Disease Detection Methodologies in Acute Lymphoblastic Leukemia (ALL): Opportunities for Improvement

Abstract

Background Minimal residual disease (MRD) is highly prognostic for outcomes in ALL, particularly prior to HCT. New methods for MRD determination in bone marrow (BM) increase the sensitivity of disease detection but remain dependent on the quality of the specimen obtained. Residual disease testing in the cerebrospinal fluid (CSF) is rarely performed. With the dual objectives to identify the role of BM biopsy in conjunction with BM aspirate and flow cytometry (FC), and also the role of CSF FC alongside cytopathology (CP), we evaluated the frequency and degree of discrepancies in disease detection based on methodologies of BM and CSF disease assessments. Design In this retrospective study of children and young adults with relapsed ALL who were enrolled on one or more phase I/II trials at the NCI between 2012 and 2018, we systematically analyzed all BM and CSF assessments that were performed. All CSF analyses included cell count, CP, and frequently FC. BM samples needed to have a biopsy, aspirate, and FC results to be included in the study. BM and CNS status were classified based on established definitions. For BM samples, a discrepancy was defined as any difference in marrow status (M1, M2, or M3). For CSF, a discrepancy was noted if the CP and the FC were discordant. BM specimens were further analyzed by extent of hemodilution and cellularity with MRD threshold set at Results From 98 patients, a total of 410 BM samples were analyzed, amongst which 61 (14.8%) samples had at least one discordant result (Figure 1). This included 26 discrepancies between biopsy and aspirate, 47 between biopsy and FC, and 39 between aspirate and FC. From these, 26 (6.3%) were deemed clinically relevant, defined as changing the assessment from a morphologic remission to more extensive involvement, and had a median discrepancy of 30.5% blasts (range, 7-91%). In 17 of these cases, the biopsy was essential to determine the full extent of disease. In 35 cases, the aspirate was reported as poor quality and in 30% the flow specimen had hemodilution deemed to be clinically significant. 351 of 418 CSF samples had a concurrent FC analysis of which, 32 (9.1%) had disease discrepancies; all cases demonstrated FC positivity with a negative CP report, with the highest discrepancy at 60% blasts (CSF WBC: 2). Conclusion MRD testing in ALL is critical, however, assessing the depth of remission is dependent on the quality of the sample acquired. This retrospective analysis highlights the frequency and degree of discrepancies observed between traditional histopathologic methods of disease assessment and FC. While the incorporation of highly-sensitive MRD testing will increase disease detection, these data support expanding methodologies for complimentary quantitation of BM and CNS involvement in patients with ALL.