The large phase III MAGRIT study investigating the MAGE-A3-specific vaccine GSK1572932A for patients with non-small cell lung cancer (NSCLC) will be completely halted following an interim analysis that demonstrated a lack of benefit.
In March, GlaxoSmithKline (GSK), the company developing the drug, reported that the immunotherapeutic failed to significantly extend disease-free survival (DFS) in two out of three patient populations. At this point, the company planned to continue the study in patients with a specific gene signature thought to predict response to therapy. However, GSK announced that further study of the interim analysis also indicated a lack of benefit in the genetic signature-positive subpopulation of patients. Following these findings, the study was stopped and unblinded will a full analysis expected at a later date.
The genetic signature in question was described in a retrospective analysis published in the Journal of Clinical Oncology
(JCO). This study examined patients from two phase II studies that investigated the vaccine. This analysis identified 84 genes that correlated with response, from mRNA samples using microarray analysis and quantitative polymerase chain reaction. The majority of the genes identified in the study were immune-related, including those in the interferon gamma pathway and specific chemokines.
This study examined the gene signature in patients with resected NSCLC who received treatment with MAGE-A3 plus the AS02B
immunostimulant. Overall, patients who tested positive for the genetic signature experienced an improvement in DFS (HR = 0.42; P
= .06) compared with negative patients (HR = 1.17; P
In September 2013, GSK announced that the same MAGE-A3 vaccine failed to extend DFS in certain patients with postsurgical melanoma when compared with a placebo in the phase III DERMA study. This study enrolled 1345 patients with MAGE-A3-positive stage IIIB or IIIC melanoma. This study was continued in select patients who expressed the gene signature. Following the cessation of the MAGRIT study, GSK announced that it still plans to continue DERMA as planned.
In the JCO
analysis, patients with melanoma who expressed the specific genetic signature saw an improvement in overall survival (OS) compared with placebo. This advantage was more pronounced when the AS15 immunostimulant was utilized (HR = .37; P
= .06) compared with AS02B
(HR = .84; P
“While we are extremely disappointed to learn that this trial did not have a positive outcome for the patients who participated in this trial, we are very grateful to its participants,” Vincent Brichard, Senior Vice-President & Head of Immunotherapeutics for GSK Vaccines, said in a statement. “We hope that the data generated in this trial will advance our understanding of the science of immunotherapeutics, and ultimately towards development of new therapies.”
The phase III placebo-controlled study assessed DFS with the MAGE-A3-directed therapy in the adjuvant setting for 2312 patients with stage IB, II, and IIIA resected NSCLC. All patients enrolled in the study expressed the MAGE-A3
gene, which occurs in approximately 35% of patients with resectable NSCLC.
Upon the cessation of the trial, the Independent Data Monitoring Committee that conducted the interim analysis noted that no specific safety concerns were caused by treatment with the MAGE-A3 vaccine.
The MAGRIT trial followed a similar design to a previous phase II study that randomized 182 patients with resected NSCLC in a 2:1 ratio to receive the MAGE-A3 immunotherapy (n = 122) or placebo (n = 60). The vaccine and placebo were administered in 13 intramuscular injections over a period of 27 months.
Treatment with the vaccine failed to statistically significantly extend disease-free interval (HR = .75, P
= .254), DFS (HR = .76, P
= .248), and OS (HR = .81, P
= .454) when compared with placebo. Despite these findings, the MAGRIT study was formed and labeled the “largest-ever” phase III lung cancer trial to study a tumor-specific approach to therapy. The study had a secondary endpoint aimed at the prospective validation of the predictive value for the 84-gene signature.