Evaluation of multiple myeloma measurable residual disease by high sensitivity flow cytometry: An international harmonized approach for data analysis

Background Multiple myeloma (MM) measurable residual disease (MRD) evaluated by flow cytometry is a surrogate for progression-free and overall survival in clinical trials. However, analysis and reporting between centers lack uniformity. We designed and evaluated a consensus protocol for MM MRD analysis to reduce inter-laboratory variation in MM MRD reporting. Methods Seventeen participants from 13 countries performed blinded analysis of the same eight de-identified flow cytometry files from patients with/without MRD using their own method (Stage 1). A consensus gating protocol was then designed following survey and discussions, and the data re-analyzed for MRD and other bone marrow cells (Stage 2). Inter-laboratory variation using the consensus strategy was reassessed for another 10 cases and compared with earlier results (Stage 3). Results In Stage 1, participants agreed on MRD+/MRD- status 89% and 68% of the time respectively. Inter-observer variation was high for total numbers of analyzed cells, total and normal plasma cells (PCs), limit of detection, lower limit of quantification, and enumeration of cell populations that determine sample adequacy. The identification of abnormal PCs remained relatively consistent. By consensus method, average agreement on MRD- status improved to 74%. Better consistency enumerating all parameters among operators resulted in near-unanimous agreement on sample adequacy. Conclusion Uniform flow cytometry data analysis substantially reduced inter-laboratory variation in reporting multiple components of the MM MRD assay. Adoption of a harmonized approach would meet an important need for conformity in reporting MM MRD for clinical trials, and wider acceptance of MM MRD as a surrogate clinical endpoint.

Automated summary

The study designed and evaluated a consensus protocol to reduce inter-laboratory variation in multiple myeloma measurable residual disease (MRD) analysis. Results showed improved consistency in MRD- status and near-unanimous agreement on sample adequacy by adopting the consensus method.