BARCELONA – Programmed death-ligand 1 (PD-L1) status in patients with advanced triple negative or HER2-positive breast cancer appears to identify distinct disease entities with varying likelihood of benefit from immune checkpoint inhibition, according to Giampaolo Bianchini, MD.
This observation, which contrasts with findings in other solid tumors and expands the road map to improved outcomes with immunotherapy for metastatic breast cancer, is based in part on new findings presented at the European Society for Medical Oncology Congress.
Among additional lessons from those findings: PD-L1 assays are not easily interchangeable, and studies with a “one size fits all” approach should be avoided,
IMPassion130 and PD-L1 assays
In the phase 3assessing nanoparticle, albumin-bound (nab)-paclitaxel chemotherapy + either the anti-PD-L1 monoclonal antibody atezolizumab or placebo for the first-line treatment of metastatic triple-negative breast cancer (mTNBC), investigators used, and validated, the to evaluate PD-L1 expression in immune cells (IC). PD-L1 positivity was defined using a 1% cutoff, meaning that PD-L1-stained IC encompassed at least 1% of the tumor area.
The trial demonstrated significantly improved progression-free survival (PFS) in the atezolizumab arm, both in the intention-to-treat (ITT) analysis (7.2 vs. 5.5 months in the placebo arm; hazard ratio, 0.80), and the PD-L1-positive subgroup (7.5 vs. 5.0 months; HR, 0.62), and the results were published in November 2018 (N Engl J Med. 2018; 379:2108-21).
“IMpassion130 is the first phase 3 trial demonstrating clinical benefit of cancer immunotherapy in patients with PD-L1-positive, metastatic triple-negative breast cancer,”, said at the congress. “The combination of atezolizumab and nab-paclitaxel is now approved in the United States and Europe for this indication.”
In addition, the SP142 antibody (which binds to PD-L1), at the 1% cutoff, predicted PFS and overall survival (OS) with atezolizumab + nab-paclitaxel, compared with nab-paclitaxel + placebo; the absolute improvement in OS in the PD-L1-positive population was 7 months (HR, 0.71), whereas no impact was seen in PFS or OS in patients who were PD-L1-negative using the SP142 assay, said Dr. Rugo, professor of hematology/oncology, and director of breast oncology and clinical trials education at the University of California, San Francisco.
Based on the IMPassion130 findings, the Food and Drug Administrationthe SP142 assay, using the 1% cutoff, as a “companion diagnostic device for selecting TNBC patients for atezolizumab.”
However, questions remain about how to best identify patients who could benefit from the atezolizumab + nab-paclitaxel combination, Dr. Rugo said.
Therefore, she and her colleagues performed a retrospective post hoc subgroup analysis of data from the trial to assess the performance and analytical concordance of the SP142 assay and two other commonly used PD-L1 immunohistochemistry (IHC) assays: the VENTANA SP263 IHC assay typically used as a companion diagnostic with durvalumab, and the Dako PD-L1 IHC 22C3 assay typically used with pembrolizumab.
In addition, the investigators assessed PD-L1 prevalence and clinical activity.
“We also included an evaluation of important factors related to PD-L1 testing and ... relationship to clinical outcome,” Dr. Rugo said.
In 614 biomarker-evaluable patients, representing 68% of the IMPassion130 ITT population, PD-L1-positive prevalence was 46% with the SP142 assay, 75% with the SP263 assay (also based on a 1% IC cutoff), and 81% with the 22C3 assay (with positivity defined as a combined positive score [CPS] of 1 or more based on an algorithm including both tumor and IC counts).
“Almost all SP142-positive cases are captured by either 22C3 or SP263. However, about a third of patients’ tumors were positive for PD-L1 using only one of the other two assays,” she noted, explaining that “this leads to suboptimal analytical concordance.”
The overall percentage agreement between SP142 and the other assays was only 64%-68%, she said.
Positive percentage agreement rates of 98% for both SP263 and 22C3 suggest that the patients identified as PD-L1 positive using the SP142 assay are captured by the other two assays. However, negative percentage agreement rates were less than 45%.
The HRs for PFS were 0.60 in SP142-positive patients, 0.64 in SP263-positive patients, and 0.68 in 22C3-positive patients, and the HRs for OS were 0.74, 0.75, and 0.78, respectively.
Subgroup analyses indicated that PFS and OS benefit with atezolizumab + nab-paclitaxel vs. nab-paclitaxel alone was greater in double-positive patients (those with SP142 positivity and either SP263 or 22C3 positivity) than in patients who were SP263-positive/SP142-negative or 22C3-positive/SP142-negative.
Dr. Rugo and her colleagues also found that the benefits with atezolizumab + nab-paclitaxel in PD-L1-positive patients were apparent regardless of the source of tissue for testing (breast or distant metastases).
They concluded that the findings of the assays are not equivalent; 22C3 and SP263 identified more patients as PD-L1 positive, and SP142-positivity was encompassed in positive tests for both.
“The clinical benefit in the 22C3-positive and the SP263-positive subgroups appear to be driven by the SP142-positive subgroup, and [SP142] identifies patients with the longest median progression-free and overall survival from the addition of atezolizumab to nab-paclitaxel,” she said “The SP142 assay with an IC cutoff of 1% or greater is the approved diagnostic test used to identify patients with metastatic triple-negative breast cancer who are most likely to benefit from the addition of the checkpoint inhibitor atezolizumab to nab-paclitaxel.”
As for whether the SP142 should be the assay of choice in other settings in which it hasn’t been validated, Dr. Rugo said it is advisable to use the assay that has been validated in a positive trial.
“That’s what we would generally do ... however, recognizing that some countries are not using SP142, and some sites may not have access, certainly you encompass that population in the patients whose tumors are positive by both other assays,” she said. “The risk is that you might overtreat, and the cost of treatment is greater.”
Excess toxicity is also a concern in that situation, she said, adding that “hopefully in the future we’ll be able to figure out ways to have even more patients benefit from the addition of immunotherapy so that won’t be an issue.”
“What this data shows is that you can feel secure that you are encompassing the patient population identified by the parent trial to benefit from the addition of atezolizumab by using either of the other two assays; you’re only missing 1% – so that’s very reasonable,” she said. “The risk is that you’re overtreating; it’s quite likely that there’s a population there that isn’t benefiting as much, but that’s a balance.”
The findings from IMPassion130 with regard to OS in the unselected population that included PD-L1-negative patients (18.7 vs. 21.0 months with vs. without atezolizumab; HR, 0.86) underscore the fact that “one size does not fit all” when it comes to immunotherapy benefit, Dr. Bianchini said.
This is further demonstrated by the post hoc analysis comparing IHC assays, he said, explaining that 63% of IMPassion130 patients who were considered PD-L1-negative based on the SP142 “actually scored as PD-L1-positive by the other tests.
“So the very clinically important question is if there is any evidence from the data that [the PD-L1-negative group] benefits in a significant way from the addition of atezolizumab,” he said. “I don’t see evidence for a clinical benefit, I see evidence to look for new biomarkers to identify a potential population who will benefit.”
The “absence of evidence is not evidence of absence,” he stressed, noting that it may be possible – with the right biomarkers – to identify PD-L1-negative patients who would benefit.
What the findings do show, however, is support for the FDA decision to approve the SP142 assay with an IC cutoff of 1% as a companion diagnostic tool, and that PD-L1 is ideally assessed using samples from both the primary and metastatic site, as the IMPassion130 data “do not inform whether PD-L1 assessment in primary and metastatic sites is equally informative,” he said.
In addition, Dr. Bianchini said the findings suggest that more information is needed about using different cutoffs for SP263 and 22C3, and he cautioned against “directly translating these finding to other disease settings or immune combinations.
“Defining new biomarkers to identify who within the PD-L1-negative group might benefit from this combination remains an unmet need,” he said. “For sure, I don’t see a space for the other tests to define this population,” he added.