Optimized Flow Cytometry Incorporating t-SNE Enables Minimal Residual Disease Assessment in a Philadelphia Chromosome-positive Acute Lymphoblastic Leukemia at or Near the Lower Detection Limit

This is the authors' abstract. We don't add key points for this paper.

Unsupervised machine learning shows significant promise for evaluating multicolor flow cytometry, particularly in advancing minimal residual disease (MRD) analysis. We implemented t-distributed stochastic neighbor embedding (t-SNE)-assisted MRD analysis in a challenging case of Philadelphia chromosome-positive acute lymphoblastic leukemia. This approach enabled us to identify MRD on an unbiased basis with a sensitivity comparable to that of genetic analysis while simultaneously reducing the workload. t-SNE is a valuable tool for the detection and verification of MRD, with its effectiveness significantly enhanced through the integration of optimized antibody panels, data preprocessing, and rigorous back-gating verification.