NEW YORK (GenomeWeb) – A study exploring the feasibility of treating melanoma patients with combination checkpoint inhibitors in the neoadjuvant setting also enabled researchers to explore immune characteristics and tumor markers that may predict response.
The study published in Nature Medicine this week and led by researchers from MD Anderson Cancer Center, randomized 23 patients with high-risk resectable melanoma to neoadjuvant treatment with just Bristol-Myers Squibb's PD-1 inhibitor Opdivo (nivolumab) or Opdivo in combination with the CTLA4 inhibitor Yervoy (ipilimumab). After completing the course of treatment, study participants, most of whom had stage III disease, had their tumors surgically removed and received adjuvant Opdivo therapy.
Because the trial enrolled patients early in their disease continuum, researchers had the opportunity to collect biological samples at different time points — at baseline (before neoadjuvant treatment) and early on in treatment — and evaluate a variety of markers predictive of therapeutic benefit using immunohistochemistry-based gene expression analysis, T cell receptor sequencing, and exome sequencing.
Researchers also had access to NanoString's GeoMx Digital Spatial Profiler through a technology access program and used the platform to explore biomarkers associated with neoadjuvant immunotherapy treatment. GeoMx DSP enables quantification of tumor and immune markers with spatial resolution, which allows researchers to focus in on tissue microenvironment, a single cell, or subcellular region within a FFPE sample.
NanoString's platform is compatible with its nCounter instruments and next-generation sequencing, but in this study the analysis was done using the former. With the help of the platform, researchers found multiple immune markers associated with response to immune checkpoint blockade, and some of these were B cell makers, noted MD Anderson's Jennifer Wargo, who is also senior author of the Nature Medicine paper.
Overall, this study demonstrated the feasibility of neoadjuvant treatment with dual checkpoint inhibitors in patients with high-risk, resectable melanoma, but Wargo said that more research is needed to identify better tolerated combination regimens and further investigate predictive biomarkers.
In the study, 25 percent of patients who received Opdivo monotherapy before surgery saw their tumors shrink compared to 73 percent of patients receiving the Yervoy/Opdivo combination. The pathologic complete response rate was 25 percent in the Opdivo arm versus 45 percent in the combination arm. Although neoadjuvant treatment with the Yervoy/Opdivo combination was associated with improved progression-free and overall survival, these findings weren't statistically significant because the trial had to be stopped early and not enough patients were enrolled.
The data safety monitoring board halted the trial after two patients receiving Opdivo monotherapy progressed to the point that their tumors could no longer be surgically removed and because patients receiving the combination treatment experienced a high rate of grade 3 toxicities. "We're working actively to modify the regimens to be more tolerable," Wargo said.
However, with access to tissue samples from patients over different points in their care, researchers had a unique window into biomarkers that may be driving disease progression and treatment response. "We have to adopt a more integrated biomarker approach, but the best way to do that in my eyes is in the context of these neoadjuvant treatments," Wargo said.
Biomarker informed treatment strategies are usually studied in heavily pretreated, advanced cancer patients. This has proved challenging for studies exploring biomarker-informed treatment strategies, for example, because participants with late-stage disease, having undergone multiple lines of treatment, have died before they can receive the test results that could have informed their care. Oncologists involved in such trials have been advocating for studying precision oncology approaches in earlier cancer settings, when patients are healthier.
In the present study, "these are patients who have earlier stage but very high-risk disease," Wargo said. "You can get better insights into the activity of these drugs in the absence of all those pre-existing treatments that may impact later responses."
For example, Wargo and colleagues conducted IHC-based immune profiling of baseline and early on-treatment tumor samples and found that responders had higher CD8-positive cell infiltration, tumor cell PD-L1 expression, as well as lymphoid markers, including granzyme B, CD4, FoxP3, CD20, and PD-1, compared to nonresponders. "Early on-treatment biopsies showed a stronger predictive signal for most of these markers compared to baseline, consistent with previously published literature," the study authors wrote.
When researchers sequenced T cell receptors, responders tended to have higher clonality in pre- and on-treatment samples — something that's been reported in other studies. But the researchers also found some treatment-related differences.
"Patients who received nivolumab monotherapy were really kind of dependent on a clonal T cell repertoire in the tumor at baseline, and they had a lot of remodeling in the course of treatment," Wargo said. "Patients on combination treatment … could respond despite not having that clonal T cell repertoire."
Wargo also highlighted that NanoString's GeoMx DSP helped researchers quantify protein expression specifically in tumor-infiltrating immune cells. "Everyone thinks T cells are critical with regard to checkpoint blockade, and no doubt they are, but what we may be seeing here is that there may be an emerging role for B cells in potentially contributing to therapeutic response to checkpoint blockade," she said, adding that these B cell markers weren't specific to either treatment and associated with response at baseline and early on treatment.
"In this particular study, the high-plex capabilities of the system also allowed independent validation of the B cell result within a single study because the association with clinical response was seen for two different B cell markers measured simultaneously — CD19 and MS4A1 (CD20)," said Sarah Warren, NanoString's director of applications. Wargo and her colleagues are following up on this finding and plan to publish on it in more detail.
NanoString's DSP platform was also used in another study, led by researchers from the Netherlands Cancer Institute and published alongside the MD Anderson-led study in Nature Medicine. This study similarly evaluated neoadjuvant treatment of 20 melanoma patients with the Yervoy/Opdivo combination and explored various biomarkers. In a statement, Christian Blank from the Netherlands Cancer Institute, who is the first author, said researchers used the DSP platform to characterize pretreatment tumor biopsies and understand the extent to which pre-existing immune status of melanoma biopsies can be predictive of treatment response.
Similar to the MD Anderson led study, this trial demonstrated that neoadjuvant treatment with Yervoy and Opdivo had promising efficacy in stage III melanoma patients, but the combination also caused significant toxicity. The MD Anderson team is now studying the efficacy of Opdivo plus BMS' LAG3 immune checkpoint inhibitor, called relatlimab, which has a better toxicity profile than the combination of a CTLA-4 and PD-1 inhibitor. They are also part of the International Neoadjuvant Melanoma Consortium, in which several melanoma centers are pooling data to harmonize the evaluation of these regimens.