Therapeutic outcome of early-phase clinical trials in multiple myeloma: a meta-analysis

Abstract

Great progress in the treatment of patients with multiple myeloma (MM) has been made due to the development of novel drugs. Patients with relapsed/refractory MM (RRMM) can be enrolled in early-phase clinical trials, but their performance across the last decade is unknown. We conducted a meta-analysis on the overall response rate (ORR) and toxicity. PubMed, Embase, and Cochrane Library were systematically searched for phase I and phase II trials investigating an experimental compound as a single agent or in combination with dexamethasone, published from January 1, 2010 to July 1, 2020. Eighty-eight articles were included, describing 61 phase I trials involving 1835 patients and 37 phase II trials involving 2644 patients. There was a high degree of heterogeneity. Using a random-effects model, the 95% CIs of the estimated ORR were 8-17% for phase I trials and 18-28% for phase II trials. There were significant subgroup differences in ORR between the years of publication in phase I trials and between drug classes in both phase I and phase II trials. The ORR in early-phase clinical trials in RRMM is substantial, especially in phase II trials, but due to high heterogeneity a general assessment of clinical benefit before participation is difficult to offer to patients.

Fig. 1. PRISMA flow diagram of the study selection process.

ITT intention-to-treat, RRMM relapsed/refractory multiple myeloma.

Fig. 2. Forest plots of overall response rates and dose-limiting toxicities in phase I trials.

A Overall response rates per included phase I trial. B Drug-limiting toxicities per included phase I trial. Forest plots are arranged by the population size of trials within each year of publication. Squares represent estimated proportions, with the size of the squares representing the weight of each trial according to the inverse variance method. Horizontal lines through the squares indicate 95% CIs. The points of the diamond indicate the 95% CI of the pooled mean. The horizontal line at the bottom indicates the prediction interval. Blank lines indicate missing information from the trial paper. *The study population in Chen et al. included two patients with Waldenström’s Macroglublinemia.

Fig. 3. Meta-analysis of overall response rates and dose-limiting toxicities in phase II trials.

A Overall response rates per included phase II trial. B Drug discontinuations - per included phase II trial. Forest plots are arranged by the population size of trials within each year of publication. Squares represent estimated proportions, with the size of the squares representing the weight of each trial according to the inverse variance method. Horizontal lines through the squares indicate 95% CIs. The points of the diamond indicate the 95% CI of the pooled mean. The horizontal line at the bottom indicates the prediction interval. Blank lines indicate missing information from the trial paper.

Fig. 4. Estimates of overall response rates for subgroups.

A Overall response rates in phase I trials by drug class. B Overall response rates in phase I trials by years of publication. C Overall response rates in phase II trials by drug class. D Overall response rates in phase II trials by years of publication. Squares represent estimate values. Horizontal lines through the squares indicate 95% CIs. IMiD immunomodulatory imide drugs, PI proteasome inhibitors, mAb monoclonal antibodies, ADC antibody–drug conjugate, ICI immune checkpoint inhibitor, Hsp90i heat-shock protein 90-inhibitor.