NEW YORK – At the American Association for Cancer Research's annual meeting on Saturday, researchers presented data showing that an immunogenic signature may improve the identification of metastatic castrate-resistant prostate cancer (mCRPC) patients who benefit from dual checkpoint inhibitor treatment.
In the ongoing, two-part, Phase II NEPTUNES study, researchers led by Mark Linch, head of the uro-oncology biology group at the University College London, are evaluating whether mCRPC patients with a distinct immunogenic signature respond more favorably to the combination of Bristol Myers Squibb's PD-1 inhibitor nivolumab (Opdivo) and its CTLA-4 inhibitor ipilimumab (Yervoy). At the meeting, Linch reported data from the first part, in which not enough mCRC patients responded to nivolumab-ipilimumab for researchers to be able to conclude that the immunogenic signature was a fully refined and robust predictor of best responders.
However, Linch believes there were enough signals to warrant further evaluations of dual checkpoint inhibitor treatment in biomarker-selected mCRPC patients. "Although we didn't quite hit our prespecified target of 12 responders … we've clearly got a signal of activity in these biomarker groups and we need more information," he said.
Typically, mCRPC patients don't respond well to single-agent checkpoint inhibitors. However, in the CheckMate-650 trial, 10 percent of patients who received two checkpoint inhibitors, nivolumab and ipilimumab, responded after docetaxel chemotherapy and 25 percent of patients responded to the combination before docetaxel. Although this suggests that some mCRPC patients are benefitting from this treatment, this analysis was conducted without using biomarkers to select a patient population.
Given that prior studies have shown that around 20 percent of mCRPC patients had tumor infiltrating lymphocytes, a biomarker of immunotherapy response, Linch and colleagues set out to explore in the NEPTUNES trial if they could use an immunogenic signature to select men who are more likely to respond to the nivolumab-ipilimumab combination.
Patients had to have archival tissue or provide a biopsy sample for assessment of the immunogenic signature, which involved assessment of defective DNA damage repair mechanisms, mismatch repair deficiencies, or high inflammatory tumor infiltrates. To derive the signature, Linch and colleagues profiled patients' tumors using University of Washington's OncoPlex cancer gene panel, evaluated whether their tumors had mismatch repair deficiency using immunohistochemistry, and used a multiplexed IHC assay to determine staining for CD4, CD8, or FOXB3 on immune cells.
Out of 184 mCRPC patients screened with this signature, 19 percent could not be evaluated due to insufficient tissue or inadequate quality or quantity of DNA. Meanwhile, testing found 62 patients, or 34 percent, to be immune signature "positive," and 87 patients, or 47 percent, were "negative."
Researchers ultimately enrolled 35 patients into one of the study's two cohorts, in which they received the dual checkpoint inhibitor regimen for six weeks followed by nivolumab for up to a year. Patients were heavily pretreated, having received more than one line of systemic therapy in the hormone sensitive or the castrate-resistant setting.
The primary endpoint in the study was a composite response rate, which factored in radiological measurements using RECIST criteria, a decline of prostate-specific antigen of at least 50 percent from baseline, and conversion of circulating tumor cells at nine weeks. Researchers also tracked overall survival and radiologic progression-free survival as secondary endpoints. They hoped to see a 40 percent composite response rate, a far better outcome than seen in the unselected CheckMate-650 trial, and prespecified that at least 12 patients would have to respond to reject a null hypothesis.
Ultimately, in the study, only 10 immune signature-positive mCRPC patients had a composite response. "This does not achieve the twelve out of thirty-five [responses] that we were hoping for, but it is certainly higher than what has previously been seen for combination checkpoint inhibitors in prostate cancer," Linch said during a presentation at the meeting. Median overall survival in this group was around 18 months and median progression-free survival was around five months.
Most patients who responded were still responding at the time of data cutoff in early January. Linch estimated that at 12 months, 70 percent of the 10 responding patients were still benefitting on the dual nivolumab-ipilimumab combination.
Researchers also considered response rates according to specific biomarkers. Among responders were four out of five patients with mismatch repair deficiency; the fifth patient had an unconfirmed partial response. Although this wasn't entirely unexpected since other studies have shown patients with mismatch repair loss respond to checkpoint inhibitors, "this was an interesting finding that there's a high penetrance of response in that particular subgroup," Linch said.
Three out of four patients with mutations in BRCA1/2 genes; three out of 10 patients who were in the trial exclusively based on high inflammatory tumor infiltrates; and one patient out of seven with a CDK12 aberration responded. No patients with altered ATM, CHD1, or CHEK2 genes responded to the combination treatment.
The most common adverse event was diarrhea in 57 percent of patients, and 20 out of 35 patients had a serious adverse event, with gastrointestinal toxicities being the most common. While there were no treatment-related deaths, only 11 patients completed all four cycles of the combination treatment. While some patients stopped treatment due to progressive disease, Linch said, many of those discontinuations were due to treatment-related toxicity.
Ultimately, Linch concluded that nivolumab and ipilimumab demonstrated antitumor activity in this cohort of mCRPC patients who were positive for an immunogenic signature. Still this study outcome "is insufficient to reject the null hypothesis," he acknowledged.
One difficulty was the high treatment discontinuation rate in this part of the trial, which led to many patients not receiving the full schedule of nivolumab-ipilimumab. "Alternative dosing schedules may be required," Linch said, noting that in the second part of NEPTUNES, which began enrolling in September, patients will receive a lower ipilimumab dose.
Researchers are also conducting additional translational biomarker analysis, for example, looking at patients' neoantigen burden and AR-V7 expression status, which they plan to present at a future conference. Linch noted that although in this cohort patients were heavily pretreated, he'd like to explore if patients earlier in their treatment trajectory will fare better on dual checkpoint inhibitor therapy. He added that researchers will learn more on the activity of this combination in mCRPC patients enrolled in other cohorts of the BMS-led CheckMate-650 trial, as well as in the second NEPTUNES cohort.
"With greater information, we'll be able to refine that biomarker," Linch said, adding that this study ultimately drives home the point that researchers should no longer be studying these drugs in all-comer populations. "Of course, we need more information about these biomarkers, but we should be directing our future studies in biomarker-selected groups," he said.