The use of measurable residual disease (MRD) assessment in multiple myeloma (MM) clinical trials has increased significantly between 2015 and 2020, but a lack of standardization limits its clinical applicability, according to a comprehensive analysis of 598 MM studies. The analysis, published in Nature, reveals that while MRD assessment is becoming more common, it is often used as an exploratory or secondary endpoint in non-randomized trials and is measured with varying methodologies and sensitivities.
The study, which examined interventional trials of myeloma with a start date between January 1, 2015, and December 31, 2020, found that MRD assessment was reported in 145 of these studies (24.2%). Of these, only a small fraction (6.2%) used MRD assessment as a stratification tool to determine treatment. The researchers performed a manual search for published data and trial protocols, finding additional data for 79 studies (54.4%). Among these, NGS-based methods were used in 38.3% of studies, and flow-based methods were used in 33.3% of studies.
Heterogeneity in MRD Assessment
The analysis highlighted significant heterogeneity in how MRD was assessed across trials. A sensitivity of 1/105 was used most often (48.3%), but the detection threshold was left unspecified in 31.7% of studies. This variability in collecting, reporting, and analysis limits the immediate clinical applicability of MRD, according to the authors.
"The current variability in collecting, reporting and analysis limits the immediate clinical applicability of MRD," the researchers stated.
Trends in MRD Usage
The proportion of trials with MRD assessment showed a clear upward trend from 13.3% in 2015 to 33.1% in 2019 and 30.0% in 2020. However, a decrease in the fraction of randomized studies with MRD assessment was observed (58.3% in 2015 to 37.0% in 2020). The number of studies that stratified treatment based on MRD showed an increasing trend (0 in 2015 to 3 in 2020).
Implications and Future Directions
The authors emphasize that comparing survival between those who achieve MRD negativity in both arms of a study to those who do not is not a valid way to establish trial-level surrogacy. They argue that surrogacy should be demonstrated by showing that between-arm differences in MRD status predict between-arm differences in survival.
The researchers also note that the lack of imaging during MRD assessment in many trials limits the understanding of disease response, as patients can have residual disease on metabolic imaging even with a complete serological remission and bone marrow MRD negativity. Ongoing technological advances beyond bone marrow NGS/NGF analysis, including imaging, mass spectrometry, and circulating tumor DNA-based methods, further complicate harmonization of MRD assessment and require further study.
Study Limitations
The study relies on data provided by sponsors to ClinicalTrials.gov, and not a review of individual protocols. As such, the high rate of “unspecified” methodology of assessment likely reflects a lack of information on ClinicalTrials.gov. The authors acknowledge that their analysis does not include recently listed protocols on ClinicalTrials.gov, such as NCT04934475/MIDAS and NCT05231629/MASTER-2.
