When a randomized controlled trial is unlikely: Propensity score analysis of blinatumomab in adults with relapsed/refractory Philadelphia chromosome–positive B‐cell acute lymphoblastic leukemia

Abstract

Over the past several years, there has been substantial progress made in the approach to treating adults with relapsed/ refractory B-cell acute lymphoblastic leukemia (B-ALL). Previously, the only options available outside the context of a clinical trial were traditional cytotoxic chemotherapy or, for those patients with Philadelphia chromosome–positive (Ph+) disease, ABL tyrosine kinase inhibitors (TKIs). Unfortunately, the expected outcomes with these approaches is poor, with rates of complete remission reported to be generally no better than 30%, a median survival measured in months, and a 5-year survival rate of typically <10%. Recent developments in the treatment of patients with relapsed/refractory B-ALL truly are at the cutting edge of oncology therapeutics. The armamentarium now includes the CD22-targeted antibody-drug conjugate inotuzumab ozogamicin; the CD3 to CD19 bispecific T-cell engager blinatumomab; the CD19-targeted chimeric antigen receptor–modified T-cell product tisagenlecleucel; and increasingly potent TKIs, including ponatinib. Both inotuzumab ozogamicin and blinatumomab were found to be superior to the investigators’ predefined choice of salvage chemotherapy in prospective randomized trials. Although the study with inotuzumab ozogamicin included both patients with Philadelphia chromosome–negative (Ph-) and Ph+ B-ALL, the corresponding trial of blinatumomab was performed specifically in those with Phdisease. A separate, single-arm, phase 2 study in patients with Ph+ B-ALL demonstrated clinical activity comparable to the experience among patients with PhB-ALL. Although one could debate the merits of these 2 strategies, they each met their fundamental goals and both agents currently are approved by the US Food and Drug Administration and international regulatory agencies for adults with Phand Ph+ relapsed/refractory B-ALL. However, blinatumomab technically has not been subjected to a prospective randomized comparison in patients with Ph+ B-ALL. And when considering the existing data (somewhat limited though they are), the potential ethical concerns, and the necessary resources needed to pursue such a trial, it appears highly improbable that we will ever truly close this gap. It is in this context that Rambaldi et al have attempted to at least narrow this gap with their analysis, which is presented in the current issue of Cancer. The authors sought to compare results from the prospective phase 2 study of blinatumomab in patients with relapsed/refractory Ph+ B-ALL with those of a historical cohort of patients selected from 2 clinical databases from centers in Italy and Spain. Because of substantial heterogeneity and variability, they compared these groups using a propensity score analysis (PSA) with Bayesian augmentation. Sparing the details of the statistical methods, this technique strives to reduce bias that may exist between groups due to disease characteristics, treatment factors, or other known predictors of outcome (although some have argued whether it is superior to more traditional multivariable methods). Ultimately, Rambaldi et al found that blinatumomab yielded a statistically significant improvement in survival and a nonsignificantly higher rate of complete remission than observed among the historical controls. They appropriately emphasized that this methodology is not a replacement for prospective randomized trials and thus they were unable to conclude that blinatumomab is superior in patients with relapsed/refractory Ph+ B-ALL. However, PSA can be a useful tool for the comparison of novel approaches with existing standards of care in patients with rare diseases, in whom the gold standard of comparison is impractical (if not impossible).