CHICAGO – according to post hoc analyses of data from three randomized CheckMate studies.
In 70 treatment-naive patients from the randomized phase 2study who received sequential treatment with the programmed death-1 (PD-1) checkpoint inhibitor nivolumab (NIVO) followed by the cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) checkpoint inhibitor ipilimumab (IPI), best overall response was modestly associated with lower baseline serum IL-6 (P = .087) and significantly associated with on-treatment IL-6 (P = .006). In 70 patients who received IPI then NIVO, best overall response was associated only with on-treatment IL-6 (P = .043), Jeffrey S. Weber, MD, PhD, reported at the annual meeting of the American Society of Clinical Oncology.
“This stimulated us to look at associations with survival ... and there apparently was a significant association with high IL-6 levels in the serum both pretreatment and on treatment in both arms, whether they got NIVO then IPI followed by NIVO maintenance, or IPI then NIVO, also followed by NIVO maintenance,” he said.
After adjusting for covariates, the hazard ratios for survival for baseline IL-6 below versus above the median were 7.81 and 1.07, in the groups, respectively. No deaths occurred in the NIVO-IPI group (thus, no HR), but the HR for survival based on on-treatment IL-6 below versus above the median in the IPI-NIVO group was 1.92.
“So initial conclusions: High baseline and on-treatment IL-6 levels in the serum were associated with poor survival,” said, deputy director of the Perlmutter Cancer Center, New York University Langone Medical Center.
This finding prompted evaluation of additional samples from the randomizedstudy, which compared dacarbazine chemotherapy (the standard of care at the time) and NIVO in 400 treatment-naive patients with BRAF wild-type disease.
Again, baseline IL-6 levels (nondetectable vs. detectable) were associated with better overall survival (OS) in both groups (adjusted HRs, 1.79 and 1.54).
“So this is not a predictive marker, this is a baseline prognostic marker,” he said.
In the international, three-arm, randomized phase 3study, which compared IPI, NIVO, and IPI+NIVO in 945 treatment-naive patients with either BRAF wild-type or BRAF mutated disease, baseline IL-6 levels (nondetectable vs. detectable) again were associated with better OS in all 3 arms (adjusted HRs, 3.13 for NIVO, 2.67 for NIVO+IPI, and 4.06 for IPI alone).
A multivariate analysis of data from the CheckMate 066 and 067 studies, with controlling for lactic acid deydrogenase, performance status, and disease stage, provided additional “impressive evidence” of IL-6 as a potent prognostic factor, Dr. Weber said.
“We then looked at CRP. I’ve always been interested in CRP because in a recent publication CRP was found to be associated with outcomes in patients who got PD-1, and the higher the CRP, the worse they did,” he said.
In CheckMate 064 there was modest association between lower baseline CRP and best overall response in both the NIVO-IPI and IPI-NIVO groups (P = .069 and 0.009, respectively), and on treatment, the association was really only seen in the IPI-NIVO group (P = .210 for NIVO-IPI and 0.015 for IPI-NIVO), in which the higher CRP levels were associated with progression or stability.
For survival, however, both baseline and on-treatment CRP levels were associated with OS; baseline serum CRP above the median was associated with shorter OS (HRs, 7.25 for NIVO-IPI and 1.53 for IPI-NIVO), and a similar trend was seen for on-treatment CRP (HRs, 1.60 and 2.0, respectively).
In CheckMate 066, the association between CRP and OS was also apparent, but not as impressive for NIVO alone (HR, 0.996) as it was for dacarbazine (HR, 1.90), and similar to CheckMate 064, higher baseline CRP levels were associated with shorter survival and were prognostic, he said.
In CheckMate 067, similar trends were seen across the treatment arms, and they were similar to those seen for IL-6, with higher baseline CRP levels (at or above median versus below) associated with shorter OS (HRs, 1.46 for NIVO, 1.26 for NIVO+IPI, and 1.48 for IPI alone).
To better understand how CRP might inhibit the effects of PD-1 and how it could have an immune effect – as also indicated by some prior data – Dr. Weber and colleagues conducted additional in vitro studies to examine the impact of exogenous CRP on T-cell function; they found that CRP affected the earliest steps in T-cell signaling and activation, thereby dampening antitumor immune responses.
Acute phase reactants such as CRP and chronic inflammatory proteins including IL-6 (which induces production of CRP from the liver) have been associated with poor prognosis in a variety of cancers, as well as with poor outcomes after anti–PD-1 or programmed death-ligand 1 (PD-L1) therapy in melanoma and other cancers, Dr. Weber said.
“In murine models of melanoma and pancreatic cancer, combined treatment with anti-IL-6 blockade and anti–PD-1/PD-L1 antibodies enhances antitumor immune responses and efficacy,” he explained, noting that the current analyses were undertaken based on those findings and on “a significant body of data” from other groups and from his own lab.
The current findings suggest that IL-6 and CRP may be prognostic for immune checkpoint inhibitor therapies in patients with melanoma, he said, adding that “blockade of IL-6 and CRP synthesis and/or activity in combination with immune checkpoint therapies may enhance responses and survival rates in patients with different cancers, including melanomas.”
To that end, an investigator-sponsored trial looking at IPI-NIVO with the IL-6–blocking antibody tocilizumab has been approved and will start accruing patients in the next few months, he said.
During a discussion of the findings at the meeting,, associate director for clinical research at the Herbert Irving Comprehensive Cancer Center at Columbia University, New York, said that “Dr. Weber and his colleagues should be commended for really trying to show what CRP does to T-cell activation, and in the studies he showed us, it’s clearly negative.”
“But IL-6 is a pleiotropic cytokine. It will be very interesting to see what happens in the prospective clinical trial that he mentioned, in terms of all the other effects on CD-4 cells, neutrophils, and macrophages,” said Dr. Drake, who also is codirector of Columbia’s Cancer Immunotherapy Program. “Nevertheless, I think the data were clear that IL-6 and CRP are negative prognostic biomarkers in melanoma.”
Of note, the development of a biomarker identified in a trial typically takes many steps, but in the case of IL-6 – and perhaps even more so for CRP – the pathway is relatively short, Dr. Drake said.
“That’s because CRP is a validated and [Food and Drug Administration]–approved test; you can order it to assess cardiovascular risk in almost any hospital in the United States, and so the analyte – this part of the qualification – is done,” he explained. “I think if this was validated prospectively we could have CRP as a negative prognostic – not predictive – biomarker in melanoma, actually.”
Dr. Weber and Dr. Drake each reported relationships with numerous companies, including stock and other ownership interests and patents, consulting or advisory roles and/or receipt of honoraria, research funding to their respective institutions, and payment for travel, accommodations, and expenses
SOURCE: Weber J et al. ASCO 2019, .