Single-Cell Multi-Omic Analysis of AML MRD Reveals Differences in Clonal Architecture between Relapse and Non-Relapse Cases

Introduction

The Mission Bio AML single-cell MRD assay quantitatively characterizes SNVs and surface protein expression simultaneously across thousands of individual cells. In contrast, bulk NGS requires averaging across the entire population, preventing co-localization of joint genetic lesions or changes in cellular immunophenotype, and leukemia-associated immunophenotyping by flow cytometry can miss residual AML cells that have the same genotype, but different immunophenotype from the diagnostic population. These methods are commonly discordant, leading to clinical questions for individual surveillance and treatment. Combining these assays with single-cell resolution overcomes these limitations and reveals subclonal populations that may harbor resistance mutations or unique therapeutic targets.

Methods

We analyzed 17 paired diagnosis/remission samples (n=34) from SWOG S0106 trial previously characterized for MRD via flow cytometry and duplex sequencing (Dillon, Haematologica 2024) using the commercial Mission Bio AML single-cell MRD assay at the RUO laboratory at Mission Bio. This assay uses CD34/CD117 magnetic column enrichment to identify potential leukemic cells followed by multiplexing of 3 samples per machine run. We focused on discrepant cases where flow cytometry was positive for MRD, but the patient did not relapse (MRD+, relapse-; n=5); and where flow was negative, but the patient did relapse (MRD-, relapse+; n=6). We also processed true positive (MRD+, relapse+; n=4) and true negative (MRD-, relapse-; n=2) cases for comparison.

Results

After demultiplexing three independent patient samples per Tapestri machine cycle, we analyzed several thousand cells/sample. As expected, residual disease was identified in MRD+, relapse+ samples and was not identified in MRD-, relapse- samples. For the MRD-, relapse+ cases, we identified MRD in 2/6 cases; with the remaining 4 cases having reported disease present below our assay’s limit of detection of 10^-4. Of the 5 MRD+, relapse- cases, 3 had residual cells with the AML genotype where the remaining 2 did not. Of note, cases that relapsed generally had a more branched clonal architecture at diagnosis than cases that did not relapse, regardless of flow results (Relapse+ vs Relapse- averages: 4.6 vs 1.3 clones/patient, respectively).

Conclusion

Single cell characterization of AML via the Mission Bio AML MRD assay enables the simultaneous assessment of SNVs and indels in genetic regions recommended by ELN guidelines along with surface immunophenotyping. Whereas bulk NGS and flow are frequently discrepant, this assay resolves genotype:phenotype ambiguity with greater resolution and without the need for averaging across the entire sample, revealing low frequency subclones or changing in the diagnostic immunophenotype that are often the cause of false negative results and recurrent leukemia. In this cohort, we observed a correlation between relapse and greater clonal diversity, presumably due to greater genetic variability. While our study included enrichment for CD34/CD117, it is possible that changes in immunophenotype from diagnosis to remission resulted in recurrent AML that was not positive for these surface markers and would be identified using a less stringent pre-enrichment protocol (e.g. lymphocyte depletion). In summary, Mission Bio has developed a novel, flexible, high-resolution single cell assay for the characterization of clonal diversity and identification of putative therapeutic targets.