Dr. Thomas E. Witzig
The PILLAR-2 trial tested the safety and efficacy of adjuvant everolimus in patients with poor-risk DLBCL (International Prognostic Index, 3 to 5) who had achieved complete remission (CR) with prior rituximab-combined chemotherapy. This patient population is known to be at a high risk of relapse. Treatment with rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) for six cycles without any maintenance therapy cures approximately 60% of poor-risk patients. However, 20% of patients do not achieve a CR with R-CHOP, and an additional 20% will experience relapse after R-CHOP.
“The pattern of relapse is that basically all of the relapses occur in the first 2 years and then the survival curve begins to level out,” said Dr. Witzig. “Therefore, the key to improving R-CHOP, which is what we were trying to do, is to straighten that event-free survival curve for the first 24 months from the diagnosis.” Earlier studies haves shown that everolimus has single-agent activity in relapsed DLBCL, with a 30% overall response rate (ORR). This led to the hypothesis that everolimus may improve outcomes after R-CHOP treatment because it targets the P13K/mTOR pathway, which is known to be dysregulated in DLBCL.
The PILLAR-2 study included 742 patients randomly assigned to receive 10 mg of everolimus daily (372 patients) or placebo (370 patients) for 1 year until disease relapse, unacceptable toxicity, or death. All patients were PET-negative prior to treatment. The median study follow-up was 50.4 months (range, 24.0 to 76.9 months).
Adjuvant everolimus did not improve DFS versus placebo (HR 0.92, 95% CI [0.69, 1.22]; log-rank p = 0.276). The 2-year DFS rates were 77.8% (95% CI [72.7, 82.1]) in the everolimus arm compared with 77% (95% CI [72.1, 81.1]) in the placebo arm. In contrast, the study protocol anticipated a 2-year DFS rate of 74% and 65% in the everolimus and placebo arms, respectively (HR 0.7).
Dr. Anas Younes
“There were trends in these selected groups and this perhaps warrants further investigation—either with this type of drug or other drugs targeting this pathway,” said Dr. Witzig.
“The adjuvant everolimus had a safety profile consistent with what we already knew about the drug,” he added. In this investigation, 65% of patients in the everolimus arm (239 patients) had a grade 3/4 adverse event, most of which were thought to be drug-related. The most common 3/4 adverse events were neutropenia (everolimus, 17%; placebo, 13%) and stomatitis (everolimus, 8%; placebo, < 1%).
“Surprisingly, 42% of the placebos had a grade 3 or 4 toxicity,” said Dr. Witzig, which highlights the continued need for placebo-controls.
Adjuvant Maintenance Failure Not Uncommon
“Other adjuvant or maintenance studies have not been successful,” said Dr. Witzig. “I think another approach to target this pathway would be to combine everolimus with R-CHOP.” He referenced phase I results from the NCCTG1085 study that showed a 96% CR rate with combined everolimus and R-CHOP.
“I think we should do more biomarker selection,” said Anas Younes, MD, of Memorial Sloan Kettering Cancer Center, who commented on the trial and offered the first of three strategies to improve the success rate of novel agents in this setting. The PILLAR-2 trial involved unselected patients. He also suggested investigation of mechanism-based combinations with different mechanisms of action or drug resistance, either in the front-line or relapsed settings. Finally, he recommends innovative clinical trial designs such as including multiple novel-novel combinations in one study and studying concurrent versus sequential combinations.
– Alice McCarthy