SAN FRANCISCO—Two new potential approaches have been identified to help determine which cancer patients are most likely to benefit from immunotherapy. Only one predictive biomarker has been approved to guide decision-making about offering immunotherapy for checkpoint inhibitors—expression of the PD-L1. Several biomarkers under investigation appear to complement or improve upon PD-L1 assays.
“It is clear that, while checkpoint inhibitors produce remarkable responses in some patients, others yield no benefit at all,” said Catherine S. M. Diefenbach, MD, Director of the Clinical Lymphoma Program at NYU Langone's Perlmutter Cancer Center, in an ASCO Perspectives article released in conjunction with the 2019 ASCO-Society for Immunotherapy of Cancer (SITC) Clinical Immuno-Oncology Symposium. “Similarly, we lack tools to evaluate biomarkers that could identify which patients will tolerate these therapies well and which patients are at risk of substantial immune toxicity.”
The results from two studies presented at the ASCO-SITC symposium show progress in identifying predictive markers. In one study, genetic alterations in plasma cell-free DNA, typically referred to as circulating tumor DNA (ctDNA), showed that mutations in the EGFR, 1ALK, PTEN, and STK11 genes conferred poor outcomes with immunotherapy, while mutations to the KRAS or TP53 genes predicted good outcomes to immunotherapy.
In the second study, liquid biopsies were used to explore whether protein profiles could be predictive for response to immunotherapies. Researchers determined that differentially expressed proteins, or those expressed at different quantities and levels in a cell, could help distinguish between patients who did or did not respond to checkpoint inhibitors.
Circulating Tumor DNA
Targeting limited, but relevant, genetic alterations in plasma ctDNA with next-generation sequencing along with early monitoring may represent a non-invasive approach to predict response to immune checkpoint inhibitors in advanced non-small cell lung cancer (NSCLC) (Abstract 103).
Researchers collected plasma samples from responders who achieved progression-free survival (PFS) of more than 6 months and non-responders who had progressive disease at first evaluation before second-line nivolumab was initiated, as well as at 1 month. The researchers, led by Nicolas Guibert, MD, of Inserm, Cancer Research Center of Toulouse in France, sequenced hotspots and coding regions from 36 genes. Molecular profiles of ctDNA and its early kinetics at 1 month were analyzed as potential early indicators of response.
Some 98 NSCLC patients were analyzed, including 86 patients who were evaluable for response; 39 patients were responders and 47 patients were non-responders. Alterations in ctDNA were detectable in 67 of 86 baseline samples (78%).
“The detection of a targetable oncogenic driver (five EGFR, one ALK) was associated with progressive disease on the first CT scan,” Guibert noted. The presence of PTEN and/or STK11 mutations was correlated with poor outcomes, while the presence of transversion mutations in KRAS and/or TP53 predicted good outcomes.
“Combining these results, patients with a low immune score derived poor outcomes (PFS 2 months), compared with patients with a high immune score (PFS 14 months),” he said.
Studying early changes in 65 specimens, molecular response was correlated with clinical outcomes—14-month PFS in patients with early ctDNA decrease as compared to 2 months in patients with an increase.
“Targeted sequencing of plasma ctDNA and its early variations can predict response to anti-PD-1,” Guibert stated.
Using Plasma Proteomic Profiling
Proteomic profiling of patient plasma samples may reveal predictors of immunotherapy response and uncover biological insights underlying primary resistance (Abstract 130).
Researchers at Massachusetts General Hospital in Boston included 58 metastatic melanoma patients receiving anti-programmed death (PD-1) therapy with pembrolizumab or nivolumab in an initial cohort. They added in 150 additional patients into a validation cohort.
For the study, the researchers collected plasma samples at baseline and several on-treatment time points. Then they analyzed the samples for 1,102 proteins. A subset of patients had single-cell RNA-sequencing performed on tumor tissue.
Across the treatment period, 70 significantly differentially expressed proteins were identified, including markers of immune activation (PD-1, CXCL9, CXCL10, CD25, and IL-17a, among others). Also, 38 significantly differentially expressed proteins were identified with on-treatment time points between anti-PD-1 responders and non-responders, including several implicated in primary or acquired resistance (IL-8, MIA, and ERBB2, among others).
“Importantly, we demonstrate the relationship of these serum biomarkers to overall and progression-free survival,” said lead author Arnav Mehta, MD, PhD, Clinical Fellow in Medicine at Mass General. “We employed statistical learning approaches to build classifiers of treatment response, leveraging early and late on-treatment time points.”
Analysis of single-cell RNA-sequencing data of tumor tissue revealed that gene expression of most proteins predictive of response were enriched among tumor myeloid cells, with the remainder of proteins being reflective of exhausted T-cell states.
In conclusion, Mehta stated: “Whole plasma proteomic profiling of anti-PD-1 treated patients revealed differentially expressed proteins between responders and non-responders that may enable a liquid biopsy to predict anti-PD-1 response. These results unveil a putative role of myeloid cells within the tumor microenvironment in anti-PD-1 response or primary resistance.”
“Predictive biomarkers are critical for identifying patients who could benefit the most from treatment. Perhaps nowhere else in oncology today are these markers more urgently needed than in the growing field of immunotherapy,” said Diefenbach. “Many of the predictive markers currently being explored in laboratories worldwide will require extensive validation. New checkpoint inhibitors continue to be developed, so a synergistic approach between therapeutic development and predictive marker validation should be the goal.”
Mark L. Fuerst is a contributing writer.