Enhanced Flow Cytometry and Next Generation Sequencing Assays for Residual B Lymphoblastic Leukemia (B-ALL) Reveal a Subset with Discordant Results Due to Leukemic Changes Post-Therapy

Background: Persistent measurable residual disease (MRD) in B lymphoblastic leukemia (B-ALL) is associated with a higher rate of relapse and poor overall survival. Residual disease that persists below the limit of detection by standard flow cytometry (0.01% of nucleated mononuclear cells) can be detected by an enhanced multiparameter flow cytometric (MFC) assay or Next Generation Sequencing (NGS) at the limit of detection of 0.001% and 0.0001%, respectively. Although some published studies compare the standard flow cytometric assay with NGS, very little data exists using a high-sensitivity flow cytometric assay in comparison with NGS in a routine clinical setting. This study addresses this gap by documenting the relative performance of MFC vs NGS to detect and quantify low levels of MRD including the concordance and discordance between the two assays.

Method: 259 pediatric bone marrow aspirate samples were prospectively and concurrently tested for the presence of MRD by both MFC and NGS. The samples were derived from the acute leukemia and bone marrow transplant services and submitted for testing at the physician's discretion. Many of the patients had been treated with anti-CD19 immunotherapy. The MFC assay used was a 10-color two-tube stain-lyse-wash assay containing reagents from the standard Children's Oncology Group (COG) B-ALL MRD assay supplemented by CD73, CD81, CD123 and CD304. CD22 and CD24 were included in both tubes to allow B-cell identification following anti-CD19 therapy. A third (denominator) tube containing a nucleic acid binding dye (Syto16) was used to standardize MRD quantitation. A target of up to 4,000,000 events was acquired on a Becton-Dickinson 12-color Lyric flow cytometer and the data analyzed using laboratory-developed software (Woodlist 3.1.5). NGS testing was performed at Adaptive Biotechnologies using the clonoSEQ assay that is FDA approved for detecting MRD in B-ALL. The clonoSEQ Clonality test identifies leukemia-associated DNA sequences; at least one clonal sequence is a prerequisite for future monitoring of MRD using the clonoSEQ Tracking MRD test.

Results: Among the 259 total samples from 61 unique patients, 202 were negative by both MFC and NGS to a level of 0.001%, and 41 samples were concordantly positive between MFC and NGS with values ranging from 68.9% to 0.0017% and having a linear relationship with a slope of 0.89 and R² of 0.80. Seven (7) samples were NGS+ and MFC-. Four (4) were from a single patient with CRLF2-rearranged B-ALL having detectable clonal sequences between 0.03% and 0.3%. The patient became NGS negative after 6 months and likely represents clonal sequences retained in myelomonocytic populations derived from the leukemia and not captured by B-cell antigen-directed MFC. For one additional sample there were insufficient B cells identified to explain the NGS result, suggesting either a clonal sequence in a non-B cell population or laboratory cross-contamination. Seven (7) samples were MFC+ and NGS- and two (2) showed changes in the leukemic clonal sequence that differed from that seen at diagnosis and so were not called MRD by NGS, while another had a poorly differentiated immunophenotype with KMT2A-R suggesting likely loss of the clonal sequence. The remaining discordant samples are a result of sample quality issues or are unexplained. Very low-level (0.0001-0.001%) clonal sequences were detected by NGS in 57 samples, all of which were negative by MFC.

Conclusion: There is an excellent overall concordance between the NGS and enhanced MFC assays to a limit of detection of 0.001% using routine clinical samples from a pediatric oncology service where anti-CD19 immunotherapy is frequently administered. MFC is able to identify a subset of discordant cases with clear residual disease that is not identifiable by NGS due to changes in clonal sequence (NGS false negatives). Conversely, NGS identifies a subset of discordant cases that likely represent persistent clonal sequences in myelomonocytic populations whose relationship to relapse risk remains to be determined (potentially NGS false positives). NGS detects very low-level (<0.001%) residual clonal sequences in 22% of patients and the clinical significance remains a matter of investigation. These data suggest that MFC and NGS are complementary and provide more powerful insight into the biology and clinical outcome of B-ALL than either assay alone.

Wood:Cellnomics LLC: Current equity holder in private company; Amgen: Consultancy.