Identifying Patients Who Respond to Immunomodulator MK-3475

Roxanne Nelson

April 08, 2014

SAN DIEGO — The data are very early, but researchers are cautiously optimistic that they have discovered a way to predict which patients will respond to the investigational immunotherapy MK-3475 (pembrolizumab, under development by Merck & Co.), which has shown promising activity in non-small cell lung cancer (NSCLC) and melanoma.

The findings were presented here at the American Association for Cancer Research 2014.

MK-3475 acts as a checkpoint inhibitor in the interaction between T-cells and cancer cells, and specifically inhibits the programmed cell death 1 (PD-1) pathway. At the meeting, researchers presented clinical trial data suggesting that the best responses to MK-3475 are in patients whose tumors express the programmed death ligand 1 (PD-L1).

An analysis of the data from melanoma clinical trials showed a "major difference in the response rates between patients with PD-L1-positive and PD-L1-negative tumors treated with MK-3475," said lead author Adil Daud, MD, codirector of the UCSF Melanoma Center and director of melanoma clinical research at the UCSF Helen Diller Family Comprehensive Cancer Center in San Francisco. "This is the largest dataset, to my knowledge, to look at PD-L1 expression in tumors from melanoma patients treated with PD-1 inhibitors, he said.

In this phase 1 clinical trial, Dr. Daud and colleagues evaluated the relation between PD-L1 tumor expression and outcome and assessed T-cell activation as a pharmacodynamic marker of MK-3475 activity.

The study cohort involved 195 patients with advance melanoma who received MK-3475 at different doses: 10 mg/kg every 2 weeks, 10 mg/kg every 3 weeks, or 2 mg/kg every 3 weeks. All patients had been pretreated, some with the immunotherapy agent ipilimumab.

Of the 125 evaluable patients, 89 had tumors that were positive for PD-L1 and 36 had tumors that were negative. Tumors were considered to be positive for PD-L1 if at least 1 in 100 tumor cells expressed the protein.

Overall response rate was better when tumors expressed PD-L1 than when they did not (46% vs 17%). Progression-free survival at 6 months was also better in patients with tumors that expressed PD-L1 (64% vs 34%).

At 1 year, more patients with tumors expressing PD-L1 were alive (86% vs 72%).

Activity was observed in patients with low PD-L1 expression and, when HLA-DR expression was used as a marker, post-treatment T-cell activation in the circulating pool increased at all drug doses tested. At 6 weeks, there was a statistically significant increase from baseline, at all doses of MK-3475, in the pooled mean percent change of activated CD4+ T-cells expressing HLA-DR antigens (17.5%) and activated CD8+ T-cells expressing HLA-DR antigens (24.0%).

There was no significant difference in overall response rate between previous ipilimumab treatment and no previous ipilimumab treatment for patients with PD-L1-positive tumors (44% vs 47%) or patients with PD-L1-negative tumors (14% vs. 17%).

Unaddressed Concerns

Although the data are encouraging, MK-3475 is not ready for prime time, said study discussant Mario Sznol, MD, clinical research program leader in the melanoma program at the Yale Cancer Center in New Haven, Connecticut.

It is important that variables such as previous therapies, type of biopsy, and variations in cutpoints are addressed, said Dr. Sznol. We know that "previous therapy can make a difference, and it appears that immunotherapies can induce higher PD-L1 levels," he noted. But, "does it make a difference in PD-L1 levels if we use fresh or frozen samples, or if we biopsy a primary or metastatic tumor?

Another concern is the multiple assays in development. "We need to know if they correlate," he said.

Patients with low PD-L1 expression, or negative tumors, might benefit from these agents as well, but they can be excluded because of the cutoff points, Dr. Sznol pointed out. This can be relevant for those with limited or no treatment options.

"We must wait for phase 3 trial to determine the benefit in the low-expression group," he said. A global assessment of the many factors in blood and tumor that can influence the effectiveness of immune interventions will likely be necessary to develop predictive biomarkers.

High Levels for Best Response in NSCLC

In preliminary findings from the NSCLC study of MK-3475, high levels of tumor PD-L1 expression were associated with tumor response, progression-free survival, and overall survival.

Patients with previously treated NSCLC have limited options and a poor prognosis, explained lead investigator Leena Gandhi, MD, PhD, assistant professor of medicine at Harvard Medical School in Boston. But "the responses we have seen among patients whose tumors express high levels of PD-L1 appear to be quite durable, which is extremely exciting."

In the study, previously treated NSCLC patients received MK-3475 10 mg/kg every 3 weeks. Tumor response was assessed every 9 weeks, and a tumor biopsy was performed in the 60 days before the first dose of MK-3475.

The response rate was better in patients with tumors that had high pretreatment PD-L1 levels than in those with low pretreatment levels (37% vs 11%).

At 6 months, more patients with high levels of PD-L1 than with low levels had no disease progression (41% vs 17%). In addition, more patients with high levels of PD-L1 were still alive (72% vs 53%).

"Data from 2 ongoing studies will be used to further explore the relation between tumor PD-L1 expression and MK-3475 activity in NSCLC patients," concluded Dr. Gandhi. She noted that 300 additional patients from the trial will be prospectively analyzed to explore the clinical utility of strong and weak cutpoints.

In the ongoing KEYNOTE-010 study comparing MK-3475 (2 mg/kg and 10 mg/kg) with docetaxel, the cutpoint will be validated, she reported.

Clinical development of MK-3475, as monotherapy and as part of combination strategies, is ongoing in multiple solid tumors and hematologic malignancies.

Trials looking at investigational immunotherapies are ongoing, said Bo H. Chao, MD, MS, assistant professor in the division of medical oncology at the Ohio State University Comprehensive Cancer Center in Columbus. "The studies are showing significant and very promising results in non-small cell lung cancer."

This specific study is actually an update; results were originally presented last year at the World Lung Cancer Conference, Dr. Chao told Medscape Medical News. It looks at the correlation between PD-L1-receptor status and outcome to see if it is a biomarker for response, he said. "Based on the preliminary data, it appears to be yes."

At the recent European Lung Cancer Conference (ELCC) by Fred Hirsch, MD, PhD, professor of medicine and pathology at the University of Colorado in Denver, presented data on this association and discussed PD-L1 testing.

At the moment, each company developing one of these agents has its own test, but is not discussing (indeed, is not allowed to discuss for competitive reasons) the details of the test with other researchers, Dr. Hirsch said at that meeting. So there is uncertainty about what antibody is being used and what cutoff points are being used; consequently, it is unclear how the results from one of these tests will compare with the others.

Collaboration and some uniformity need to be introduced into this field. This will turn into a nightmare if these drugs are approved with individual companion tests, leading to pathologists having to perform many different variations of the same test, he noted.

Dr. Hirsch reminded the ELCC audience that the correlation between PD-L1 positivity and drug response has not been definitely established. He showed waterfall plots of responses to new agents to demonstrate that patients who have tested positive are responding, but responses in patients who have tested negative often appear to be just as good. It is not clear how useful such testing will be as a predictor for drug response, he said.

Both studies were funded by Merck. Dr. Gandhi has disclosed no relevant financial relationships. Dr. Daud reports serving on the advisory boards of Merck and GlaxoSmithKline PLC.

American Association for Cancer Research (AACR) 2014: Abstracts CT104 and CT105. Presented April 6, 2014.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.