Atezolizumab in Combination With Carboplatin and Survival Outcomes in Patients With Metastatic Triple-Negative Breast Cancer

Key Points Question Is the efficacy of carboplatin increased with atezolizumab in metastatic triple-negative breast cancer (TNBC), and are there clinical or molecular correlates associated with response? Findings In this randomized clinical trial including 106 patients, combining atezolizumab with carboplatin significantly improved progression-free and overall survival of patients with metastatic TNBC; tumor-infiltrating lymphocytes, mutation burden, obesity, and increased blood glucose levels were associated with response to anti–programmed cell death ligand 1 therapy. Meaning Adding atezolizumab to carboplatin provided a clinically meaningful benefit to patients with metastatic TNBC.

T riple-negative breast cancer (TNBC) is a heterogeneous disease characterized by increased genomic instability, immune infiltration, 1 and high programmed death ligand 1 (PD-L1) expression. 2 These features provided the rationale for early clinical studies evaluating the efficacy of anti-PD-L1 monoclonal antibody atezolizumab as monotherapy or in combination with nab-paclitaxel in advanced TNBC. 3,4The durable responses, particularly in PD-L1-positive tumors, led to the phase 3 IMpassion130 trial and accelerated US Food and Drug Administration (FDA) approval for PD-L1-positive metastatic TNBC. 3 A subsequent phase 3 trial evaluated atezolizumab in combination with paclitaxel; however, this combination did not improve progression-free survival (PFS) or overall survival (OS) vs paclitaxel alone in PD-L1-positive metastatic TNBC. 5 The reasons for this are unclear and suggest that anti-PD-L1 therapy efficacy may differ when combined with different chemotherapy.Anti-PD-L1 therapy was first approved in combination with taxane chemotherapy. 6Unlike taxanes, platinum agents are DNA intercalating agents and generate neoantigens that may stimulate an immune response. 7Therefore, the study reported herein, TBCRC 043 was designed to prospectively evaluate the efficacy of adding the anti-PD-L1 antibody atezolizumab to carboplatin therapy.
TNBC is a heterogeneous disease with varying immune cell composition and may have differential sensitivity to immune checkpoint therapies. 8,9Although several biomarkers, such as tumor mutation burden (TMB) 10,11 and tumor-infiltrating lymphocytes (TILs), 12 are associated with better response, there is a need to identify patients who are unlikely to respond to immunotherapy and spare them from severe immunerelated toxic effects. 13We thus present clinical trial results as well as molecular correlates of response to atezolizumab in combination with carboplatin.

Study Design and Participants
TBCRC 043 is a prospective, multicenter, randomized, doubleblind phase 2 trial (NCT03206203) investigating the efficacy of atezolizumab in combination with carboplatin vs carboplatin alone in patients with metastatic TNBC.Eligible patients had either clinical stage IV or metastatic invasive breast cancer negative for estrogen receptor (<10%), progesterone receptor (<10%), and ERBB2 (immunohistochemical analysis, ≥0/1 or fluorescence in situ hybridization, <2.0).Patients with an Eastern Cooperative Oncology Group performance status of 0 to 1, measurable disease, adequate hematologic, kidney, hepatic, and cardiac function, 0 to 1 prior treatments for metastatic disease, and no prior carboplatin in the metastatic setting or prior immune-oncology treatment were eligible.Patients were not stratified by PD-L1 status.
The protocol (Supplement 1) was approved by ethical and institutional review boards (IRB#160633) at the participating institutions, and all patients provided written informed consent and did not receive financial compensation.Patients were screened and enrolled at participating centers of the Translational Breast Cancer Research Consortium (TBCRC).Data were collected and reviewed by the Clinical Trials Office and Data and Safety Monitoring Committee.

Treatments and End Points
Patients were randomized to receive intravenous carboplatin area under the curve (AUC) 6 alone or in combination with atezolizumab, 1200 mg, every 3 weeks until disease progression, unacceptable toxic effects, or withdrawal of consent.On disease progression (clinically/Response Evaluation Criteria in Solid Tumors), patients on the carboplatin-alone arm were allowed to cross over to receive atezolizumab alone after undergoing a metastatic biopsy if reasonably safe (eFigure 1 in Supplement 2).
The primary end point was PFS, the time from the first day of treatment to disease progression or death.Secondary end points were overall response rate (ORR; CR + PR) and clinical benefit rate (CBR; CR + PR + stable disease ≥6 months), duration of response (DOR), and OS.Adverse events were graded using the National Cancer Institute Common Terminology Criteria for Adverse Events, V5.0.

Statistical Analysis
The primary PFS end point was powered for a 1-sided test of the hazard ratio (HR) for PFS at α = .10.The Kaplan-Meier method was used to estimate time-to-event curves and medians for PFS, OS, and DOR.Stratified log-rank tests were used to determine significance.Cox proportional hazards regression models were used to estimate HRs and 95% CIs.Data cutoff was October 2021.Additional details of the statistical analysis are given in the eMethods in Supplement 2. The analysis was performed using R statistical software (version 4.3.0,R Project for Statistical Computing) and followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guidelines.

Key Points
Question Is the efficacy of carboplatin increased with atezolizumab in metastatic triple-negative breast cancer (TNBC), and are there clinical or molecular correlates associated with response?
Findings In this randomized clinical trial including 106 patients, combining atezolizumab with carboplatin significantly improved progression-free and overall survival of patients with metastatic TNBC; tumor-infiltrating lymphocytes, mutation burden, obesity, and increased blood glucose levels were associated with response to anti-programmed cell death ligand 1 therapy.
Meaning Adding atezolizumab to carboplatin provided a clinically meaningful benefit to patients with metastatic TNBC.
tients were excluded, and 106 were randomized to receive atezolizumab plus carboplatin (n = 56) or carboplatin (n = 50) (Figure 1).Five patients receiving carboplatin and 2 on the combination arm were removed from the study for reasons other than disease progression.

Safety and Tolerability
The median (range) duration of treatment for patients receiving atezolizumab plus carboplatin was 17.4 (1.4-90.3)weeks, whereas for single-agent carboplatin it was 15.4 (3.0-72.1)weeks.The combination was generally well tolerated, and toxic effects were consistent with previous reports. 5,6The most common adverse events (AEs, >1) on the combination arm were thrombocytopenia, anemia, lymphocytopenia, nausea, fatigue, and increased liver enzymes (eTable 2 in Supplement 2).Compared with carboplatin, atezolizumab plus carboplatin was associated with a higher incidence of grade 3/4 serious AEs (41% vs 8%).Grade 3 immune-related AEs occurred in 10 patients (5%) receiving the combination, and among possible immune-related AEs were individual cases of colitis, hypothyroidism, retinopathy, and infective myositis (eTable 2 in Supplement 2).Even with increased serious AEs, the withdrawal of study drugs was similar between arms, with 3 patients (6%) receiving carboplatin and 2 patients (4%) receiving the combination coming off study for toxic effects.Crossover patients receiving atezolizumab monotherapy had lower serious AEs (21%) than the combination arm.

Molecular Correlates
This trial was designed for extensive exploratory biomarker analysis to identify patient subgroups with response to atezolizumab that included: (1) tumor and immune PD-L1 expression, (2) nonsynonymous mutation burden rate, (3) TNBC molecular subtype, and (4) lymphocytes composition by multichannel immunofluorescence (eMethods in Supplement 2).
To identify differentially expressed transcripts and evaluate TNBC molecular subtypes associated with response, we performed RNA sequencing on pretreatment tumor biopsies.After adjusting for metastatic tissue site and sample type (fresh frozen vs formalin-fixed paraffin-embedded), we identified 82 differentially expressed transcripts between patients responding (partial response or stable disease) to carboplatin plus atezolizumab and those with progressive disease (eFigure 6A-6C, eTable 5 in Supplement 3).Gene ontology analysis of transcripts elevated in responders showed enrichment in insulin-like growth factor (IGF) transport and uptake, and lipid  pathways (eTable 5 in Supplement 3).IGF2, a growth factor capable of binding the insulin receptor and implicated in diabetes, 16 was among the highest transcripts expressed in responding patient tumors (eFigure 6D in Supplement 2).TNBC subtyping was performed on RNA sequencing from 102 pretreatment biopsies, which resulted in 25% basal-like 1 (BL1), 26% mesenchymal (M), 33% basal-like 2 (BL2), and 15% luminal androgen receptor (LAR) (eFigure 6E in Supplement 2).This distribution in metastatic disease differs from primary TNBC in the Cancer Genome Atlas (35% BL1, 30% M, 20% BL2, and 15% LAR) and likely reflects the reduction of the chemosensitive BL1 subtype from prior chemotherapy treatments (eFigure 6F in Supplement 2). 17,18All subtypes tended to benefit more with the addition of atezolizumab, except for the LAR subtype (Figure 4; eFigure 6G in Supplement 2).
Because variants generate neoantigens, we performed whole genome DNA sequencing on pretreatment tumor tis-sues to identify somatic variants.The median nonsynonymous mutation rate ranged from 9.1-371.0mutations/ megabase (MB) with a median (SD) of 15.7 (40.1) mut/Mb (eFigure 7A in Supplement 2).There was no association between TMB and PFS for patients receiving carboplatin (log-rank P = .63);however, patients with high TMB (>25 mut/ Mb) who received atezolizumab and carboplatin had better PFS (log-rank P = .05)(supporting data in eFigure 7B and 7C in Supplement 2).

Discussion
In this study, adding atezolizumab to carboplatin therapy increased PFS from 2.2 to 4.1 months.This increase is similar to adding pembrolizumab (5.6 to 7.5 months) to a gemcitabinecarboplatin therapy. 23The combination of atezolizumab and carboplatin also increased OS in IIT patients with metastatic disease from 8.6 to 12.6 months.
Patients receiving atezolizumab monotherapy after disease progression on carboplatin had similar OS with fewer toxic effects than those receiving the combination, suggesting sequential chemotherapy and immunotherapy may be another option for those patients in which clinical toxic effects management is critical.However, additional studies are needed for validation because this study was not powered for that comparison and treatment crossover may inherently select for patients with aggressive disease.
It is unclear which biomarkers are useful for predicting immune checkpoint therapy response.Atezolizumab was initially approved for PD-L1-positive metastatic TNBC, 6 and pembrolizumab showed benefit in patients with PD-L1-positive disease. 24However, benefit of atezolizumab added to paclitaxel was not observed in patients with PD-L1-positive disease in the IMpassion131 study. 5In this study, PD-L1 status did not affect PFS or OS with immunotherapy.However, the incidence of PD-L1-positive tumors (22%) was substantially lower than the 41%, 45%, and 38% reported in the IMpassion130, 6 IMpassion131, 5 and KEYNOTE-355 clinical trials, 24 respectively.Lower PD-L1 positivity may be attributed to the proportion of biopsies evaluated from primary vs metastatic sites.In this study, most (75%) were metastatic biopsies compared with 48% in the IMpassion131 trial. 5Furthermore, the PD-L1 positivity rates vary by metastatic biopsy location, with liver, skin, and bone metastases displaying much lower rates than other sites. 25egardless, there was a significant increase in PFS and OS in PD-L1-negative tumors consistent with the KEYNOTE-522 trial, in which pembrolizumab provided a benefit regardless of PD-L1 status. 26ike other studies, we observed increased survival benefit for patients receiving the combination whose tumors had higher pretreatment TILs.Multi-immunofluorescence analyses demonstrated that although CD8 T cells were associated with shorter PFS on both arms, activated (GZMB + ) T cells were associated with improved PFS only on the combination arm.Furthermore, CD4 T cells were associated with increased PFS, while FOXP3-positive regulatory T cells were associated with decreased PFS for patients receiving the combination.
Mutation burden has been associated with immunotherapy response retrospectively 27 and prospectively. 26We observed a similar association in patients with TMB-H tumors, having increased PFS on both arms.Activating PIK3CA mutations were the only variant associated with a lack of response to immunotherapy.However, PIK3CA variants were also associated with decreased clinical benefit with carboplatin alone, consistent with a prior study 28 in which patients with PIK3CA variants had a lower response to anthracycline-based neoadjuvant chemotherapy.
In this study, we observed an association of obesity and uncontrolled blood glucose levels with better responses to PD-L1 immunotherapy.Obesity has been associated with increased efficacy of PD-1/PD-L1 blockade and attributed to T-cell dysfunction, 29 and retrospective studies have shown that patients with obesity with breast cancer achieved the highest benefit. 30This benefit could be attributed to higher adipose tis-   sue composition in the breast and augmented by metabolic syndrome conditions such as type 2 diabetes.Changes in blood glucose levels in fasting-mimicking diets profoundly affect peripheral blood cell composition by reducing immunosuppressive myeloid and regulatory T-cell compartments. 31Patients with with obesity with elevated blood glucose levels may have more immunosuppressive immune cells upon which anti-PD1/PD-L1 therapies can act.Future studies are needed to validate these findings and delineate the effects of blood glucose and obesity.

Limitations
Although preliminary associations were observed in the molecular subgroup analyses, caution should be exercised in interpretation because these were retrospective, exploratory, and limited by small patient numbers and require further investigation in larger cohorts to draw definitive conclusions.

Conclusions
In this multicenter phase 2 randomized clinical trial, adding atezolizumab to carboplatin significantly increased PFS and OS in patients with metastatic TNBC.Increased TILs, higher TMB, obesity, and uncontrolled blood glucose levels were associated with a decreased risk of progression, whereas tumor HLA-A expression, PIK3CA variants, and the LAR subtype were associated with a greater risk of progression for patients receiving the combination.Crossover patients receiving sequential chemotherapy and immunotherapy had fewer toxic effects than patients receiving the combination, suggesting sequential treatment could be considered for patients whose treatment-related toxic effects are of concern.

Figure 3 .
Figure 3. Subgroup Analysis of Progression-Free Survival (PFS) in Patients Treated With Carboplatin or Carboplatin Plus Atezolizumab

Figure 4 .
Figure 4. Subgroup Analysis of Progression of Patients Between Treatment Arms Figure 5. Subgroup Analysis of Molecular Correlates Associated With Progression-Free Survival Forest shows the hazard ratio (HR) for the progression of patients between variable cutoffs on each arm stratified by TNBC type, DNA mutation, and immunofluorescence markers.BL1 indicates basal-like 1; BL2, basal-like 2; M, mesenchymal; LAR, luminal androgen receptor; TMB, tumor mutation burden; CPS, combined positive score.Atezolizumab in Combination With Carboplatin In Metastatic TNBC Original Investigation Research jamaoncology.com(Reprinted) JAMA Oncology Published online December 14, 2023 E7 Downloaded from jamanetwork.comby guest on 12/29/2023