Bayesian Network Meta-Analysis (NMA): Safety of Bortezomib, Thalidomide or Lenalidomide Containing Regimens for Transplant Ineligible Multiple Myeloma (MM) Patients

Background
MM is a diagnosis of the elderly, with a median age of 70. Novel agents, such as proteasome inhibitors (bortezomib, V), and immunomodulatory agents (thalidomide, T; lenalidomide, R) have improved response rates and survival in transplant-ineligible patients; however, the disease is incurable and relapse is inevitable. Since older patients are particularly vulnerable to adverse events (AEs), the safety profile, particularly of multidrug regimens, is one of the most important considerations in treatment selection.

While numerous phase 3 studies have been done in this patient population, each study is limited to a comparison of no more than 2 or 3 regimens. Given the cost and time to conduct phase 3 studies, Bayesian NMA is a powerful tool that has been used to extract additional information from precious resources. The goal of our study is to compare the safety and preliminary efficacy of induction regimens for transplant ineligible MM patients using data from published phase 3 studies.

Methods
To identify phase 3 clinical trials comparing induction regimens in transplant ineligible patients, we conducted a systematic literature search using Pubmed, Google Scholar, Embase, ASH and ASCO meeting abstracts with the MeSH terms “multiple myeloma”, “randomized controlled trial” and “bortezomib” or “lenalidomide” or “thalidomide” or equivalent. 18 trials met our inclusion criteria. Among these trials, 9 trials did not have sufficient data on AEs (neutropenia, thrombocytopenia, venous thromboembolism and neuropathy grade 3 or higher by CTCAE) to be statistically analyzed. Data were pooled and reported as odds ratio (OR) for the 9 trials incorporated in the analysis. Bayesian NMA under the random-effects model was performed with STATA ver. 14.0.

Results
Because MP was the most frequently incorporated regimen in the 9 studies, it was utilized as the reference for NMA. The pooled OR of treatment A to have AEs compared to treatment B and 95 % confidence interval (CI) is presented in the table. The pooled OR of Rd 18 compared to MP was 0.57 and statistically significant. Continuous Rd (Rd) and VTP trended towards less AEs than MP (pooled OR 0.63 and 0.6, respectively) but were not statistically significant. The remaining regimens trended to have more AEs than MP. Rd, Rd 18 and VTP were associated with significantly less AEs than VMP with OR 0.31, 0.29, 0.3 respectively.

Existing evidence has demonstrated that the triplet regimens MPT, MPR and VMP have longer progression free survival (PFS) than MP. Rd is better than MPT and VMPT is better than VMP in terms of PFS. There is no head to head comparison to determine the best regimen among Rd, MPT, VMP, VMPT. We are contacting primary researchers of original trials to obtain individual patient data in order to conduct NMA for PFS and cost effectiveness.

Conclusion
Rd and VTP trend safer than MP and are significantly safer than VMP. Based on the systematic review, we do not have sufficient data to select the best regimen in terms of PFS. Further statistical analysis for efficacy will be available upon receipt of individual patient data.