumors that do not harbor BRCA alterations, Dent explained. Notably, the majority of clinical trials do not include those patients, which “poses a real challenge for us in clinical practice,” she underscored.

The double-blind, placebo-controlled, randomized phase 3 IMpassion132 trial enrolled patients with unresectable locally advanced or metastatic TNBC who previously received an anthracycline and taxane for early disease and experienced progressive disease (PD) within 1 year after their last curative-intent treatment. They could not have previously received chemotherapy for advanced disease, and their PD-L1 status needed to be available.

Participants were randomly assigned 1:1 to investigator-selected chemotherapy in the form of gemcitabine at 1000 mg/m2 plus carboplatin area under the curve 2 mg/mL/min on days 1 and 8 every 21 days, or capecitabine at 1000 mg/m2 twice daily on days 1 to 14 every 21 days with or without atezolizumab at 1200 mg once every 3 weeks. Treatment continued until PD or intolerable toxicity.

Patients were stratified by visceral metastases (lung and/or liver), CT backbone, and PD-L1 status during all-comer enrollment.

OS served as the trial’s primary end point. The trial design called for hierarchical testing: those patients who were PD-L1­ positive were evaluated, and if positive, modified for the ITT population.

“This trial started enrolling in January 2018; this was an all-comer enrollment that as I mentioned, were stratified by PD-L1 status using the SP142 assay. In August of 2019, we saw the results of the Impassion130 study [NCT02425891], which clearly showed us that the benefits of immune checkpoint inhibition were largely driven by those patients who had PD-L1–positive metastatic TNBC, which is similar to what we’ve observed in the KEYNOTE-355 study [NCT02819518] using pembrolizumab [Keytruda] but using the [PD-L1 IHC] 22C3 assay. So, at this point, enrollment was restricted to just further enroll those patients who were PD-L1 positive, as measured by the SP142 assay, and this was completed in August 2023.”

The primary analysis was planned when a prespecified number of 247 deaths occurred in the PD-L1–positive subgroup, with a target HR of 0.70. If positive, investigators would then look at the modified ITT population, which included patients irrespective of PD-L1 status.

The median patient age was 48 years in the chemotherapy-alone and atezolizumab arms, respectively. Slightly more than half of patients had an ECOG performance status of 0 (57% vs 62%) and 18% in both arms had prior exposure to platinum-based chemotherapy; 27% and 29% of patients, respectively previously received capecitabine. More than half of patients had a disease-free interval (DFS) of less than 6 months (69% vs 66%) and lung and/or liver metastases (62% vs 60%). Investigator-selected chemotherapy was carboplatin/gemcitabine for 73% and capecitabine for 27% of those in both arms.

“When looking at the forest plots, looking at the subgroups of patients with PD-L1–positive TNBC, I want to draw your attention to a couple of things. First of all, looking at the chemotherapy backbone, this is hypothesis generating, but it is interesting to note in the combination of carboplatin/gemcitabine in the placebo arm, we see a median OS of 9.9 months, which increased to 12.6 months in those patients who also received atezolizumab, a hazard ratio of 0.81, but this didn’t cross 1,” Dent said. “The other thing I wanted to highlight was the DFI. What you can see is in those patients, and remember, this is about three-quarters of the patients who had a DFI of less than 6 months, the median OS was only 9.4 months in the control arm and 11.3 months in the atezolizumab arm. But you can see in both of these subgroups, whether they relapsed less than 6 months or greater than 6 months, there was no statistically significant improvement in OS with the addition of atezolizumab.”

The median progression-free survival in the placebo arm was 3.6 months (95% CI, 3.4-4.2) vs 4.2 months (95% CI, 3.7-5.6) in the atezolizumab arm (stratified HR, 0.84; 95% CI, 0.67-1.06). The unconfirmed objective response rates in the respective arms were 28% (95% CI, 21%-36%) and 40% (95% CI, 32%-48%), respectively, translating to a percentage difference of 11%. The median duration of response in the placebo arm was 4.1 months (95% CI, 3.5-5.8) vs 6.6 months (95% CI, 4.6-8.0) in the atezolizumab arm (HR, 0.73; 95% CI, 0.48-1.11). Data for these secondary end points in the modified ITT population proved to be consistent, according to Dent.

Safety was evaluated in a total of 587 patients (placebo arm, n = 294; atezolizumab arm, n = 293). Any-grade adverse effects (AEs) occurred in 96% of those in the placebo arm and 96% in the atezolizumab arm; these effects were treatment related for 91% and 90% of patients, respectively. Treatment-related grade 3 or 4 AEs occurred in 71% and 67% of patients, respectively, and were grade 3 or 4 for 65% and 62% of patients, respectively. Two patients in the placebo arm experienced a treatment-related grade 5 AE; 4 patients in the atezolizumab arm experienced grade 5 AEs, only 1 of which was determined to be treatment related. AEs of special interest occurred in 54% and 61% of patients, respectively. AEs led to treatment discontinuation for 11% of those in the placebo arm and 15% of those in the atezolizumab arm.

“No new safety signals were observed,” Dent concluded.

Reference

Dent R, André F, Gonçalves A, et al. Impassion132 double-blind randomised phase III trial of chemotherapy with or without atezolizumab for early relapsing unresectable locally advanced or metastatic triple-negative breast cancer. Ann Oncol. Published online May 15, 2024. doi:10.1016/j.annonc.2024.04.001