Abstract It is well agreed that measurable residual disease (MRD) testing in acute leukemia patients following induction chemotherapy has profound prognostic implications. Flow cytometry is well suited to be the mainstay for MRD testing due to the heterogenous immunophenotypic and molecular natures of leukemia. Current flow cytometry MRD testing involves examination of more than 15 antigens, requiring multiple tubes to be run for each patient as current clinical flow cytometers are limited to examining 12 parameters per tube. Because multiple tubes are required, hematopathologists must infer antigen expression between tubes, which can be difficult. As MRD testing has such prognostic implications and defines treatment stratifications and outcomes, MRD testing requires very high sensitivity. While several groups have improved sensitivity by increasing the number of cells analyzed, we propose that increasing the antigenic targets in a single immunophenotyping assay will improve sensitivity by enhancing our ability to differentiate normal from abnormal hematopoietic cells. To test this, we utilized spectral flow cytometry, which allows for targeting more parameters with the same number of lasers for a single instrument. We developed a 23-color flow cytometry panel which was designed to thoroughly examine myelomonocytic differentiation and included multiple other antigenic targets which are known to be aberrantly expressed on leukemia cells. The panel also included a live-dead stain to exclude dead cells from analysis. All data was collected with a 3-laser Cytek Northern Lights spectral flow cytometer, which can analyze 10 million events in ~10 minutes. We tested this panel on more than 50 bone marrow aspirate samples, including normal samples, multiple myeloid leukemias, myelodysplastic samples, as well as more than 20 aspirate samples which also had MRD testing completed at a reference laboratory. Our results for MRD testing had excellent concordance with results from the gold-standard reference laboratory method. As our method only required one tube to be run per sample, we were routinely able to analyze at least 10 million cells per sample. Using in silico modelling from our data, we determined the sensitivity of our assay to be an average of 0.005% (range 0.002 – 0.015%) depending on the specific immunophenotype, when analyzing 10 million cells. As we examined all marker expression in a single tube, we were able to model sensitivity as if only 10-parameters were collected per tube and inference was required between tubes for analysis as is the current standard practice. We found that our 23-color assay was significantly more sensitive (p =0.02) compared to running multiple 10-color tubes and inferring between samples. Overall, higher parameter flow cytometric assays (>20 markers) allow for a more sensitive and robust MRD analysis than current gold-standard methods and should be explored for utilization in the clinical setting.
Read full abstract